LECTURE NOTES ON

INTERMEDIATE FLUID MECHANICS

Joseph M. Powers

Department of Aerospace and Mechanical Engineering

University of Notre Dame
Notre Dame, Indiana 46556-5637
USA

updated
05 December 2025, 9:05am
2

© 05 December 2025, Joseph M. Powers. All rights reserved.

Contents

Preface 11

I Governing equations: development 13

1 Philosophy of rational continuum mechanics 15
1.1 Mechanics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
1.2 Continuum mechanics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
1.3 Rational continuum mechanics . . . . . . . . . . . . . . . . . . . . . . . . . . 18
1.4 Notions from Newtonian continuum mechanics . . . . . . . . . . . . . . . . . 18

2 Geometry 23
2.1 Vectors and Cartesian tensors . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.1.1 Gibbs and Cartesian index notation . . . . . . . . . . . . . . . . . . . 23
2.1.2 Rotation of axes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.1.3 Vectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.1.4 Tensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
[Link] Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
[Link] Alternating symbol . . . . . . . . . . . . . . . . . . . . . . . 33
[Link] Some secondary definitions . . . . . . . . . . . . . . . . . . 34
[Link].1 Transpose . . . . . . . . . . . . . . . . . . . . . . . 34
[Link].2 Symmetric . . . . . . . . . . . . . . . . . . . . . . . 34
[Link].3 Anti-symmetric . . . . . . . . . . . . . . . . . . . . 35
[Link].4 Decomposition . . . . . . . . . . . . . . . . . . . . 35
[Link] Tensor inner product . . . . . . . . . . . . . . . . . . . . . . 35
[Link] Dual vector of a tensor . . . . . . . . . . . . . . . . . . . . . 36
[Link] Tensor product: two tensors . . . . . . . . . . . . . . . . . . 37
[Link] Vector product: vector and tensor . . . . . . . . . . . . . . 38
[Link].1 Pre-multiplication . . . . . . . . . . . . . . . . . . 38
[Link].2 Post-multiplication . . . . . . . . . . . . . . . . . . 39
[Link] Dyadic product: two vectors . . . . . . . . . . . . . . . . . . 39
[Link] Contraction . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
[Link] Vector cross product . . . . . . . . . . . . . . . . . . . . . . 40

3
4 CONTENTS

[Link] Vector associated with a plane . . . . . . . . . . . . . . . . 40

2.2 Solution of linear algebra equations . . . . . . . . . . . . . . . . . . . . . . . 41
2.3 Eigenvalues, eigenvectors, and tensor invariants . . . . . . . . . . . . . . . . 43
2.4 Grad, div, curl, etc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
2.4.1 Gradient operator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
2.4.2 Divergence operator . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
2.4.3 Curl operator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
2.4.4 Laplacian operator . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
2.4.5 Time derivative . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
2.4.6 Relevant theorems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
[Link] Fundamental theorem of calculus . . . . . . . . . . . . . . . 55
[Link] Gauss’s theorem . . . . . . . . . . . . . . . . . . . . . . . . 55
[Link] Stokes’ theorem . . . . . . . . . . . . . . . . . . . . . . . . . 56
[Link] A useful identity . . . . . . . . . . . . . . . . . . . . . . . . 57
[Link] Leibniz’s rule: general transport theorem for arbitrary regions 57
[Link].1 Material region: Reynolds transport theorem . . . 58
[Link].2 Fixed region . . . . . . . . . . . . . . . . . . . . . . 58
[Link].3 Scalar function . . . . . . . . . . . . . . . . . . . . 58
2.5 General coordinate transformations . . . . . . . . . . . . . . . . . . . . . . . 59

3 Kinematics 67
3.1 Lagrangian description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
3.2 Eulerian description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
3.3 Material derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
3.4 Streamlines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
3.5 Pathlines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
3.6 Streaklines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
3.7 Kinematic decomposition of motion . . . . . . . . . . . . . . . . . . . . . . . 77
3.7.1 Translation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
3.7.2 Solid body rotation and straining . . . . . . . . . . . . . . . . . . . . 79
[Link] Solid body rotation . . . . . . . . . . . . . . . . . . . . . . . 81
[Link] Straining . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
[Link].1 Extensional straining . . . . . . . . . . . . . . . . . 82
[Link].2 Shear straining . . . . . . . . . . . . . . . . . . . . 82
[Link].3 Principal axes of strain rate . . . . . . . . . . . . . 83
[Link].4 Extensional strain rate quadric . . . . . . . . . . . 84
3.8 Expansion rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
3.9 Invariants of the strain rate tensor . . . . . . . . . . . . . . . . . . . . . . . . 88
3.10 Invariants of the velocity gradient tensor . . . . . . . . . . . . . . . . . . . . 89
3.11 Two-dimensional kinematics . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
3.11.1 General two-dimensional flows . . . . . . . . . . . . . . . . . . . . . . 89
[Link] Rotation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89

© 05 December 2025, Joseph M. Powers. All rights reserved.

CONTENTS 5

[Link] Extension . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
[Link] Shear . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
[Link] Expansion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
3.11.2 Relative motion along 1 axis . . . . . . . . . . . . . . . . . . . . . . . 91
3.11.3 Relative motion along 2 axis . . . . . . . . . . . . . . . . . . . . . . . 92
3.11.4 Uniform flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
3.11.5 Pure rigid body rotation . . . . . . . . . . . . . . . . . . . . . . . . . 94
3.11.6 Pure extensional motion (a compressible flow) . . . . . . . . . . . . . 95
3.11.7 Pure shear straining . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
3.11.8 Ideal corner flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
3.11.9 Couette flow: shear + rotation . . . . . . . . . . . . . . . . . . . . . . 98
3.11.10 Ideal irrotational vortex: extension + shear . . . . . . . . . . . . . . 99
3.12 Three-dimensional kinematics: summary . . . . . . . . . . . . . . . . . . . . 100
3.13 Kinematics as a dynamical system . . . . . . . . . . . . . . . . . . . . . . . . 100
3.13.1 General three-dimensional analysis . . . . . . . . . . . . . . . . . . . 100
3.13.2 Two-dimensional limit . . . . . . . . . . . . . . . . . . . . . . . . . . 114

4 Evolution axioms 125

4.1 Mass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
4.2 Linear momenta . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
4.2.1 Statement of the principle . . . . . . . . . . . . . . . . . . . . . . . . 129
4.2.2 Surface forces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
4.2.3 Final form of linear momenta equation . . . . . . . . . . . . . . . . . 135
4.3 Angular momenta . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
4.4 Energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
4.4.1 Total energy term . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
4.4.2 Work term . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
4.4.3 Heat transfer term . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
4.4.4 Conservative form of the energy equation . . . . . . . . . . . . . . . . 143
4.4.5 Secondary forms of the energy equation . . . . . . . . . . . . . . . . . 144
[Link] Enthalpy-based conservative formulation . . . . . . . . . . . 144
[Link] Mechanical energy equation . . . . . . . . . . . . . . . . . . 145
[Link] Thermal energy equation . . . . . . . . . . . . . . . . . . . 146
[Link] Non-conservative energy equation . . . . . . . . . . . . . . . 146
[Link] Energy equation in terms of enthalpy . . . . . . . . . . . . . 147
[Link] Energy equation in terms of entropy . . . . . . . . . . . . . 148
4.5 Entropy inequality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
4.6 Integral forms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
4.6.1 Mass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
[Link] Fixed region . . . . . . . . . . . . . . . . . . . . . . . . . . 153
[Link] Material region . . . . . . . . . . . . . . . . . . . . . . . . . 153
[Link] Moving solid enclosure with holes . . . . . . . . . . . . . . . 153

© 05 December 2025, Joseph M. Powers. All rights reserved.

6 CONTENTS

4.6.2 Linear momenta . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155

4.6.3 Energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
4.6.4 General expression . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
4.7 Summary of axioms in differential form . . . . . . . . . . . . . . . . . . . . . 156
4.7.1 Conservative form . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
[Link] Cartesian index form . . . . . . . . . . . . . . . . . . . . . . 156
[Link] Gibbs form . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
[Link] Non-orthogonal index form . . . . . . . . . . . . . . . . . . 157
4.7.2 Non-conservative form . . . . . . . . . . . . . . . . . . . . . . . . . . 158
[Link] Cartesian index form . . . . . . . . . . . . . . . . . . . . . . 158
[Link] Gibbs form . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
[Link] Non-orthogonal index form . . . . . . . . . . . . . . . . . . 159
4.7.3 Physical interpretations . . . . . . . . . . . . . . . . . . . . . . . . . 159
4.8 Incompleteness of the axioms . . . . . . . . . . . . . . . . . . . . . . . . . . 160

5 Constitutive equations 161

5.1 Frame and material indifference . . . . . . . . . . . . . . . . . . . . . . . . . 161
5.2 Second law restrictions and Onsager relations . . . . . . . . . . . . . . . . . 162
5.2.1 Weak form of the Clausius-Duhem inequality . . . . . . . . . . . . . . 162
[Link] Non-physical motivating example . . . . . . . . . . . . . . . 162
[Link] Real physical effects . . . . . . . . . . . . . . . . . . . . . . 165
5.2.2 Strong form of the Clausius-Duhem inequality . . . . . . . . . . . . . 165
5.3 Fourier’s law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
5.4 Stress-strain rate relation for a Newtonian fluid . . . . . . . . . . . . . . . . 173
5.4.1 Underlying experiments . . . . . . . . . . . . . . . . . . . . . . . . . 173
5.4.2 Analysis for isotropic Newtonian fluid . . . . . . . . . . . . . . . . . . 174
[Link] Diagonal component . . . . . . . . . . . . . . . . . . . . . . 186
[Link] Off-diagonal component . . . . . . . . . . . . . . . . . . . . 186
5.4.3 Stokes’ assumption . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
5.4.4 Second law restrictions . . . . . . . . . . . . . . . . . . . . . . . . . . 187
[Link] One-dimensional systems . . . . . . . . . . . . . . . . . . . 188
[Link] Two-dimensional systems . . . . . . . . . . . . . . . . . . . 188
[Link] Three-dimensional systems . . . . . . . . . . . . . . . . . . 189
5.5 Equations of state . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191

6 Governing equations: summary and special cases 195

6.1 Boundary and interface conditions . . . . . . . . . . . . . . . . . . . . . . . . 195
6.2 Complete set of compressible Navier-Stokes equations . . . . . . . . . . . . . 196
6.2.1 Conservative form . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
[Link] Cartesian index form . . . . . . . . . . . . . . . . . . . . . . 196
[Link] Gibbs form . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
6.2.2 Non-conservative form . . . . . . . . . . . . . . . . . . . . . . . . . . 197

© 05 December 2025, Joseph M. Powers. All rights reserved.

CONTENTS 7

[Link] Cartesian index form . . . . . . . . . . . . . . . . . . . . . . 197

[Link] Gibbs form . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
6.3 Fluid statics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198
6.3.1 Cartesian index form . . . . . . . . . . . . . . . . . . . . . . . . . . . 198
6.3.2 Gibbs form . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198
6.4 Incompressible Navier-Stokes equations with constant properties . . . . . . . 199
6.4.1 Mass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
6.4.2 Linear momenta . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200
6.4.3 Energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200
6.4.4 Poisson equation for pressure . . . . . . . . . . . . . . . . . . . . . . 202
6.4.5 Vector potential . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
6.4.6 Summary of incompressible constant property equations . . . . . . . 204
6.4.7 Limits for one-dimensional diffusion . . . . . . . . . . . . . . . . . . . 204
6.5 Euler equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
6.5.1 Conservative form . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206
[Link] Cartesian index form . . . . . . . . . . . . . . . . . . . . . . 206
[Link] Gibbs form . . . . . . . . . . . . . . . . . . . . . . . . . . . 206
6.5.2 Non-conservative form . . . . . . . . . . . . . . . . . . . . . . . . . . 206
[Link] Cartesian index form . . . . . . . . . . . . . . . . . . . . . . 206
[Link] Gibbs form . . . . . . . . . . . . . . . . . . . . . . . . . . . 206
6.5.3 Alternate forms of the energy equation . . . . . . . . . . . . . . . . . 207
6.6 Dimensionless compressible Navier-Stokes equations . . . . . . . . . . . . . . 208
6.6.1 Mass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
6.6.2 Linear momenta . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
6.6.3 Energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
6.6.4 Thermal state equation . . . . . . . . . . . . . . . . . . . . . . . . . . 214
6.6.5 Caloric state equation . . . . . . . . . . . . . . . . . . . . . . . . . . 215
6.6.6 Upstream conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
6.6.7 Reduction in parameters . . . . . . . . . . . . . . . . . . . . . . . . . 215
6.7 First integrals of linear momenta . . . . . . . . . . . . . . . . . . . . . . . . 215
6.7.1 Bernoulli’s equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216
[Link] Irrotational case . . . . . . . . . . . . . . . . . . . . . . . . 217
[Link] Steady case . . . . . . . . . . . . . . . . . . . . . . . . . . . 218
[Link].1 Streamline integration . . . . . . . . . . . . . . . . 218
[Link].2 Lamb surfaces . . . . . . . . . . . . . . . . . . . . . 219
[Link] Irrotational, steady, incompressible case . . . . . . . . . . . 220
6.7.2 Crocco’s theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
[Link] Stagnation enthalpy variation . . . . . . . . . . . . . . . . . 221
[Link] Extended Crocco’s theorem . . . . . . . . . . . . . . . . . . 223
[Link] Traditional Crocco’s theorem . . . . . . . . . . . . . . . . . 224
6.8 Fluid analog to Maxwell-Heaviside equations . . . . . . . . . . . . . . . . . . 225
6.8.1 Development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225

© 05 December 2025, Joseph M. Powers. All rights reserved.

8 CONTENTS

6.8.2 The fluid Maxwell-Heaviside equations . . . . . . . . . . . . . . . . . 228

6.8.3 Statics with q and J . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
6.8.4 Statics with no q and J . . . . . . . . . . . . . . . . . . . . . . . . . . 232
6.8.5 Dynamics with q and J . . . . . . . . . . . . . . . . . . . . . . . . . . 235
6.8.6 Dynamics with no q or J . . . . . . . . . . . . . . . . . . . . . . . . . 237

II Governing equations: solutions 239

7 Vortex dynamics 241
7.1 Transformations to cylindrical coordinates . . . . . . . . . . . . . . . . . . . 241
7.1.1 Centripetal and Coriolis accelerations . . . . . . . . . . . . . . . . . . 242
7.1.2 Grad and div for cylindrical systems . . . . . . . . . . . . . . . . . . 245
[Link] Grad . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247
[Link] Div . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247
[Link] Alternate derivations . . . . . . . . . . . . . . . . . . . . . . 248
7.1.3 Incompressible Navier-Stokes equations in cylindrical coordinates . . 250
7.2 Ideal rotational vortex . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251
7.3 Ideal irrotational vortex . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 254
7.4 Helmholtz vorticity transport equation . . . . . . . . . . . . . . . . . . . . . 257
7.4.1 General development . . . . . . . . . . . . . . . . . . . . . . . . . . . 257
7.4.2 Incompressible conservative body force limit . . . . . . . . . . . . . . 259
[Link] Isotropic, Newtonian, constant viscosity . . . . . . . . . . . 259
[Link] Two-dimensional, isotropic, Newtonian, constant viscosity . 259
7.4.3 Physical interpretations . . . . . . . . . . . . . . . . . . . . . . . . . 260
[Link] Baroclinic (non-barotropic) effects . . . . . . . . . . . . . . 260
[Link] Bending and stretching of vortex tubes . . . . . . . . . . . . 261
7.5 Kelvin’s circulation theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . 263
7.6 Potential flow of ideal point vortices . . . . . . . . . . . . . . . . . . . . . . . 264
7.6.1 Two interacting ideal vortices . . . . . . . . . . . . . . . . . . . . . . 265
7.6.2 Image vortex . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 266
7.6.3 Vortex sheets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 266
7.6.4 Potential of an ideal irrotational vortex . . . . . . . . . . . . . . . . . 268
7.6.5 Interaction of multiple vortices . . . . . . . . . . . . . . . . . . . . . 270
7.6.6 Pressure field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273
[Link] Single stationary vortex . . . . . . . . . . . . . . . . . . . . 273
[Link] Group of N vortices . . . . . . . . . . . . . . . . . . . . . . 273
7.7 Streamlines and vortex lines at walls . . . . . . . . . . . . . . . . . . . . . . 274

8 Potential flow 279

8.1 Stream functions and velocity potentials . . . . . . . . . . . . . . . . . . . . 280
8.2 Mathematics of complex variables . . . . . . . . . . . . . . . . . . . . . . . . 283
8.2.1 Euler’s formula . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283

© 05 December 2025, Joseph M. Powers. All rights reserved.

CONTENTS 9

8.2.2 Polar and Cartesian representations . . . . . . . . . . . . . . . . . . . 284

8.2.3 Cauchy-Riemann equations . . . . . . . . . . . . . . . . . . . . . . . 288
8.3 Elementary complex potentials . . . . . . . . . . . . . . . . . . . . . . . . . 291
8.3.1 Uniform flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291
8.3.2 Sources and sinks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 292
8.3.3 Point vortices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 293
8.3.4 Superposition of sources . . . . . . . . . . . . . . . . . . . . . . . . . 295
8.3.5 Flow in corners . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296
8.3.6 Doublets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 303
8.3.7 Rankine half body . . . . . . . . . . . . . . . . . . . . . . . . . . . . 304
8.3.8 Flow over a cylinder . . . . . . . . . . . . . . . . . . . . . . . . . . . 306
8.4 Forces induced by potential flow . . . . . . . . . . . . . . . . . . . . . . . . . 310
8.4.1 Contour integrals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 310
[Link] Simple pole . . . . . . . . . . . . . . . . . . . . . . . . . . . 312
[Link] Constant potential . . . . . . . . . . . . . . . . . . . . . . . 312
[Link] Uniform flow . . . . . . . . . . . . . . . . . . . . . . . . . . 312
[Link] Quadrupole . . . . . . . . . . . . . . . . . . . . . . . . . . . 313
8.4.2 Laurent series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313
8.4.3 Pressure distribution for steady flow . . . . . . . . . . . . . . . . . . 314
8.4.4 Blasius force theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . 315
8.4.5 Kutta-Zhukovsky lift theorem . . . . . . . . . . . . . . . . . . . . . . 318
8.5 Heat transfer analog . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321
8.5.1 Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 322
8.5.2 Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323
8.5.3 Conservation of thermal energy . . . . . . . . . . . . . . . . . . . . . 336
8.5.4 Biharmonic equation . . . . . . . . . . . . . . . . . . . . . . . . . . . 337

9 One-dimensional compressible flow 343

9.1 Thermodynamics of general compressible fluids . . . . . . . . . . . . . . . . 344
9.1.1 Maxwell relation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 344
9.1.2 Internal energy from thermal equation of state . . . . . . . . . . . . . 345
9.1.3 Sound speed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 351
9.1.4 Adiabatic gamma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 355
9.2 Generalized one-dimensional equations . . . . . . . . . . . . . . . . . . . . . 357
9.2.1 Mass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 358
9.2.2 Linear momentum . . . . . . . . . . . . . . . . . . . . . . . . . . . . 359
9.2.3 Energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 360
9.2.4 Summary of equations . . . . . . . . . . . . . . . . . . . . . . . . . . 365
[Link] Unsteady conservative form . . . . . . . . . . . . . . . . . . 365
[Link] Unsteady non-conservative form . . . . . . . . . . . . . . . . 365
[Link] Steady conservative form . . . . . . . . . . . . . . . . . . . 366
[Link] Steady non-conservative form . . . . . . . . . . . . . . . . . 366

© 05 December 2025, Joseph M. Powers. All rights reserved.

10 CONTENTS

9.2.5 Influence coefficients . . . . . . . . . . . . . . . . . . . . . . . . . . . 368

9.3 Flow with area change . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 370
9.3.1 Isentropic Mach number relations . . . . . . . . . . . . . . . . . . . . 370
9.3.2 Sonic properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 374
9.3.3 Effect of area change . . . . . . . . . . . . . . . . . . . . . . . . . . . 375
9.3.4 Choking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 377
9.4 Normal shock waves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 379
9.4.1 Rankine-Hugoniot equations . . . . . . . . . . . . . . . . . . . . . . . 380
9.4.2 Rayleigh line . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 382
9.4.3 Hugoniot curve . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 382
9.4.4 Solution procedure for general equations of state . . . . . . . . . . . 384
9.4.5 Calorically perfect ideal gas solutions . . . . . . . . . . . . . . . . . . 384
9.4.6 Acoustic limit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 391
9.5 Flow with area change and normal shocks . . . . . . . . . . . . . . . . . . . 392
9.5.1 Converging nozzle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 392
9.5.2 Converging-diverging nozzle . . . . . . . . . . . . . . . . . . . . . . . 393
9.6 Method of characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 395
9.6.1 Inviscid one-dimensional equations . . . . . . . . . . . . . . . . . . . 395
9.6.2 Homeoentropic flow of a calorically perfect ideal gas . . . . . . . . . . 401
9.6.3 Simple waves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 403
9.6.4 Centered rarefaction . . . . . . . . . . . . . . . . . . . . . . . . . . . 406
9.6.5 Simple compression . . . . . . . . . . . . . . . . . . . . . . . . . . . . 408
9.6.6 Two interacting expansions . . . . . . . . . . . . . . . . . . . . . . . 408
9.6.7 Wall interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 409
9.6.8 Shock tube . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 409
9.6.9 Inviscid Bateman-Burgers’ equation solution . . . . . . . . . . . . . . 412
9.6.10 Viscous Bateman-Burgers’ equation solution . . . . . . . . . . . . . . 419

10 Viscous incompressible laminar flow 423

10.1 Fully developed, one-dimensional solutions . . . . . . . . . . . . . . . . . . . 423
10.1.1 Pressure gradient-driven flow in a slot . . . . . . . . . . . . . . . . . . 424
10.1.2 Couette flow with pressure gradient . . . . . . . . . . . . . . . . . . . 434
10.2 Similarity solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 439
10.2.1 Stokes’ first problem . . . . . . . . . . . . . . . . . . . . . . . . . . . 439
10.2.2 Blasius boundary layer . . . . . . . . . . . . . . . . . . . . . . . . . . 455

Bibliography 473

© 05 December 2025, Joseph M. Powers. All rights reserved.

Preface

These are lecture notes for AME 60635, Intermediate Fluid Mechanics, taught in the De-
partment of Aerospace and Mechanical Engineering of the University of Notre Dame. Most
students are beginning graduate students and advanced engineering undergraduates. Typi-
cally they have completed one undergraduate fluids course as well as courses in linear algebra
and differential equations. The course provides a survey of continuum fluid mechanics. Part
I gives an extensive development of the compressible Navier-Stokes equations. Part II focuses
on their solution in various limits: vorticity dynamics, compressible flow, potential flow, and
viscous laminar flow. The emphasis is on fluid physics and the mathematics necessary to ef-
ficiently describe the physics. The notes make no attempt to address three important topics:
1) discrete computational models of the continuum physics, 2) turbulent fluid mechanics,
or 3) molecular dynamics. The notes do provide the foundation for later courses that ad-
dress computational fluid dynamics (CFD) as well as turbulence; courses that address the
molecular nature of fluids are enlightened by understanding of the continuum limit.
While there is rigor in the development, it is not absolute. The student should call
on other sources for a full description. Much of the development and notation follows
Panton (2024). Other material is drawn from a variety of sources. A detailed bibliog-
raphy is provided. The notes, along with information on the course, can be found at
[Link] At this stage, anyone is free to du-
plicate the notes. A significantly updated and expanded version is also available in Powers,
Mechanics of Fluids, Cambridge U. Press, 2024.
The notes have been transposed from written notes I composed in developing this course
in 1992 and a related course in viscous fluid flow in 1991. Many enhancements have been
made, and thanks go to many students and faculty who have pointed out errors. It is likely
that there are more waiting to be discovered; I would be happy to hear from you regarding
these or suggestions for improvement.

Joseph M. Powers
powers@[Link]
[Link]
University of Notre Dame
Notre Dame, Indiana; USA
© Joseph M. Powers. 05 December 2025. All rights reserved.

11
12 CONTENTS

© 05 December 2025, Joseph M. Powers. All rights reserved.

 Part I

Governing equations: development

13
Chapter 1

Philosophy of rational continuum

mechanics

see Powers, Chapter 1,

see Truesdell, Chapter 1,
see Panton, Chapter 1,
see Paolucci, Chapter 1,
see Eringen, Introduction.

We will study in these notes the mechanics of a fluid, defined as a material that continuously
deforms under the influence of an applied shear stress. Such a definition allows both liquids
and gases to be considered fluids. We seek to present an approach to fluid mechanics founded
on the general principles of rational continuum mechanics. These general principles apply
to all continuous materials: solids, liquids, and gases. The first four chapters will be quite
general and may be applied to all continuous materials. The remaining chapters are specific
to fluids.
There are many paths to understanding fluid mechanics, and good arguments can be
made for each. A typical first undergraduate class will combine a mix of basic equations,
coupled with strong physical motivations, and allows the student to develop a knowledge
that is of great practical value, often driven strongly by intuition. Such an approach works
well within the confines of the intuition we develop in everyday life. It often fails when the
engineer moves into unfamiliar territory. For example, lack of fundamental understanding
of high Mach number flows led to many aircraft and rocket failures in the 1950’s. In such
cases, a return to the formalism of a careful theory, one that clearly exposes the strengths
and weaknesses of all assumptions, is invaluable in both understanding the true fluid physics,
and applying that knowledge to engineering design.
Probably the most formal of approaches is that of the school of thought advocated most
clearly by Truesdell,1 who forcefully advocated for rational continuum mechanics. Truesdell
1
Clifford Ambrose Truesdell, III, 1919-2000, American continuum mechanician and natural philosopher.
Taught at Indiana and Johns Hopkins Universities.

15
16 CHAPTER 1. PHILOSOPHY OF RATIONAL CONTINUUM MECHANICS

developed a broadly based theory that encompassed all materials that could be regarded
as continua, including solids, liquids, and gases, in the limit when averaging volumes were
sufficiently large so that the micro- and nanoscopic structure of these materials was unimpor-
tant. For fluids (both liquid and gas), such length scales are often on the order of microns,
while for solids, it may be somewhat smaller, depending on the type of molecular structure.
The difficulty of the Truesdellian approach is that it is burdened with a difficult notation
and tends to become embroiled in proofs and philosophy, that while ultimately useful, can
preclude learning basic fluid mechanics in the time scale of the human lifetime.
In this course, we will attempt to steer between the fallible pragmatism of undergraduate
fluid mechanics and the harsh formalism of the Truesdellian school. The material will pay
some due homage to rational continuum mechanics and will be geared towards a basic under-
standing of fluid behavior. We shall first spend some time carefully developing the governing
equations for a compressible viscous fluid. We shall then study representative solutions of
these equations in a wide variety of physically motivated limits in order to understand how
the basic conservation principles of mass, linear momenta,2 angular momenta, and energy,
coupled with constitutive relations, influence the behavior of fluids.

1.1 Mechanics
Mechanics is the broad superset of the topic matter of this course. Mechanics is the science
that seeks an explanation for the motion of bodies based upon models grounded in well de-
fined axioms. Axioms, as in geometry, are statements that cannot be proved; they are useful
insofar as they give rise to results that are consistent with our empirical observations. A
hallmark of science has been the struggle to identify the smallest set of axioms that are suf-
ficient to describe our universe. When we find an axiom to be inconsistent with observation,
it must be modified or eliminated. A familiar example of this is the Michelson-3 Morley4
experiment, that motivated Einstein5 to modify the Newtonian6 axioms of conservation of
mass and energy into a conservation of mass-energy.
In Truesdell’s exposition on mechanics, he suggests the following hierarchy:
• bodies exist,
2
Throughout these notes, we use the less common plural “momenta” as a reminder that in our three-
dimensional world, there are three scalar components of the singular “momentum.”
3
Albert Abraham Michelson, 1852-1931, Prussian born American physicist, graduate of the U.S. Naval
Academy and faculty member at Case School of Applied Science, Clark University, and University of Chicago.
4
Edward Williams Morley, 1838-1923, New Jersey-born American physical chemist, graduate of Williams
College, professor of chemistry at Western Reserve College.
5
Albert Einstein, 1879-1955, German physicist who developed the theory of relativity and made funda-
mental contributions to quantum mechanics and Brownian motion in fluid mechanics; spent later life in the
United States.
6
Sir Isaac Newton, 1642-1727, English physicist and mathematician and chief figure of the scientific rev-
olution of the seventeenth and eighteenth centuries. Developed calculus, theories of gravitation and motion
of bodies, and optics. Educated at Cambridge University and holder of the Lucasian chair at Cambridge. In
civil service as Warden of the Mint, he became the terror of counterfeiters, sending many to the gallows.

© 05 December 2025, Joseph M. Powers. All rights reserved.

1.2. CONTINUUM MECHANICS 17

• bodies are assigned to place,

• geometry is the theory of place,

• change of place in time is the motion of the body,

• a description of the motion of a body is kinematics,

• motion is the consequence of forces,

• study of forces on a body is dynamics.

There are many subsets of mechanics, e.g. statistical mechanics, quantum mechanics,
continuum mechanics, fluid mechanics, or solid mechanics. Auto mechanics, while a legit-
imate topic for study, does not generally fall into the class of mechanics we consider here,
though the intersection of the two sets is not the empty set.

1.2 Continuum mechanics

Early mechanicians, such as Newton, dealt primarily with point masses and finite collections
of particles. In one sense this is because such systems are the easiest to study, and it
makes more sense to grasp the simple before the complex. External motivation was also
present in the 18th century, that had a martial need to understand the motion of cannonballs
and a theological need to understand the motion of planets. The discipline that considers
systems of this type is often referred to as classical mechanics. Mathematically, such systems
are generally characterized by a finite number of ordinary differential equations, and the
properties of each particle (e.g. position, velocity) are taken to be functions of time only.
Continuum mechanics, generally attributed to Euler,7 considers instead an infinite num-
ber of particles. In continuum mechanics every physical property (e.g. velocity, density,
pressure) is taken to be a function of both time and space. There is an infinitesimal prop-
erty variation from point to point in space. While variations are generally continuous, finite
numbers of surfaces of discontinuous property variation are allowed. This models, for ex-
ample, the contact between one continuous body and another. Point discontinuities are
not allowed, however. Finite valued material properties are required. Mathematically, such
systems are characterized by a finite number of partial differential equations in which the
properties of the continuum material are functions of both space and time. It is possible to
show that a partial differential equation can be thought of as an infinite number of ordinary
differential equations, so this is consistent with our model of a continuum as an infinite
number of particles.
7
Leonhard Euler, 1707-1783, Swiss-born mathematician and physicist who served in the court of Cather-
ine I of Russia in St. Petersburg, regarded by many as one of the greatest mechanicians.

© 05 December 2025, Joseph M. Powers. All rights reserved.

18 CHAPTER 1. PHILOSOPHY OF RATIONAL CONTINUUM MECHANICS

1.3 Rational continuum mechanics

The modifier “rational” was first applied by Truesdell to continuum mechanics to distinguish
the formal approach advocated by his school, from less formal, though mainly not irrational,
approaches to continuum mechanics. Rational continuum mechanics is developed with tools
similar to those that Euclid8 used for his geometry: formal definitions, axioms, and theorems,
all accompanied by careful language and proofs. This course will generally follow the less
formal, albeit still rigorous, approach of Panton (2024), including the adoption of much of
Panton’s notation.

1.4 Notions from Newtonian continuum mechanics

The following are useful notions from Newtonian continuum mechanics. Here we use Newto-
nian to distinguish our mechanics from Einsteinian or relativistic mechanics. Newton himself
did not study continuum mechanics; however, notions from his studies of the mechanics of
discrete sets of point masses extend to the mechanics of continua.

• Space is three-dimensional and independent of time.

• An inertial frame is a reference frame in which the laws of physics are invariant; further,
a body in an inertial frame with zero net force acting upon it does not accelerate.
• A Galilean9 transformation specifies how to transform from one inertial frame to an-
other inertial frame moving at constant velocity relative to the original frame. If a
second inertial frame has constant velocity vo = uo i + vo j + wo k relative to the original
inertial frame, the Galilean transformation (x, y, z, t) → (x′ , y ′, z ′ , t′ ) is as follows

x′ = x − uo t, (1.1)
y′ = y − vo t, (1.2)
z′ = z − wo t, (1.3)
t′ = t. (1.4)

This must be accompanied with a transformation of the velocities

u′ = u − uo , (1.5)
v ′ = v − vo , (1.6)
w ′ = w − wo . (1.7)
8
Euclid, Greek geometer of profound influence who taught in Alexandria, Egypt, during the reign of
Ptolemy I Soter, who ruled 323-283 BC.
9
Galileo Galilei, 1564-1642, Pisa-born Italian astronomer, physicist, and developer of experimental meth-
ods, first employed a pendulum to keep time, builder and user of telescopes used to validate the Copernican
view of the universe, developer of the principle of inertia and relative motion.

© 05 December 2025, Joseph M. Powers. All rights reserved.

1.4. NOTIONS FROM NEWTONIAN CONTINUUM MECHANICS 19

• Control volumes are useful; we will study three varieties:

– Fixed: constant in space,

– Material: no flux of mass through boundaries, can deform,
– Arbitrary: can move, can deform, can have different fluid contained within.

• Control surfaces enclose control volumes; they have the same three varieties:

– Fixed,
– Material,
– Arbitrary.

• Density is a material property, not used in classical mechanics, that only considers
point masses. We can define density ρ as
PN
i=1 mi
ρ = lim . (1.8)
V →0 V
Here V is the volume of the space considered, N is the number of particles contained
within the volume, and mi is the mass of the ith particle. We can define a length scale
L associated with the volume V to be L = V 1/3 . In commonly encountered physical
scenarios, we expect the density to vary with distance on a macroscale, approach a
limiting value at the microscale, and become ill-defined below a cutoff scale below
which molecular effects are important. That is to say, when V becomes too small, such
that only a few molecules are contained within it, we expect wild oscillations in ρ.
We will in fact assume that matter can be modeled as a continuum: the limit in which
discrete changes from molecule to molecule can be ignored and distances and times
over which we are concerned are much larger than those of the molecular scale. This
will enable the use of calculus in our continuum thermodynamics.
Continuum mechanics will treat macroscopic effects only and ignore individual molec-
ular effects. For example molecules bouncing off a wall exchange momentum with the
wall and induce pressure. We could use Newtonian mechanics for each particle collision
to calculate the net force on the wall. Instead our approach amounts to considering
the average over space and time of the net effect of millions of collisions on a wall.
The continuum theory can break down in important applications where the length
and time scales are of comparable magnitude to molecular time scales. Important
applications where the continuum assumption breaks down include

– rarefied gas dynamics of the outer atmosphere (relevant for low orbit space vehi-
cles), and
– nano-scale heat transfer (relevant in cooling of computer chips).

© 05 December 2025, Joseph M. Powers. All rights reserved.

20 CHAPTER 1. PHILOSOPHY OF RATIONAL CONTINUUM MECHANICS

To get some idea of the scales involved, we note that for air at atmospheric pressure
and temperature that the time and distance between molecular collisions provides the
limits of the continuum. Under these conditions, we observe for air that
– length > 0.1 µm, and
– time > 0.1 ns,
will be sufficient to admit the continuum assumption. For denser gases, these cutoff
scales are smaller. For lighter gases, these cutoff scales are larger. A sketch of a possible
density variation in a gas near atmospheric pressure is given in Fig. 1.1.

100
variation on the
sub-continuum variation on the
molecular scale continuum scale
1

0.01

Figure 1.1: Sketch of possible density variation of a gas near atmospheric pressure.

Details of collision theory can be found in advanced texts such as that of Vincenti and
Kruger (1965), pp. 12-26. They show for air that the mean free path λ is well modeled
by the equation:
M
λ= √ . (1.9)
2πN ρd2
Here M is the molecular mass, N is Avogadro’s number, and d is the molecular
diameter.

Example 1.1
Find the variation of mean free path with density for air.

We turn to Vincenti and Kruger (1965) for numerical parameter values, that are seen to be
M = 28.9 kg/kmole, N = 6.02252 × 1023 molecule/mole, d = 3.7 × 10−10 m. Thus,
 

kg kmole
28.9 kmole 1 1000 mole
λ = √
, (1.10)
2π 6.02252 × 1023 molecule
mole ρ (3.7 × 10−10 m)2

© 05 December 2025, Joseph M. Powers. All rights reserved.

1.4. NOTIONS FROM NEWTONIAN CONTINUUM MECHANICS 21

kg
7.8895 × 10−8 molecule m2
= . (1.11)
ρ
The unit “molecule” is not really a dimension, but really is literally a “unit,” that may well be
thought of as dimensionless. Thus, we can safely say
kg
7.8895 × 10−8 m2
λ= . (1.12)
ρ
A plot of the variation of mean free path λ as a function of ρ is given in Fig. 1.2. Vincenti

0.001 0.01 0.1 1 10

Figure 1.2: Mean free path length, λ, as a function of density, ρ, for air.

and Kruger (1965) go on to consider an atmosphere with density of ρ = 1.288 kg/m3 . For this
density
kg
7.8895 × 10−8 m2
λ = kg
, (1.13)
1.288 m3
−8
= 6.125 × 10 m, (1.14)
−2
= 6.125 × 10 µm. (1.15)

Vincenti and Kruger (1965) also show the mean molecular speed under these conditions is roughly
c = 500 m/s, so the mean time between collisions, τ , is

λ 6.125 × 10−8 m
τ∼ = = 1.225 × 10−10 s. (1.16)
c 500 ms

Density is an example of a scalar property. We shall have more to say later about
scalars. For now we say that a scalar property associates a single number with each
point in time and space. We can think of this by writing the usual notation ρ(x, y, z, t),
that indicates ρ has functional variation with position and time.

© 05 December 2025, Joseph M. Powers. All rights reserved.

22 CHAPTER 1. PHILOSOPHY OF RATIONAL CONTINUUM MECHANICS

• Other properties are not scalar, but are vector properties. For example the velocity
vector
v(x, y, z, t) = u(x, y, z, t)i + v(x, y, z, t)j + w(x, y, z, t)k, (1.17)
associates three scalars u, v, w with each point in space and time. We will see that a
vector can be characterized as a scalar associated with a particular direction in space.
Here we use a boldfaced notation for a vector. This is known as Gibbs10 notation. We
will soon study an alternate notation, developed by Einstein, and known as Cartesian11
index notation.

• Other properties are not scalar or vector, but are what is know as tensorial. The
relevant properties are called tensors. The best known example is the stress tensor,
whose physics and mathematics will be fully described in Ch. 4.2.2. One can think of
a tensor as a quantity that associates a vector with a plane inclined at a selected angle
passing through a given point in space. An example is the viscous stress tensor τ , that
is best expressed as a three by three matrix with nine components:
 
τxx (x, y, z, t) τxy (x, y, z, t) τxz (x, y, z, t)
τ (x, y, z, t) =  τyx (x, y, z, t) τyy (x, y, z, t) τyz (x, y, z, t)  . (1.18)
τzx (x, y, z, t) τzy (x, y, z, t) τzz (x, y, z, t)

10
Josiah Willard Gibbs, 1839-1903, American physicist and chemist with a lifelong association with Yale
University who made fundamental contributions to vector analysis, statistical mechanics, thermodynamics,
and chemistry. Studied in Europe in the 1860s. Probably one of the few great American scientists of the
nineteenth century.
11
René Descartes, 1596-1650, French mathematician and philosopher of great influence. A great doubter
of existence who nevertheless concluded, “I think, therefore I am.” Developed analytic geometry.

© 05 December 2025, Joseph M. Powers. All rights reserved.

Chapter 2

Geometry

see Powers, Chapter 2,

see Panton, Chapter 3,
see Aris, Chapters 1-3, 7, Appendices A and B,
see Hughes and Gaylord, Appendix,
see Segel, Chapters 1 and 2,
see Yih, Appendix 2,
see Paolucci, Chapter 2,
see Powers and Sen, Chapters 1 and 2.

Here we outline some fundamental mathematical principles of geometry that are necessary
to understand continuum mechanics as it will be presented here. In that we will focus on
vectors and tensors and their calculus, we are considering aspects of differential geometry.

2.1 Vectors and Cartesian tensors

2.1.1 Gibbs and Cartesian index notation
Gibbs notation for vectors and tensors is the most familiar from undergraduate courses.
It typically uses boldface, arrows, underscores, or overbars to denote a vector or a tensor.
Unfortunately, it also hides some of the structures that are actually present in the equations.
Einstein realized this in developing the theory of general relativity and developed a useful
alternate, index notation. In these notes we will focus on what is known as Cartesian
index notation, that is restricted to Cartesian coordinate systems. Einstein also developed
a more general index system for non-Cartesian systems. We will briefly touch on this in our
summaries of our equations later in this chapter but refer the reader to books such as that
of Aris (1962) for a full exposition. While it can seem difficult at the outset, in the end
many agree that the use of index notation actually simplifies many common notions in fluid
mechanics. Moreover, its use in the archival literature is widespread, so to be conversant in
fluid mechanics, one must know index notation. Table 2.1 summarizes the correspondences

23
24 CHAPTER 2. GEOMETRY

Quantity Common Gibbs Cartesian Matrix

Parlance Index
zeroth order tensor scalar a a ( a )
a1
 a2 
first order tensor vector a ai  . 
 .. 
 an 
a11 a12 . . . a1n
 a21 a22 . . . a2n 
second order tensor tensor A aij  . .. .. .. 
 .. . . . 
an1 an2 . . . ann
third order tensor tensor - aijk -
fourth order tensor tensor - aijkl -
.. .. .. ..
. . . . -

Table 2.1: Scalar, vector, and tensor notation conventions.

between Gibbs, Cartesian index, and matrix notation. Here we adopt a convention for the
Gibbs notation, that we will find at times conflicts with other conventions, in which italics
font (a) indicates a scalar, bold font (a) indicates a vector, upper case bold (A) indicates a
second order tensor1 . In Cartesian index notation, their is no need to use anything except
italics, as all terms are thought of as scalar components of a more expansive structure, with
the structure indicated by the presence of subscripts.
The essence of the Cartesian index notation is as follows. We can represent a three-
dimensional vector a as a linear combination of scalars and orthonormal basis vectors:

a = ax i + ay j + az k. (2.1)

We choose now to associate the subscript 1 with the x direction, the subscript 2 with the
y direction, and the subscript 3 with the z direction. Further, we replace the orthonormal
basis vectors i, j, and k, by e1 , e2 , and e3 . Then the vector a is represented by
 
X3 a1
a = a1 e1 + a2 e2 + a3 e3 = 
ai ei = ai ei = ai = a2  . (2.2)
i=1 a3

Following Einstein, we have adopted the convention that a summation is understood to exist
when two indices, known as dummy indices, are repeated, and have further left the explicit
representation of basis vectors out of our final version of the notation. We have also included
a representation of a as a 3 × 1 column vector. We adopt the standard that all vectors can
1
Following longstanding fluid mechanics tradition, we will break this convention for the viscous stress
tensor, τ .

© 05 December 2025, Joseph M. Powers. All rights reserved.

2.1. VECTORS AND CARTESIAN TENSORS 25

be thought of as column vectors. Often in matrix operations, we will need row vectors.
They will be formed by taking the transpose, indicated by a superscript T , of a column
vector. In the interest of clarity, full consistency with notions from matrix algebra, as well
as transparent translation to the conventions of necessarily meticulous (as well as popular)
software tools such as MATLAB, we will scrupulously use the transpose notation. This comes
at the expense of a more cluttered set of equations at times. We also note that most authors
do not explicitly use the transpose notation, but its use is implicit.

2.1.2 Rotation of axes

The Cartesian index notation is developed to be valid under transformations from one Carte-
sian coordinate system to another Cartesian coordinate system. It is not applicable to either
general orthogonal systems (such as cylindrical or spherical) or non-orthogonal systems. It
is straightforward, but tedious, to develop a more general system to handle generalized co-
ordinate transformations, and Einstein did just that as well. For our purposes however, the
simpler Cartesian index notation will suffice.
We will consider a coordinate transformation that is a simple rotation of axes. This trans-
formation preserves all angles; hence, right angles in the original Cartesian system will be
right angles in the rotated, but still Cartesian system. It also preserves lengths of geometric
features, with no stretching. We will require, ultimately, that whatever theory we develop
must generate results in which physically relevant quantities such as temperature, pressure,
density, and velocity magnitude, are independent of the particular set of coordinates with
which we choose to describe the system. To motivate this, let us consider a two-dimensional
rotation from an unprimed system to a primed system. So, we seek a transformation that
maps (x1 , x2 )T → (x′1 , x′2 )T . We will rotate the unprimed system counterclockwise through
an angle α to achieve the primed system.2 The rotation is sketched in Fig. 2.1. It is easy
to show that the angle β = π/2 − α. Here a point P is identified by a particular set of co-
ordinates (x∗1 , x∗2 ). One of the keys to all of continuum mechanics is realizing that while the
location (or velocity, or stress, ...) of P may be represented differently in various coordinate
systems, ultimately it must represent the same physical reality. Straightforward geometry
shows the following relation between the primed and unprimed coordinate systems for x′1
′
x∗1 = x∗1 cos α + x∗2 cos β. (2.3)
More generally, we can say for an arbitrary point that
x′1 = x1 cos α + x2 cos β. (2.4)
We adopt the following notation
2
This is an example of a so-called alias transformation. In such a transformation, the coordinate axes
transform, but the underlying object remains unchanged. So a vector may be considered to be invariant, but
its representation in different coordinate systems may be different. Alias transformations are most common
in continuum mechanics. In contrast, an alibi transformation is one in which the coordinate axes remain
fixed, but the object transforms. This mode of thought is most common in fields such as robotics. In short,
alias rotates the axes, but not the body; alibi rotates the body, but not the axes.

© 05 December 2025, Joseph M. Powers. All rights reserved.

26 CHAPTER 2. GEOMETRY

Figure 2.1: Sketch of coordinate transformation that is a rotation of axes.

• (x1 , x′1 ) denotes the angle between the x1 and x′1 axes,
• (x2 , x′2 ) denotes the angle between the x2 and x′2 axes,
• (x3 , x′3 ) denotes the angle between the x3 and x′3 axes,
• (x1 , x′2 ) denotes the angle between the x1 and x′2 axes,
.
• ..
Thus, in two-dimensions, we have

x′1 = x1 cos(x1 , x′1 ) + x2 cos(x2 , x′1 ). (2.5)

In three dimensions, this extends to

x′1 = x1 cos(x1 , x′1 ) + x2 cos(x2 , x′1 ) + x3 cos(x3 , x′1 ). (2.6)

Extending this analysis to calculate x′2 and x′3 gives

x′2 = x1 cos(x1 , x′2 ) + x2 cos(x2 , x′2 ) + x3 cos(x3 , x′2 ), (2.7)

x′3 = x1 cos(x1 , x′3 ) + x2 cos(x2 , x′3 ) + x3 cos(x3 , x′3 ). (2.8)

These can be written in matrix form as

 
cos(x1 , x′1 ) cos(x1 , x′2 ) cos(x1 , x′3 )
( x′1 x′2 x′3 ) = ( x1 x2 x3 )  cos(x2 , x′1 ) cos(x2 , x′2 ) cos(x2 , x′3 )  . (2.9)
cos(x3 , x′1 ) cos(x3 , x′2 ) cos(x3 , x′3 )

© 05 December 2025, Joseph M. Powers. All rights reserved.

2.1. VECTORS AND CARTESIAN TENSORS 27

If we use the shorthand notation, for example, that ℓ11 = cos(x1 , x′1 ), ℓ12 = cos(x1 , x′2 ), etc.,
we have  
ℓ11 ℓ12 ℓ13
( x′ x′ x′3 ) = ( x1 x2 x3 )  ℓ21 ℓ22 ℓ23  . (2.10)
| 1 {z2 } | {z }
ℓ31 ℓ32 ℓ33
x′T xT | {z }
Q
3
In Gibbs notation, defining the matrix of ℓ’s to be Q, and recalling that all vectors are
taken to be column vectors, we can alternatively say4
x′T = xT · Q. (2.11)
Taking the transpose of both sides and recalling the useful identities that (A · b)T = bT · AT
and (AT )T = A, we can also say
x′ = QT · x. (2.12)
We call Q = ℓij the matrix of direction cosines and QT = ℓji the rotation matrix. It can be
shown that coordinate systems that satisfy the right hand rule require further that
det Q = 1. (2.13)
Matrices Q that have | det Q| = 1 are associated with volume-preserving transformations.
Matrices Q that have det Q > 0, are orientation-preserving transformations. Matrices Q
that have det Q = 1 are thus volume- and orientation-preserving, and can be thought
of a rotations. A matrix that had determinant −1 would be volume-preserving but not
orientation-preserving. It could be considered as a reflection. A matrix Q composed of
orthonormal column vectors, with | det Q| = 1 (thus either rotation or reflection matrices)
is commonly known as orthogonal, though perhaps “orthonormal” would have been a more
descriptive nomenclature. Another way to think of the matrix of direction cosines ℓij = Q
is as a matrix of orthonormal basis vectors in its columns:
 . .. .. 
.. . .
ℓij = Q =  n(1) n(2) n(3)  . (2.14)
.. .. ..
. . .
In a result that is both remarkable and important, it can be shown that the transpose of
an orthogonal matrix is its inverse:
QT = Q−1 . (2.15)
3
Panton (2024) has a different notation for the direction cosines ℓij and employs Q for a different purpose;
our usage is probably more common in the broader literature.
4
The more commonly used alternate convention of not explicitly using the transpose notation for vectors
would instead have our x′T = xT · Q written as x′ = x · Q. In fact, our use of the transpose notation
is strictly viable only for Cartesian coordinate systems, while many will allow Gibbs notation to represent
vectors in non-Cartesian coordinates, for which the transpose operation is ill-suited. However, realizing that
these notes will primarily focus on Cartesian systems, and that such operations relying on the transpose
are useful notions from linear algebra, it will be employed in an overly liberal fashion in these notes. The
alternate convention still typically applies, where necessary, the transpose notation for tensors, so it would
also hold that x′ = QT · x.

© 05 December 2025, Joseph M. Powers. All rights reserved.

28 CHAPTER 2. GEOMETRY

Thus, we have
Q · QT = QT · Q = I. (2.16)
The equation x′T = xT · Q is really a set of three linear equations. For instance, the first
is
x′1 = x1 ℓ11 + x2 ℓ21 + x3 ℓ31 . (2.17)
More generally, we could say that
x′j = x1 ℓ1j + x2 ℓ2j + x3 ℓ3j . (2.18)
Here j is a so-called “free index,” that for three-dimensional space takes on values j = 1, 2, 3.
Some rules of thumb for free indices are
• A free index can appear only once in each additive term.
• One free index (e.g. k) may replace another (e.g. j) as long as it is replaced in each
additive term.
We can simplify Eq. (2.18) further by writing
3
X
x′j = xi ℓij . (2.19)
i=1

This is commonly written in the following form:

x′j = xi ℓij . (2.20)
We again note that it is to be understood that whenever an index is repeated, as has the
index i here, that a summation from i = 1 to i = 3 is to be performed and that i is the
“dummy index.” Some rules of thumb for dummy indices are
• dummy indices can appear only twice in a given additive term,
• a pair of dummy indices, say i, i, can be exchanged for another, say j, j, in a given
additive term with no need to change dummy indices in other additive terms.
We define the Kronecker5 delta, δij as

0, i 6= j,
δij = (2.21)
1, i = j.
This is effectively the identity matrix I:
 
1 0 0
δij = I =  0 1 0  . (2.22)
0 0 1
5
Leopold Kronecker, 1823-1891, German mathematician, critic of set theory, who stated “God made the
integers; all else is the work of man.”

© 05 December 2025, Joseph M. Powers. All rights reserved.

2.1. VECTORS AND CARTESIAN TENSORS 29

Direct substitution proves that what is effectively the law of cosines can be written as
ℓij ℓkj = δik . (2.23)
This is also equivalent to Eq. (2.16), Q · QT = I.

Example 2.1
Show for the two-dimensional system described in Fig. 2.1 that ℓij ℓkj = δik holds.

Expanding for the two-dimensional system, we get

ℓi1 ℓk1 + ℓi2 ℓk2 = δik . (2.24)
First, take i = 1, k = 1. We get then
ℓ11 ℓ11 + ℓ12 ℓ12 = δ11 = 1, (2.25)
cos α cos α + cos(α + π/2) cos(α + π/2) = 1, (2.26)
cos α cos α + (− sin(α))(− sin(α)) = 1, (2.27)
cos2 α + sin2 α = 1. (2.28)
This is obviously true. Next, take i = 1, k = 2. We get then
ℓ11 ℓ21 + ℓ12 ℓ22 = δ12 = 0, (2.29)
cos α cos(π/2 − α) + cos(α + π/2) cos(α) = 0, (2.30)
cos α sin α − sin α cos α = 0. (2.31)
This is obviously true. Next, take i = 2, k = 1. We get then
ℓ21 ℓ11 + ℓ22 ℓ12 = δ21 = 0, (2.32)
cos(π/2 − α) cos α + cos α cos(π/2 + α) = 0, (2.33)
sin α cos α + cos α(− sin α) = 0. (2.34)
This is obviously true. Next, take i = 2, k = 2. We get then
ℓ21 ℓ21 + ℓ22 ℓ22 = δ22 = 1, (2.35)
cos(π/2 − α) cos(π/2 − α) + cos α cos α = 1, (2.36)
sin α sin α + cos α cos α = 1. (2.37)
Again, this is obviously true.

Using this, we can easily find the inverse transformation back to the unprimed coordinates
via the following operations:
ℓkj x′j = ℓkj xi ℓij , (2.38)
= ℓij ℓkj xi , (2.39)
= δik xi , (2.40)
′
ℓkj xj = xk , (2.41)
ℓij x′j = xi , (2.42)
xi = ℓij x′j . (2.43)

© 05 December 2025, Joseph M. Powers. All rights reserved.

30 CHAPTER 2. GEOMETRY

The Kronecker delta is also known as the substitution tensor as it has the property that
application of it to a vector simply substitutes one index for another:

xk = δki xi . (2.44)

For students familiar with linear algebra, it is easy to show that the matrix of direction
cosines, ℓij , is a rotation matrix. Each of its columns is a vector that is orthogonal to the
other column vectors. Additionally, each column vector is itself normal. Such a matrix has a
Euclidean norm of unity, and three eigenvalues that have magnitude of unity. Its determinant
is +1, that renders it a rotation; in contrast a reflection matrix would have determinant of
−1. Operation of a rotation matrix on a vector rotates it, but does not stretch it.

2.1.3 Vectors
Three scalar quantities vi where i = 1, 2, 3 are scalar components of a vector if they transform
according to the following rule
vj′ = vi ℓij , (2.45)
under a rotation of axes characterized by direction cosines ℓij . In Gibbs notation, we would
say
v′T = vT · Q, (2.46)
or alternatively
v′ = QT · v. (2.47)

We can also say that a vector associates a scalar with a chosen direction in space by an
expression that is linear in the direction cosines of the chosen direction.

Example 2.2
Consider the set of scalars that describe the velocity in a two-dimensional Cartesian system:
 
vx
vi = , (2.48)
vy

where we return to the typical x, y coordinate system. Determine if vi is a vector.

In a rotated coordinate system, using the same notation of Fig. 2.1, we find that

vx′ = vx cos α + vy cos(π/2 − α) = vx cos α + vy sin α, (2.49)

vy′ = vx cos(π/2 + α) + vy cos α = −vx sin α + vy cos α. (2.50)

This is linear in the direction cosines, and satisfies the definition for a vector.

© 05 December 2025, Joseph M. Powers. All rights reserved.

2.1. VECTORS AND CARTESIAN TENSORS 31

Example 2.3
Do two arbitrary scalars, say the quotient of pressure and density and the product of specific heat
at constant volme and temperature, (p/ρ, cv T )T , form a vector?

If this quantity is a vector, then we can say

 
p/ρ
vi = . (2.51)
cv T

This pair of numbers has an obvious physical meaning in our unrotated coordinate system. If the
system were a calorically perfect ideal gas (CPIG), the first component would represent the difference
between the enthalpy and the internal energy, and the second component would represent the internal
energy. And if we rotate through an angle α, we arrive at a transformed quantity of
p
v1′ = cos α + cv T cos(π/2 − α), (2.52)
ρ
p
v2′ = cos(π/2 + α) + cv T cos(α). (2.53)
ρ

This quantity does not have any known physical significance, and so it seems that these quantities do
not form a vector.

We have the following vector algebra

• Addition

– wi = ui + vi (Cartesian index notation)

– w = u + v (Gibbs notation)

• Dot product (inner product)

– ui vi = b (Cartesian index notation)

– uT · v = b (Gibbs notation)
– both notations require u1 v1 + u2 v2 + u3 v3 = b.

While ui and vi have scalar components that change under a rotation of axes, their in-
ner product (or dot product) is a true scalar and is invariant under a rotation of axes.

Example 2.4
Demonstrate invariance of the dot product uT · v = b by subjecting vectors u and v to a rotation.

© 05 December 2025, Joseph M. Powers. All rights reserved.

32 CHAPTER 2. GEOMETRY

Under rotation, our vectors transform as u′ = QT · u, v′ = QT · v. Thus Q · u′ = Q · QT · u = u.

and Q · v′ = Q · QT · v = v. Then consider the dot product

uT · v = b, (2.54)
′ T ′
(Q · u ) · (Q · v ) = b, (2.55)
u′T · QT · Q ·v′ = b, (2.56)
| {z }
=I
u′T · I · v′ = b, (2.57)
′T ′
u ·v = b. (2.58)

The inner product is invariant under rotation.

Here we have in the Gibbs notation explicitly noted that the transpose is part of the
inner product. Most authors in fact assume the inner product of two vectors implies the
transpose and do not write it explicitly, writing the inner product simply as u · v ≡ uT · v.

2.1.4 Tensors
[Link] Definition
A second order tensor, or a rank two tensor, is nine scalar components that under a rotation
of axes transformation according to the following rule:

Tij′ = ℓki ℓlj Tkl . (2.59)

We could also write this in an expanded form as

3 X
X 3 3 X
X 3
Tij′ = ℓki ℓlj Tkl = ℓTik Tkl ℓlj . (2.60)
k=1 l=1 k=1 l=1

In these expressions, i and j are both free indices; while k and l are dummy indices. The
notation ℓTik is unusual and rarely used. It does allow us to see the correspondence to GIbbs
notation. The Gibbs notation for this transformation is easily shown to be

T′ = QT · T · Q. (2.61)

Analogously to our conclusion for a vector, we say that a tensor associates a vector with
each direction in space by an expression that is linear in the direction cosines of the chosen
direction. For a given tensor Tij , the first subscript is associated with the face of a unit cube
(hence the mnemonic device, first-face); the second subscript is associated with the vector
components for the vector on that face.

© 05 December 2025, Joseph M. Powers. All rights reserved.

2.1. VECTORS AND CARTESIAN TENSORS 33

Tensors can also be expressed as matrices. All rank two tensors are two-dimensional
matrices, but not all matrices are rank two tensors, as they do not necessarily satisfy the
transformation rules. We can say
 
T11 T12 T13
Tij =  T21 T22 T23  . (2.62)
T31 T32 T33
The first row vector, ( T11 T12 T13 ), is the vector associated with the 1 face. The second
row vector, ( T21 T22 T23 ), is the vector associated with the 2 face. The third row vector,
( T31 T32 T33 ), is the vector associated with the 3 face.

Example 2.5
Consider how the equation A · x = b transforms under rotation.

Using
A′ = QT · A · Q, (2.63)
′ T
x = Q · x, (2.64)
b′ = QT · b, (2.65)
we see that by pre-multiplying all equations by Q, and post-multiplying the tensor equation by QT
that
A = Q · A′ · QT , (2.66)
x = Q · x′ , (2.67)
b = Q · b′ , (2.68)
giving us
Q · A′ · QT · Q · x′ = Q · b′ , (2.69)
| {z } | {z } | {z }
A x b
Q · A′ · x′ = Q · b′ , (2.70)
QT · Q · A′ · x′ = Q T · Q · b′ , (2.71)
A′ · x′ = b′ . (2.72)
Obviously, the form is invariant under rotation.

We also have the following items associated with tensors.

[Link] Alternating symbol

The alternating symbol, ǫijk , will soon be seen to be useful, especially when we introduce
the vector cross product. It is defined as follows

 1 if ijk = 123, 231, or 312,
ǫijk = 0 if any two indices identical, (2.73)

−1 if ijk = 321, 213, or 132.

© 05 December 2025, Joseph M. Powers. All rights reserved.

34 CHAPTER 2. GEOMETRY

Another way to remember this is to start with the sequence 123, that is positive. A sequential
permutation, say from 123 to 231, retains the positive nature. A trade, say from 123 to 213,
gives a negative value. It can be thought of as a matrix of vectors:
      
0 0 0
 0  0   −1  
 
 
 0  1 0 
 0 0 1 
 
ǫijk = 

 0  0 0 .
 (2.74)
 −1 
    0  0 
 0 −1 0 
 
 1  0  0 
0 0 0

An identity that will be used extensively

ǫijk ǫilm = δjl δkm − δjm δkl , (2.75)

can be proved a number of ways, including tedious direct substitution for all values of
i, j, k, l, m.

[Link] Some secondary definitions

[Link].1 Transpose The transpose of a second rank tensor, denoted by a superscript
T , is found by exchanging elements about the diagonal. In shorthand index notation, this is
simply
(Tij )T = Tji . (2.76)
Written out in full, if  
T11 T12 T13

Tij = T21 T22 T23  , (2.77)
T31 T32 T33
then  
T11 T21 T31
T 
Tij = Tji = T12 T22 T32  . (2.78)
T13 T23 T33

[Link].2 Symmetric A tensor Dij is symmetric iff

Dij = Dji , (2.79)

D = DT . (2.80)

A symmetric tensor has only six independent scalars. We will reserve D for tensors that are
symmetric. We will see that D is associated with the deformation of a fluid element.

© 05 December 2025, Joseph M. Powers. All rights reserved.

2.1. VECTORS AND CARTESIAN TENSORS 35

[Link].3 Anti-symmetric A tensor Rij is anti-symmetric iff

Rij = −Rji , (2.81)

R = −RT . (2.82)

An anti-symmetric tensor must have zeroes on its diagonal and only three independent
scalars on off-diagonal elements. We will reserve R for tensors that are anti-symmetric. We
will see that R is associated with the rotation of a fluid element. But R is not a rotation
matrix.

[Link].4 Decomposition An arbitrary tensor Tij can be separated into a symmetric

and anti-symmetric pair of tensors:
1 1 1 1
Tij = Tij + Tij + Tji − Tji . (2.83)
2 2 2 2
Rearranging, we get
1 1
Tij = (Tij + Tji ) + (Tij − Tji ) . (2.84)
|2 {z } |2 {z }
symmetric anti−symmetric

The first term must be symmetric, and the second term must be anti-symmetric. This is
easily seen by considering applying this to any matrix of actual numbers. If we define the
symmetric part of the matrix Tij by the following notation
1
T(ij) = (Tij + Tji ) , (2.85)
2
and the anti-symmetric part of the same matrix by the following notation
1
T[ij] = (Tij − Tji ) , (2.86)
2
we then have
Tij = T(ij) + T[ij] . (2.87)

[Link] Tensor inner product

The tensor inner product of two tensors Tij and Sji is defined as follows

Tij Sji = a, (2.88)

where a is a scalar. In Gibbs notation, we would say

T : S = a. (2.89)

We also can prove

T : S = S : T = tr (T · S) = tr (S · T) = a. (2.90)

© 05 December 2025, Joseph M. Powers. All rights reserved.

36 CHAPTER 2. GEOMETRY

It is easily shown, and will be important in upcoming derivations, that the tensor inner
product of any symmetric tensor D with any anti-symmetric tensor R is the scalar zero:

Dij Rji = 0, (2.91)

D : R = 0. (2.92)

Example 2.6
For all 2 × 2 matrices, prove the tensor inner product of general symmetric and anti-symmetric
tensors is zero.

Take    
a b 0 d
D= , R= . (2.93)
b c −d 0
By definition then

D : R = Dij Rji = D11 R11 + D12 R21 + D21 R12 + D22 R22 , (2.94)
= a(0) + b(−d) + bd + c(0), (2.95)
= 0. QED. (2.96)

The theorem is proved.6 The proof can be extended to arbitrary square matrices.

Further, if we decompose a tensor into its symmetric and anti-symmetric parts, Tij =
T(ij) + T[ij] and take T(ij) = Dij = D and T[ij] = Rij = R, so that T = D + R, we note the
following common term can be expressed as a tensor inner product with a dyadic product:

xi Tij xj = xT · T · x, (2.97)
xi (T(ij) + T[ij])xj = xT · (D + R) · x, (2.98)
xi T(ij) xj = xT · D · x, (2.99)
T(ij) xi xj = D : xxT . (2.100)

[Link] Dual vector of a tensor

We define the dual vector, di , of a tensor Tjk as follows7
1 1 1
di = ǫijk Tjk = ǫijk T(jk) + ǫijk T[jk] . (2.101)
2 2 | {z } 2
=0

6
The common abbreviation QED at the end of the proof stands for the Latin quod erat demonstrandum,
“that which was to be demonstrated.”
7
There is a lack of uniformity in the literature in this area. First, note this definition differs from that
given by Panton (2024) by a factor of 1/2. It is closer, but not identical, to the approach found in Aris
(1962), p. 25.

© 05 December 2025, Joseph M. Powers. All rights reserved.

2.1. VECTORS AND CARTESIAN TENSORS 37

The term ǫijk is anti-symmetric for any fixed i; for example for i = 1, we have
   
ǫ111 ǫ112 ǫ113 0 0 0
ǫ1jk =  ǫ121 ǫ122 ǫ123  =  0 0 1. (2.102)
ǫ131 ǫ132 ǫ133 0 −1 0

Thus, when its tensor inner product is taken with the symmetric T(jk) , the result must be
the scalar zero. Hence, we also have
1
di = ǫijk T[jk] . (2.103)
2
Let us find the inverse relation for di , Starting with Eq. (2.101), we take the inner product
of di with ǫilm to get
1
ǫilm di = ǫilm ǫijk Tjk . (2.104)
2
Employing Eq. (2.75) to eliminate the ǫ’s in favor of δ’s, we get
1
ǫilm di = (δlj δmk − δlk δmj ) Tjk , (2.105)
2
1
= (Tlm − Tml ), (2.106)
2
= T[lm] . (2.107)

Hence,
T[lm] = ǫilm di . (2.108)
Note that  
0 d3 −d2
T[lm] 
= ǫ1lm d1 + ǫ2lm d2 + ǫ3lm d3 = −d3 0 d1  . (2.109)
d2 −d1 0
And we can write the decomposition of an arbitrary tensor as the sum of its symmetric
part and a factor related to the dual vector associated with its anti-symmetric part:

Tij = T(ij) + ǫkij dk . (2.110)

|{z} |{z} | {z }
arbitrary tensor symmetric part anti−symmetric part

[Link] Tensor product: two tensors

The tensor product between two arbitrary tensors yields a third tensor. For second order
tensors, we have the tensor product in Cartesian index notation as

Sij Tjk = Pik . (2.111)

Note that j is a dummy index, i and k are free indices, and that the free indices in each
additive term are the same. In that sense they behave somewhat as dimensional units, that

© 05 December 2025, Joseph M. Powers. All rights reserved.

38 CHAPTER 2. GEOMETRY

must be the same for each term. In Gibbs notation, the equivalent tensor product is written
as
S · T = P. (2.112)
In contrast to the tensor inner product, that has two pairs of dummy indices and two dots,
the tensor product has one pair of dummy indices and one dot. The tensor product is
equivalent to matrix multiplication in matrix algebra.
An important property of tensors is that, in general, the tensor product does not commute,
S · T 6= T · S. In the most formal manifestation of Cartesian index notation, one should
also not commute the elements, and the dummy indices should appear next to another in
adjacent terms as shown. However, it is of no great consequence to change the order of
terms so that we can write Sij Tjk = Tjk Sij . That is in Cartesian index notation, elements do
commute. But, in Cartesian index notation, the order of the indices is extremely important,
and it is this order that does not commute: Sij Tjk 6= Sji Tjk in general. The version presented
for Sij Tjk in Eq. (2.111), in which the dummy index j is juxtaposed between each term, is
slightly preferable as it maintains the order we find in the Gibbs notation.

Example 2.7
For two general 2 × 2 tensors, S and T, find the tensor product.

The tensor product is

    
S11 S12 T11 T12 S11 T11 + S12 T21 S11 T12 + S12 T22
S·T = = . (2.113)
S21 S22 T21 T22 S21 T11 + S22 T21 S21 T12 + S22 T22
Compare with the commutation:
    
T11 T12 S11 S12 S11 T11 + S21 T12 S12 T11 + S22 T12
T·S = = . (2.114)
T21 T22 S21 S22 S11 T21 + S21 T22 S12 T21 + S22 T22
T
Clearly S · T 6= T · S. It can be shown that if both S and T are symmetric, that S · T = (T · S) .

[Link] Vector product: vector and tensor

The product of a vector and tensor, again that does not in general commute, comes in two
flavors, pre-multiplication and post-multiplication, both important, and given in Cartesian
index and Gibbs notation next:

[Link].1 Pre-multiplication
uj = vi Tij = Tij vi , (2.115)
uT = vT · T 6= T · v. (2.116)
In the Cartesian index notation here, the first form is preferred as it has a correspondence
with the Gibbs notation, but both are correct representations given our summation conven-
tion.

© 05 December 2025, Joseph M. Powers. All rights reserved.

2.1. VECTORS AND CARTESIAN TENSORS 39

[Link].2 Post-multiplication

wi = Tij vj = vj Tij , (2.117)

w = T · v 6= vT · T. (2.118)

[Link] Dyadic product: two vectors

As opposed to the inner product between two vectors, that yields a scalar, we also have the
dyadic product, that yields a tensor. In Cartesian index and Gibbs notation, we have

Tij = ui vj = vj ui , (2.119)
T = uvT 6= vuT . (2.120)

Notice there is no dot in the dyadic product; the dot is reserved for the inner product.

Example 2.8
Find the dyadic product between two general two-dimensional vectors. Show the dyadic product
does not commute in general, and find the condition under which it does commute.

Take    
u1 v1
u= , v= . (2.121)
u2 v2

Then    
T u1 u1 v1 u1 v2
uv = ui vj = ( v1 v2 ) = . (2.122)
u2 u2 v1 u2 v2

Compare this to the commuted operation, vuT :

   
v1 v1 u1 v1 u2
vuT = vi uj = ( u1 u2 ) = . (2.123)
v2 v2 u1 v2 u2

By inspection, we see the operations in general do not commute. They do commute if v2 /v1 = u2 /u1 .
So in order for the dyadic product to commute, u and v must be parallel.
It is easily seen that the dyadic product vvT is a symmetric tensor. For the two-dimensional
system, we would have
   
v1 v1 v1 v1 v2
vvT = vi vj = ( v1 v2 ) = . (2.124)
v2 v2 v1 v2 v2

© 05 December 2025, Joseph M. Powers. All rights reserved.

40 CHAPTER 2. GEOMETRY

[Link] Contraction
We contract a general tensor, that has all of its subscripts different, by setting one subscript
to be the same as the other. A single contraction will reduce the order of a tensor by two.
For example the contraction of the second order tensor Tij is Tii , that indicates a sum is to
be performed:
Tii = T11 + T22 + T33 , (2.125)
tr T = T11 + T22 + T33 . (2.126)
So, in this case the contraction yields a scalar. In matrix algebra, this particular contraction
is the trace of the matrix.

[Link] Vector cross product

The vector cross product is defined in Cartesian index and Gibbs notation as
wi = ǫijk uj vk , (2.127)
w = u × v. (2.128)
Expanding for i = 1, 2, 3 gives
w1 = ǫ123 u2 v3 + ǫ132 u3 v2 = u2 v3 − u3 v2 , (2.129)
w2 = ǫ231 u3 v1 + ǫ213 u1 v3 = u3 v1 − u1 v3 , (2.130)
w3 = ǫ312 u1 v2 + ǫ321 u2 v1 = u1 v2 − u2 v1 . (2.131)

[Link] Vector associated with a plane

We often have to select a vector that is associated with a particular direction. Now for any
direction we choose, there exists an associated unit vector and normal plane. Recall that
our notation has been defined so that the first index is associated with a face or direction,
and the second index corresponds to the components of the vector associated with that face.
If we take ni to be a unit normal vector associated with a given direction and normal plane,
and we have been given a tensor Tij , the vector tj associated with that plane is given in
Cartesian index and Gibbs notation by
tj = ni Tij , (2.132)
tT = nT · T, (2.133)
t = TT · n. (2.134)
A sketch of a Cartesian element with the tensor components sketched on the proper face is
shown in 2.2.

Example 2.9
Find the vector associated with the 1 face, t(1) , as shown in Fig. 2.2,

© 05 December 2025, Joseph M. Powers. All rights reserved.

2.2. SOLUTION OF LINEAR ALGEBRA EQUATIONS 41

Figure 2.2: Sample Cartesian element that is aligned with coordinate axes, along with
tensor components and vectors associated with each face.

T
We first choose the unit normal associated with the x1 face, that is the vector ni = (1, 0, 0) . The
associated vector is found by doing the actual summation

tj = ni Tij = n1 T1j + n2 T2j + n3 T3j . (2.135)

Now n1 = 1, n2 = 0, and n3 = 0, so for this problem, we have

(1)
tj = T1j . (2.136)

2.2 Solution of linear algebra equations

We briefly discuss the solution of linear algebra equations of the form

Aij xj = bi , (2.137)
A · x = b. (2.138)

Full details can be found in any text addressing linear algebra, e.g. Powers and Sen (2015).
Let us presume that A is a known square matrix of dimension N × N, x is an unknown
column vector of length N, and b is a known column vector of length N. The following can
be proved:

© 05 December 2025, Joseph M. Powers. All rights reserved.

42 CHAPTER 2. GEOMETRY

• A unique solution for x exists iff det A 6= 0.

• If det A = 0, solutions for x may or may not exist; if they exist, they are not unique.

• Cramer’s8 rule, a method involving the ratio of determinants discussed in linear algebra
texts, can be used to find x; other methods exist, such as Gaussian elimination.
Let us consider a few examples for N = 2.

Example 2.10
Use Cramer’s rule to solve a general linear algebra problem with N = 2.

Consider then     
a11 a12 x1 b1
= . (2.139)
a21 a22 x2 b2
The solution from Cramer’s rule involves the ratio of determinants. We get

b1 a12 a11 b1
b2 a22 b1 a22 − b2 a12 a21 b2 b2 a11 − b1 a21
x1 = = , x2 = = . (2.140)
a11 a12 a 11 a22 − a12 a21 a11 a12 a 11 a22 − a12 a21
a21 a22 a21 a22

If b1 , b2 6= 0 and det A = a11 a22 − a12 a21 6= 0, there is a unique nontrivial solution for x. If b1 = b2 = 0
and det A = a11 a22 − a12 a21 6= 0, we must have x1 = x2 = 0. Obviously, if det A = a11 a22 − a12 a21 = 0,
we cannot use Cramer’s rule to compute as it involves division by zero. But we can salvage a non-unique
solution if we also have b1 = b2 = 0, as we shall see.

Example 2.11
Find any and all solutions for
    
1 2 x1 0
= . (2.141)
2 4 x2 0

Certainly (x1 , x2 )T = (0, 0)T is a solution. But maybe there are more. Cramer’s rule gives

0 2 1 0
0 4 0 2 0 0
x1 = = , x2 = = . (2.142)
1 2 0 1 2 0
2 4 2 4

This is indeterminate! But the more robust Gaussian elimination process allows us to use row operations
(multiply the top row by −2 and add to the bottom row) to rewrite the original equation as
    
1 2 x1 0
= . (2.143)
0 0 x2 0
8
Gabriel Cramer, 1704-1752, Swiss mathematician at University of Geneva.

© 05 December 2025, Joseph M. Powers. All rights reserved.

2.3. EIGENVALUES, EIGENVECTORS, AND TENSOR INVARIANTS 43

By inspection, we get an infinite number of solutions, given by the one-parameter family of equations

x1 = −2s, x2 = s, s ∈ R1 . (2.144)

We could also eliminate s and say that x1 = −2x2 . The solutions are linearly dependent. In terms of
the language of vectors, we find the solution to be a vector of fixed direction, with arbitrary magnitude.
In terms of a unit normal vector, we could write the solution as
!
− √25
x=s 1 , s ∈ R1 . (2.145)
√
5

Example 2.12
Find any and all solutions for
    
1 2 x1 1
= . (2.146)
2 4 x2 0

Cramer’s rule gives

1 2 1 1
0 4 4 2 0 −2
x1 = = , x2 = = . (2.147)
1 2 0 1 2 0
2 4 2 4

There is no value of x that satisfies A · x = b!9

2.3 Eigenvalues, eigenvectors, and tensor invariants

For a given tensor Tij , it is possible to select a plane for which the vector from Tij associated
with that plane points in the same direction as the normal associated with the chosen plane.
In fact for a three-dimensional element, it is possible to choose three planes for which the
vector associated with the given planes is aligned with the unit normal associated with those
planes. We can think of this as finding a rotation as sketched in Fig. 2.3.
9
There is however, in some sense a best solution, that is, an x of minimum length that also minimizes
||A · x − b||. Using the pseudoinverse procedure described in Powers and Sen (2015), Ch. 7, we find there
exists a non-unique set of x = (1/25 − 2s, 2/25 + s)T , s ∈ R1 , for which
√ for all values of s, the so-called
error norm takes on the same minimum value, e = ||A · x − b|| = 2/ 5. For s = 0, we then get the “best”
x = (1/25, 2/25)T in that this x minimizes the error and is itself of minimum length.

© 05 December 2025, Joseph M. Powers. All rights reserved.

44 CHAPTER 2. GEOMETRY

rotate

Figure 2.3: Sample Cartesian element that is rotated so that its faces have vectors that are
aligned with the unit normals associated with the faces of the element.

Mathematically, we can enforce this condition by requiring that

ni Tij = λnj . (2.148)

| {z } |{z}
vector associated with chosen direction scalar multiple of chosen direction

Here λ is an as of yet unknown scalar. The vector ni could be a unit vector, but does not
have to be. We can rewrite this as

ni Tij = λni δij . (2.149)

In Gibbs notation, this becomes nT · T = λnT · I. In mathematics, this is known as a left

eigenvalue problem. Solutions ni that are non-trivial are known as left eigenvectors. We
can also formulate this as a right eigenvalue problem by taking the transpose of both sides
to obtain TT · n = λI · n. Here we have used the fact that IT = I. We note that the left
eigenvectors of T are the right eigenvectors of TT . Eigenvalue problems are quite general
and arise whenever an operator operates on a vector to generate a vector that leaves the
original unchanged except in magnitude.
We can rearrange to form
ni (Tij − λδij ) = 0. (2.150)
In matrix notation, this can be written as
 
T11 − λ T12 T13
( n1 n2 n3 )  T21 T22 − λ T23  = ( 0 0 0 ) . (2.151)
T31 T32 T33 − λ

A trivial solution to this equation is (n1 , n2 , n3 ) = (0, 0, 0). But this is not interesting.
As suggested by our understanding of Cramer’s rule, we can get a non-unique, non-trivial

© 05 December 2025, Joseph M. Powers. All rights reserved.

2.3. EIGENVALUES, EIGENVECTORS, AND TENSOR INVARIANTS 45

solution if we enforce the condition that the determinant of the coefficient matrix be zero.
As we have an unknown parameter λ, we have sufficient degrees of freedom to accomplish
this. So, we require
T11 − λ T12 T13
T21 T22 − λ T23 = 0. (2.152)
T31 T32 T33 − λ
We know from linear algebra that such an equation for a third order matrix gives rise to a
characteristic polynomial for λ of the form10
(1) (2) (3)
λ3 − IT λ2 + IT λ − IT = 0, (2.153)

(1) (2) (3)

where IT , IT , IT are scalars that are functions of all the scalars Tij . The IT ’s are known
as the invariants of the tensor Tij . They can be shown, following a detailed analysis, to be
given by
(1)
IT = Tii = tr T, (2.154)
(2) 1 1


IT = (Tii Tjj − Tij Tji ) = (tr T)2 − tr (T · T) = (det T) tr T−1 , (2.155)
2 2
1

= T(ii) T(jj) + T[ij] T[ij] − T(ij) T(ij) , (2.156)
2
(3)
IT = ǫijk T1i T2j T3k = det T. (2.157)

Here “det” denotes the determinant. It can also be shown that if λ(1) , λ(2) , λ(3) are the three
eigenvalues, then the invariants can also be expressed as
(1)
IT = λ(1) + λ(2) + λ(3) , (2.158)
(2)
IT = λ(1) λ(2) + λ(2) λ(3) + λ(3) λ(1) , (2.159)
(3)
IT = λ(1) λ(2) λ(3) . (2.160)

In general these eigenvalues, and consequently, the eigenvectors are complex. Addition-
ally, in general the eigenvectors are non-orthogonal. If, however, the matrix we are consid-
ering is symmetric, that is often the case in fluid mechanics, it can be formally proven that
all the eigenvalues are real and all the eigenvectors are real and orthogonal. If for instance,
our tensor is the stress tensor, we will show that it is symmetric in the absence of external
couples. The eigenvectors of the stress tensor can form the basis for an intrinsic coordinate
system that has its axes aligned with the principal stress on a fluid element. The eigenvalues
themselves give the value of the principal stress. This is actually a generalization of the
familiar Mohr’s11 circle from solid mechanics.
10
We employ a slightly more common form here than the similar Eq. (3.66) of Panton (2024).
11
Christian Otto Mohr, 1835-1918, Holstein-born German civil engineer, railroad and bridge designer.

© 05 December 2025, Joseph M. Powers. All rights reserved.

46 CHAPTER 2. GEOMETRY

Figure 2.4: Sketch of stresses being applied to a cubical fluid element. The thinner lines
with arrows are the components of the stress tensor; the thicker lines on each face represent
the vector associated with the particular face.

Example 2.13
Find the principal axes and principal values of stress if the stress tensor is
 
1 0 0
Tij =  0 1 2  . (2.161)
0 2 1

A sketch of these stresses is shown on the fluid element in Fig. 2.4. We take the eigenvalue problem

ni Tij = λnj , (2.162)

= λni δij , (2.163)
ni (Tij − λδij ) = 0. (2.164)

This becomes for our problem

 
1−λ 0 0
( n1 n2 n3 )  0 1−λ 2  = (0 0 0). (2.165)
0 2 1−λ
For a non-trivial solution for ni , we must have
1−λ 0 0
0 1−λ 2 = 0. (2.166)
0 2 1−λ
This gives rise to the polynomial equation

(1 − λ) ((1 − λ)(1 − λ) − 4) = 0. (2.167)

© 05 December 2025, Joseph M. Powers. All rights reserved.

2.3. EIGENVALUES, EIGENVECTORS, AND TENSOR INVARIANTS 47

This has three solutions

λ = 1, λ = −1, λ = 3. (2.168)
Notice all eigenvalues are real, that we expect because the tensor is symmetric.
Now let us find the eigenvectors (aligned with the principal axes of stress) for this problem First, it
can easily be shown that when the vector product of a vector with a tensor commutes when the tensor
is symmetric. Although this is not a crucial step, we will use it to write the eigenvalue problem in a
slightly more familiar notation:

ni (Tij − λδij ) = 0 =⇒ (Tij − λδij ) ni = 0, because scalar components commute. (2.169)

Because of symmetry, we can now commute the indices to get

(Tji − λδji ) ni = 0, because indices commute if symmetric. (2.170)

Expanding into matrix notation, we get

    
T11 − λ T21 T31 n1 0
 T12 T22 − λ T32   n2  =  0  . (2.171)
T13 T23 T33 − λ n3 0

We have taken the transpose of T . Substituting for Tji and considering the eigenvalue λ = 1, we get
    
0 0 0 n1 0
 0 0 2   n2  =  0  . (2.172)
0 2 0 n3 0

We get two equations 2n2 = 0, and 2n3 = 0; thus, n2 = n3 = 0. We can satisfy all equations with an
arbitrary value of n1 . It is always the case that an eigenvector will have an arbitrary magnitude and a
well-defined direction. Here we will choose to normalize our eigenvector and take n1 = 1, so that the
eigenvector is  
1
nj =  0  for λ = 1. (2.173)
0
Geometrically, this means that the original 1 face already has an associated vector that is aligned with
its normal vector.
Now consider the eigenvector associated with the eigenvalue λ = −1. Again substituting into the
original equation, we get     
2 0 0 n1 0
 0 2 2   n2  =  0  . (2.174)
0 2 2 n3 0
This is simply the system of equations

2n1 = 0, (2.175)
2n2 + 2n3 = 0, (2.176)
2n2 + 2n3 = 0. (2.177)

Clearly n1 = 0. We could take n2 = 1 and n3 = −1 for a non-trivial solution. Alternatively, let’s

normalize and take  
0
√
2
nj =  2√  . (2.178)
− 22

© 05 December 2025, Joseph M. Powers. All rights reserved.

48 CHAPTER 2. GEOMETRY

Finally consider the eigenvector associated with the eigenvalue λ = 3. Again substituting into the
original equation, we get     
−2 0 0 n1 0
 0 −2 2   n2  =  0  . (2.179)
0 2 −2 n3 0
This is the system of equations

−2n1 = 0, (2.180)
−2n2 + 2n3 = 0, (2.181)
2n2 − 2n3 = 0. (2.182)

Clearly again n1 = 0. We could take n2 = 1 and n3 = 1 for a non-trivial solution. Once again, let us
normalize and take  
√0
2
nj =  √2  . (2.183)
2
2
In summary, the three eigenvectors and associated eigenvalues are
 
1
(1)
nj = 0 for λ(1) = 1, (2.184)
0
 
√0
(2) 2 
nj =  √ 2 for λ(2) = −1, (2.185)
− 2
2
 
√0
(3) 2
nj =  √2
 for λ(3) = 3. (2.186)
2
2

The eigenvectors are mutually orthogonal, as well as normal. We say they form an orthonormal set of
vectors. Their orthogonality, as well as the fact that all the eigenvalues are real can be shown to be
a direct consequence of the symmetry of the original tensor. A sketch of the principal stresses on the
element rotated so that it is aligned with the principal axes of stress is shown on the fluid element in
Fig. 2.5. The three orthonormal eigenvectors when cast into a matrix, form an orthogonal matrix Q,
and calculation reveals that det Q = 1, so that it is a rotation matrix.
 . .. ..   
.. . . 1 √0 √0
 (1)  2 2 
Q = n n(2) n(3)  =  0 2
√ √2
. (2.187)
.. .. .. 0 − 2 2
. . . 2 2

Example 2.14
For a given stress tensor, that we will take to be symmetric though the theory applies to non-
symmetric tensors as well,  
1 2 4
Tij = T =  2 3 −1  , (2.188)
4 −1 1

© 05 December 2025, Joseph M. Powers. All rights reserved.

2.3. EIGENVALUES, EIGENVECTORS, AND TENSOR INVARIANTS 49

Figure 2.5: Sketch of fluid element rotated to be aligned with axes of principal stress, along
with magnitude of principal stress. The 1 face projects out of the page.

(1) (2) (3)

find the three basic tensor invariants of stress IT , IT , and IT , and show they are truly invariant
when the tensor is subjected to a rotation with direction cosine matrix of
 q 
√1 2 √1
3
 6 6

ℓij = Q =  √1 1
− √3 √1 . (2.189)
3 3
√1 0 − √12
2

Calculation reveals that

det Q = 1, (2.190)
and that Q · QT = I, so that QT is a rotation matrix. The eigenvalues of T, that are the principal
values of stress, are easily calculated to be

λ(1) = 5.28675, λ(2) = −3.67956, λ(3) = 3.39281. (2.191)

The three invariants of Tij are

 
1 2 4
(1)
IT = tr T = tr  2 3 −1  = 1 + 3 + 1 = 5, (2.192)
4 −1 1
(2) 1

IT = (tr T)2 − tr (T · T) , (2.193)
2
  2   
1 2 4 1 2 4 1 2 4
1   
=  tr 2 3 −1  − tr  2 3 −1   2 3 −1  , (2.194)
2
4 −1 1 4 −1 1 4 −1 1
  
21 4 6
1 2
= 5 − tr  4 14 4  , (2.195)
2
6 4 18
1
= (25 − 21 − 14 − 18), (2.196)
2
= −14, (2.197)

© 05 December 2025, Joseph M. Powers. All rights reserved.

50 CHAPTER 2. GEOMETRY
 
1 2 4
(3)
IT = det T = det  2 3 −1  = −66. (2.198)
4 −1 1
Now when we rotate the tensor T, we get a transformed tensor given by
 √1  q
√1 √1   √1 2 √1

q6 3 2 1 2 4 3
   6 6

T′ = QT · T · Q =  23 − √1 0  2 3 −1   √1 − √13 √1 , (2.199)
3 3 3
√1 √1 − √1 4 −1 1 √1 0 − √12
 6 3 2
 2
4.10238 2.52239 1.60948
=  2.52239 −0.218951 −2.91291  . (2.200)
1.60948 −2.91291 1.11657

We then seek the tensor invariants of T′ . Leaving out some of the details, that are the same as those
for calculating the invariants of the T, we find the invariants indeed are invariant:
(1)
IT = 4.10238 − 0.218951 + 1.11657 = 5, (2.201)
(2) 1 2
IT = (5 − 53) = −14, (2.202)
2
(3)
IT = −66. (2.203)

Finally, we verify that the stress invariants are indeed related to the principal values (the eigenvalues
of the stress tensor) as follows
(1)
IT = λ(1) + λ(2) + λ(3) = 5.28675 − 3.67956 + 3.39281 = 5, (2.204)
(2) (1) (2) (2) (3) (3) (1)
IT = λ λ +λ λ +λ λ , (2.205)
= (5.28675)(−3.67956) + (−3.67956)(3.39281) + (3.39281)(5.28675) = −14, (2.206)
(3) (1) (2) (3)
IT = λ λ λ = (5.28675)(−3.67956)(3.39281) = −66. (2.207)

Example 2.15
For a given two-dimensional stress tensor, that here we will take to be asymmetric,
 
2 1
Tij = T = , (2.208)
2 2
(1) (2)
find the two basic tensor invariants of stress IT and IT and show they are truly invariant when the
tensor is subjected to a rotation with direction cosine matrix of
1 1
!
√ √
2 2
ℓij = Q = . (2.209)
− √12 √1
2

Calculation reveals that det Q = 1 and that Q · QT = I, so that QT is a rotation matrix. The
eigenvalue problem induces the condition
T11 − λ T12
= 0. (2.210)
T21 T22 − λ

© 05 December 2025, Joseph M. Powers. All rights reserved.

2.4. GRAD, DIV, CURL, ETC. 51

This give the characteristic polynomial

(T11 − λ)(T22 − λ) − T12 T21 = 0, (2.211)
2
λ − (T11 + T22 )λ + (T11 T22 − T12 T21 ) = 0, (2.212)
(1) (2)
λ2 − IT λ + IT = 0. (2.213)
Here we have for the two-dimensional system, the two invariants
(1)
IT = T11 + T22 = λ(1) + λ(2) = tr T, (2.214)
(2) (1) (2)
IT = T11 T22 − T12 T21 = λ λ = det T. (2.215)
For this system, the eigenvalues of T are easily calculated to be
√ √
λ(1) = 2 + 2, λ(2) = 2 − 2. (2.216)
The two invariants of T are
 
(1) 2 1
IT = tr T = tr = 2 + 2 = 4, (2.217)
2 2
 
(2) 2 1
IT = det T = det = 2(2) − 1(2) = 2. (2.218)
2 2
Now when we rotate the tensor T, we get a transformed tensor given by
!  !  
√1 − √12 √1 √1 1
′ T 2 2 1 2 2 − 21
T = Q ·T·Q = = 12 7 . (2.219)
√1 √1 2 2 − 2
√1 √1
2 2 2 2 2

By inspection, we see the invariants of T′ are indeed the same as those of T:

(1) 1 7
IT = + = 4, (2.220)
 2  
2  
(2) 1 7 −1 1
IT = − = 2. (2.221)
2 2 2 2
Finally, we also see
(1) √ √
IT = λ(1) + λ(2) = (2 + 2) + (2 − 2) = 4, (2.222)
 √  √ 
(2)
IT = λ(1) λ(2) = 2 + 2 2 − 2 = 2. (2.223)

2.4 Grad, div, curl, etc.

Thus far, we have mainly dealt with the algebra of vectors and tensors. Now let us consider
the calculus. For now, let us consider variables that are a function of the spatial vector xi .
We shall soon allow variation with time t also. We will typically encounter quantities such
as
• φ(xi ) → a scalar function of the position vector,
• vj (xi ) → a vector function of the position vector, or
• Tjk (xi ) → a tensor function of the position vector.

© 05 December 2025, Joseph M. Powers. All rights reserved.

52 CHAPTER 2. GEOMETRY

2.4.1 Gradient operator

The gradient operator, sometimes denoted by “grad,” is motivated as follows. Consider
φ(xi ), that when written in full is

φ(xi ) = φ(x1 , x2 , x3 ). (2.224)

Taking a derivative using the chain rule gives

∂φ ∂φ ∂φ
dφ = dx1 + dx2 + dx3 . (2.225)
∂x1 ∂x2 ∂x3

Following Panton (2024), we define a non-traditional, but useful further notation ∂i for the
partial derivative
 ∂   
∂x1 ∂1
∂ ∂ ∂ ∂
∂i ≡ = e1 + e2 + e3 = ∇ = grad =  ∂
∂x2
 =  ∂2  . (2.226)
∂xi ∂x1 ∂x2 ∂x3 ∂ ∂3
∂x3

So the chain rule is actually

dφ = ∂1 φ dx1 + ∂2 φ dx2 + ∂3 φ dx3 , (2.227)

that is written using our summation convention as

dφ = ∂i φ dxi . (2.228)

After commuting so as to juxtapose the i subscript, we have

dφ = dxi ∂i φ. (2.229)

In Gibbs notation, we say

dφ = dxT · ∇φ = dxT · grad φ. (2.230)
We can also take the transpose of both sides, recalling that the transpose of a scalar is the
scalar itself, to obtain

T
(dφ)T = dxT · ∇φ , (2.231)
dφ = (∇φ)T · dx, (2.232)
= ∇T φ · dx. (2.233)

Here we expand ∇T as ∇T = (∂1 , ∂2 , ∂3 ). When ∂i or ∇ operates on a scalar, it is known as

the gradient operator. The gradient operator operating on a scalar function gives rise to a
vector function.

© 05 December 2025, Joseph M. Powers. All rights reserved.

2.4. GRAD, DIV, CURL, ETC. 53

We next describe the gradient operator operating on a vector. For vectors in Cartesian
index and Gibbs notation, we have, following a similar analysis12
dvi = dxj ∂j vi = ∂j vi dxj , (2.234)
dvT = dxT · ∇vT , (2.235)
dv = (∇vT )T · dx, (2.236)
= (grad v)T · dx. (2.237)
Here the quantity ∂j vi is the gradient of a vector, that is a tensor. So the gradient operator
operating on a vector raises its order by one. The Gibbs notation with transposes suggests
properly that the gradient of a vector can be expanded as
   
∂1 ∂1 v1 ∂1 v2 ∂1 v3
∇vT =  ∂2  ( v1 v2 v3 ) =  ∂2 v1 ∂2 v2 ∂2 v3  . (2.238)
∂3 ∂3 v1 ∂3 v2 ∂3 v3
Lastly we consider the gradient operator operating on a tensor. For tensors in Cartesian
index notation, we have, following a similar analysis
dTij = dxk ∂k Tij = ∂k Tij dxk . (2.239)
Here the quantity ∂k Tij is a third order tensor. So the gradient operator operating on a
tensor raises its order by one as well. The Gibbs notation is not straightforward as it can
involve something akin to the transpose of a three-dimensional matrix.

2.4.2 Divergence operator

The contraction of the gradient operator on either a vector or a tensor is known as the
divergence, sometimes denoted by “div.” For the divergence of a vector, we have
∂i vi = ∂1 v1 + ∂2 v2 + ∂3 v3 = ∇T · v = div v. (2.240)
The divergence of a vector is a scalar. A vector field that is divergence-free, ∇T · v = 0, is
defined as solenoidal.
For the divergence of a second order tensor, we have
∂i Tij = ∂1 T1j + ∂2 T2j + ∂3 T3j = ∇T · T = div T. (2.241)
The divergence operator operating on a tensor gives rise to a row vector. We will sometimes
have to transpose this row vector in order to arrive at a column vector, e.g. we will have

T
need for the column vector ∇T · T . We note that, as with the vector inner product, most
texts assume the transpose operation is understood and write the divergence of a vector or
tensor simply as ∇ · v or ∇ · T.
12
A more common approach, not using the transpose notation, would be to say here for the Gibbs notation
that dv = dx·∇v. However, this is only works if we consider dv to be a row vector, as dx·∇v is a row vector.
All in all, while at times clumsy, the transpose notation allows a for great deal of clarity and consistency
with matrix algebra.

© 05 December 2025, Joseph M. Powers. All rights reserved.

54 CHAPTER 2. GEOMETRY

2.4.3 Curl operator

The curl operator is the derivative analog to the cross product. We write it in the following
three ways:

ωi = ǫijk ∂j vk , (2.242)
ω = ∇ × v, (2.243)
ω = curl v. (2.244)

Expanding for i = 1, 2, 3 gives

ω1 = ǫ123 ∂2 v3 + ǫ132 ∂3 v2 = ∂2 v3 − ∂3 v2 , (2.245)

ω2 = ǫ231 ∂3 v1 + ǫ213 ∂1 v3 = ∂3 v1 − ∂1 v3 , (2.246)
ω3 = ǫ312 ∂1 v2 + ǫ321 ∂2 v1 = ∂1 v2 − ∂2 v1 . (2.247)

2.4.4 Laplacian operator

The Laplacian13 operator can operate on a scalar, vector, or tensor function. It is a simple
combination of first the gradient followed by the divergence. It yields a function of the same
order as that which it operates on. For its most common operation on a scalar, it is denoted
as follows:
∂i ∂i φ = ∇T · ∇φ = ∇2 φ = div grad φ. (2.248)
In viscous fluid flow, we will have occasion to have the Laplacian operate on vector:

T
T
∂i ∂i vj = ∇T · ∇vT = ∇2 v T = ∇2 v = div grad v. (2.249)

2.4.5 Time derivative

Following Panton (2024), we will employ a useful but unusual notation for the partial deriva-
tive with respect to time:
∂
∂o ≡ , (2.250)
∂t
that will be used extensively later.

2.4.6 Relevant theorems

We will use several theorems that are developed in vector calculus. Here we give the simplest
of motivations, and simply present them. The reader should consult a standard mathematics
text for detailed derivations.
13
Pierre-Simon Laplace, 1749-1827, Normandy-born French astronomer of humble origin. Educated at
Caen, taught in Paris at École Militaire.

© 05 December 2025, Joseph M. Powers. All rights reserved.

2.4. GRAD, DIV, CURL, ETC. 55

[Link] Fundamental theorem of calculus

The fundamental theorem of calculus is as follows
Z x=b Z x=b  
dφ
f (x) dx = dx = φ(b) − φ(a). (2.251)
x=a x=a dx

It effectively says that to find the integral of a function f (x), that is the area under the
curve, it suffices to find a function φ, whose derivative is f , and evaluate φ at each endpoint,
and take the difference to find the area under the curve.

[Link] Gauss’s theorem

Gauss’s14 theorem is the analog of the fundamental theorem of calculus extended to volume
integrals. It applies to tensor functions of arbitrary order and is as follows:
Z Z
∂i (Tijk...) dV = ni Tijk... dA. (2.252)
V A

Here V is an arbitrary volume, dV is the element of volume, A is the surface that bounds
V , ni is the outward unit normal to A, and Tjk.. is an arbitrary tensor function. The surface
integral is analogous to evaluating the function at the end points in the fundamental theorem
of calculus. In Gibbs notation, we have
Z Z
T
∇ · T dV = nT · T dA. (2.253)
V A

If we take Tjk... to be the scalar of unity (whose derivative must be zero), Gauss’s theorem
reduces to Z
ni dA = 0. (2.254)
A

That is the unit normal to the surface integrated over the surface, cancels to zero when the
entire surface is included. We will use Gauss’s theorem extensively. It allows us to convert
sometimes difficult volume integrals into easier interpreted surface integrals. It is often useful
to use this theorem as a means of toggling back and forth from one form to another.

Example 2.16
Demonstrate the validity of Gauss’s theorem for the tensor field
 
x1 x2 x1
T =  x2 x2 x3  , (2.255)
x3 x3 x2
14
Carl Friedrich Gauss, 1777-1855, Brunswick-born German mathematician, considered the founder of
modern mathematics. Worked in astronomy, physics, crystallography, optics, biostatistics, and mechanics.
Studied and taught at Göttingen.

© 05 December 2025, Joseph M. Powers. All rights reserved.

56 CHAPTER 2. GEOMETRY

where the volume under consideration is the unit cube defined on the domain x1 ∈ [0, 1], x2 ∈ [0, 1],
x3 ∈ [0, 1].

We first note that T is asymmetric. We easily see that

 
x1 x2 x1
∇T · T = ( ∂1 ∂2 ∂3 )  x2 x2 x3  , (2.256)
x3 x3 x2
= (3 2 1). (2.257)

Integrating the constant row vector over the unit cube, we find
Z
∇T · T dV = ( 3 2 1 ) . (2.258)
V

Then, we can evaluate the surface integral on each of the six faces and perform a set of six surface
integrals. Leaving out the details, we do so, and find
Z
nT · T dA = ( 3 2 1 ) . (2.259)
A

This verifies Gauss’s theorem for this case. For asymmetric tensors such as our T, we need to be careful
about commuting operators. For example, for this problem
 
Z 2
T · n dA =  2  6= ( 3 2 1 ) . (2.260)
A 0

[Link] Stokes’ theorem

Stokes’15 theorem is as follows.
Z I
ni ǫijk ∂j vk dA = vi αi ds. (2.261)
A C

Once again A is a bounding surface and ni is its outward unit normal. The integral with the
circle through it denotes a closed contour integral with respect to arc length s, and αi is the
unit tangent vector to the bounding curve C. In Gibbs notation, it is written as
Z I
T
n · ∇ × v dA = vT · α ds. (2.262)
A C
15
Sir George Gabriel Stokes, 1819-1903, Irish-born British physicist and mathematician, holder of the Lu-
casian chair of Mathematics at Cambridge University, developed, simultaneously with Navier, the governing
equations of fluid motion, in a form that was more robust than that of Navier.

© 05 December 2025, Joseph M. Powers. All rights reserved.

2.4. GRAD, DIV, CURL, ETC. 57

[Link] A useful identity

It is easy to show that a useful identity involving ∇, v and ω = ∇ × v is as follows:

 
1
vj ∂j vi = ∂i vj vj − ǫijk vj ωk , (2.263)
2
 
T

1 T
v ·∇ v = ∇ v · v − v × ω. (2.264)
2

This is easily proved by considering the right hand side of Eq. (2.263), expanding, and using
Eqs. (2.243) and then (2.75):
 
1
∂i vj vj − ǫijk vj ωk = vj ∂i vj − ǫijk vj ǫklm ∂l vm , (2.265)
2 | {z }
=ωk
= vj ∂i vj − ǫkij ǫklm vj ∂l vm , (2.266)
= vj ∂i vj − (δil δjm − δim δjl ) vj ∂l vm , (2.267)
= vj ∂i vj − vj ∂i vj +vj ∂j vi , (2.268)
| {z }
=0
= vj ∂j vi , QED. (2.269)

[Link] Leibniz’s rule: general transport theorem for arbitrary regions

Leibniz’s16 rule relates time derivatives of integral quantities to a form that distinguishes
changes that are happening within the boundaries to changes due to fluxes through bound-
aries. This is the foundation of the so-called control volume approach. Using the nomen-
clature of Whitaker (1968), p. 92, we also call Leibniz’s rule the general transport theorem.
Leibniz’s rule applied to an arbitrary tensorial function is as follows:
Z Z Z
d
Tjk...(xi , t) dV = ∂o Tjk... dV + nl wl Tjk... dA. (2.270)
dt Va (t) Va (t) Aa (t)

• Va (t) → arbitrary region that is a time-dependent volume,

• Aa (t) → bounding surface of the arbitrary moving volume,

• wl → velocity vector of points on the moving surface,

• nl → unit normal to moving surface.

16
Gottfried Wilhelm von Leibniz, 1646-1716, Leipzig-born German philosopher and mathematician. In-
vented calculus independent of Newton and employed a superior notation to that of Newton.

© 05 December 2025, Joseph M. Powers. All rights reserved.

58 CHAPTER 2. GEOMETRY

Say we have the special case in which Tjk... = 1; then Leibniz’s rule reduces to
Z Z Z
d
dV = ∂o (1) dV + nk wk (1) dA, (2.271)
dt Va (t) Va (t) Aa (t)
Z
dVa
= nk wk dA. (2.272)
dt Aa (t)

This simply says the total volume of the region, that we call Va , changes in response to net
motion of the bounding surface.

[Link].1 Material region: Reynolds transport theorem In the special case where
the volume contains the same fluid particles, the velocity of the boundary is the fluid particle
velocity, wl = vl , and our general transport theorem becomes, again using the nomenclature
of Whitaker (1968), p. 92, the Reynolds17 transport theorem:
Z Z Z
d
Tjk...(xi , t) dV = ∂o Tjk... dV + nl vl Tjk... dA. (2.273)
dt Vm (t) Vm (t) Am (t)

The term Vm (t) and Am (t) denote the time-dependent material region and material surface
to denote that the geometry in question always contains the same material particles.

[Link].2 Fixed region In the special case where the volume is fixed in time, the velocity
of the boundary is zero, wl = 0, and our general transport theorem becomes
Z Z
d
Tjk...(xi , t) dV = ∂o Tjk... dV. (2.274)
dt V V

In this case there is no time-dependency of the fixed volume V .

[Link].3 Scalar function In the special case where Tjk... is a scalar function f , Leibniz’s
rule reduces to
Z Z Z
d
f (xi , t) dV = ∂o f (xi , t) dV + nl wl f (xi , t) dA. (2.275)
dt Va (t) Va (t) Aa (t)

Further, considering one-dimensional cases only, we can then say

Z Z x=b(t)
d x=b(t) db da
f (x, t) dx = ∂o f dx + f (b(t), t) − f (a(t), t). (2.276)
dt x=a(t) x=a(t) dt dt
As in the fundamental theorem of calculus, for the one-dimensional case, we do not have to
evaluate a surface integral; instead, we simply must consider the function at its endpoints.
Here db/dt and da/dt are the velocities of the bounding surface and analogous to wk . The
terms f (b(t), t) and f (a(t), t) are equivalent to evaluating Tjk.. on Aa (t).
17
Osborne Reynolds, 1842-1912, Belfast-born British engineer and physicist, educated in mathematics at
Cambridge, first professor of engineering at Owens College, Manchester, did fundamental experimental work
in fluid mechanics and heat transfer.

© 05 December 2025, Joseph M. Powers. All rights reserved.

2.5. GENERAL COORDINATE TRANSFORMATIONS 59

2.5 General coordinate transformations

Here we introduce, following Aris (1962), Powers and Sen (2015), Ch. 1.6, and many others,
some standard notation from tensor analysis for general coordinate transformations. This
extends our analysis of Sec. 2.1.2, that was confined to simple rotation transformations. In
this notation, both sub- and superscripts are needed to distinguish between what are known
as covariant and contravariant vectors, that are really different mathematical representations
of the same quantity, just cast onto different basis vectors. The basis vectors may or may not
be orthonormal. They may not even be orthogonal. All they must be is linearly independent.
In brief, we start with a general transformation from the non-Cartesian coordinate, defined
here as xi , to the Cartesian coordinate, defined here as ξ k :

ξ 1 = ξ 1 (x1 , x2 , x3 ), (2.277)
ξ 2 = ξ 2 (x1 , x2 , x3 ), (2.278)
ξ 3 = ξ 3 (x1 , x2 , x3 ). (2.279)

As an example, this form includes the transformation from a non-Cartesian cylindrical co-
ordinate system to a Cartesian system; this will be taken up in detail in Ch. 7.1. We could
also say

ξ k = ξ k (xi ), ξ = ξ(x). (2.280)

Local linearization of the transformation gives

 1  ∂ξ1 ∂ξ1 ∂ξ 1  
dξ ∂x1 ∂x2 ∂x3 dx1
 dξ 2  =  ∂ξ21 ∂ξ22 ∂ξ 2   dx2  , (2.281)
∂x ∂x ∂x3
dξ 3 ∂ξ 3
1
∂ξ 3 ∂ξ 3 dx3
| ∂x {z2
∂x ∂x3 }
J
dξ = J · dx. (2.282)

Here the local Jacobian matrix J is defined as

 ∂ξ1 ∂ξ1 ∂ξ1 
∂x1 ∂x2 ∂x3 k
2
J =  ∂ξ 1 ∂ξ 2 ∂ξ 2  = ∂ξ . (2.283)
∂x ∂x2 ∂x3 ∂xi
∂ξ 3 ∂ξ 3 ∂ξ 3
∂x1 ∂x2 ∂x3

The chain rule for partial differentiation can be used to represent the gradient as
 ∂   ∂ξ1 ∂ξ2 ∂ξ3   ∂ 
∂x1 ∂x1 ∂x1 ∂x1 ∂ξ 1
 ∂  =  ∂ξ12 ∂ξ 2 ∂ξ 3  ∂ , (2.284)
∂x2 ∂x ∂x2 ∂x2 ∂ξ 2
∂ ∂ξ 1 ∂ξ 2 ∂ξ 3 ∂
∂x3 ∂ξ 3
| {z } | ∂x3 {z3
∂x ∂x3 } | {z }
∇x JT ∇ξ

∇ x = J T · ∇ξ . (2.285)

© 05 December 2025, Joseph M. Powers. All rights reserved.

60 CHAPTER 2. GEOMETRY

Inverting, we find

−1
∇ξ = J T · ∇x . (2.286)

This can be directly compared with Eq. (2.282). In the special case for which the trans-

−1
−1
formation is a rotation, we have J = Q and thus JT = QT = Q. In this case, we
recover the simpler dξ = Q · dx and ∇ξ = Q · ∇x .
For Cartesian systems, we must have the classical formula for differential distance ds:

2
2
2
ds2 = dξ 1 + dξ 2 + dξ 3 = dξ T · dξ. (2.287)

The differential distance must be an invariant in either coordinate representation, so we must

have

ds2 = dξ T · dξ = (J · dx)T · (J · dx) = dxT · |JT{z· J} ·x = dxT · G · dx. (2.288)

G

Here we have defined the metric tensor

G = JT · J, (2.289)
∂ξ k ∂ξ k
gij = . (2.290)
∂xi ∂xj
One can also show that
1 imn jpq
g ij = ǫ ǫ gmp gnq , (2.291)
2
gik g kj = δij . (2.292)

We also have the Jacobian determinant J defined as

√ ∂ξ k
J= g = det i = det J. (2.293)
∂x
A vector’s contravariant representation is given by v i . Its covariant representation is given
by vi . The metric tensor links one representation to the other via

vj = v i gij , v i = g ij vj . (2.294)

In the remaining paragraphs of this chapter, we present some slightly modified text first
presented by Powers and Sen (2015) in their Ch. 1.6.5 to better understand the nature of
vectors in terms of linear combinations of covariant and contravariant basis vectors. The
only requirement we place on the basis vectors is linear independence: they must point
in different directions. They need not be unit vectors, and their lengths may differ from
one another. Consider the non-orthogonal basis vectors e1 , e2 , aligned with the x1 and x2
directions shown in Fig. 2.6a. The non-orthogonal basis vectors need not be unit vectors;

© 05 December 2025, Joseph M. Powers. All rights reserved.

2.5. GENERAL COORDINATE TRANSFORMATIONS 61

a) b)

Figure 2.6: Example of a vector v represented in a non-orthogonal coordinate system along

with its a) basis vectors, and b) dual basis vectors. (Figure adapted from Fig. 1.16 of Powers
and Sen (2015)).

hence, ||e1 || =
6 1, ||e2 || =
6 1 is permitted. However, they must be linearly independent; hence
e1 × e2 6= 0. The vector v can then be written as

v = v 1 e1 + v 2 e2 (2.295)

Here v 1 and v 2 are the contravariant components of v. And the vectors e1 and e2 are the
contravariant basis vectors, even though they are subscripted. The entity v is best thought of
as either an entity unto itself or perhaps as a column vector whose components are Cartesian.
In matrix form, we can think of v as

 .   .   . ..   
.. .. .. .
1  2    v1
v = v  e1  + v  e2  =  e1 e2  . (2.296)
.. .. .. .. v2
. . . .
| {z }
J

The matrix of basis vectors really acts as a local Jacobian matrix, J, that relates the Cartesian
and non-orthogonal representations of v.
Vectors that comprise a dual or reciprocal basis have two characteristics: they are or-
thogonal to the all the original basis vectors with different indices, and the dot product of
each dual vector with respect to the original vector of the same index must be unity. The
covariant basis vectors e1 , e2 are dual to e1 , e2 , as shown in Fig. 2.6b. Specifically, we have

© 05 December 2025, Joseph M. Powers. All rights reserved.

62 CHAPTER 2. GEOMETRY

T T T T
e1 · e2 = 0, e2 · e1 = 0, e1 · e1 = 1, and e2 · e2 = 1. In matrix form, this is
 . ..  
 T  .
. . 
· · · e1 ···   1 0
T  e1 e2  = = I. (2.297)
· · · e2 ··· . . 0 1
| {z } .. ..
J−1 | {z }
J

Obviously, the matrix of dual vectors can be formed by inverting the matrix of the original
basis vectors. We can also represent v as

v = v1 e1 + v2 e2 . (2.298)

In matrix form, we can think of v as

 .   .   . ..   
.. .. .. . v1
v = v1  e1  + v2  e2  =  e1 e2  . (2.299)
.. .. .. .. v2
. . . .
| {z }
(J−1 )T

We might also say

 T 
T · · · e1 ···
v = ( v1 v2 ) T . (2.300)
· · · e2 ···
| {z }
J−1

Because the magnitude of v is independent of its coordinate system, we can say

 . ..   
 T  .
. .
· · · e1 ···   v1
vT · v = ( v1 v2 ) T e
 1 e 2 , (2.301)
· · · e2 ··· . . v2
| {z } .. ..
J−1 | {z }
J
| {z }
I

and thus
 
T v1
v · v = ( v1 v2 ) = vi v i . (2.302)
v2

Now we can also transpose Eq. (2.296) to obtain

 
T 1 2 · · · eT1 · · ·
v = (v v ) . (2.303)
· · · eT2 · · ·
| {z }
JT

© 05 December 2025, Joseph M. Powers. All rights reserved.

2.5. GENERAL COORDINATE TRANSFORMATIONS 63

Now combining this with Eq. (2.296) to form vT · v, we also see

 . ..   
  .. .
· · · eT1 ···   v1
vT · v = ( v 1 v2 )  e1 e2  . (2.304)
· · · eT2 ··· . . v2
| {z } .
. .
.
JT | {z }
J

Now with the metric tensor gij = G = JT · J, we have

 
T 1 2 v1
v · v = (v v )·G· . (2.305)
v2

One can now compare Eq. (2.305) with Eq. (2.302) to infer the covariant components v1 and
v2 . For the same vector v, the covariant components are different than the contravariant
components. Thus, for example,

vi = ( v 1 v 2 ) · G = v j gij = gij v j . (2.306)

Deducing from Eq. (2.292) that g ij = G−1 , we also see

g ij vi = v j . (2.307)

In Cartesian coordinates, a basis and its dual are the same, and so also are the contravari-
ant and covariant components of a vector. For this reason Cartesian vectors and tensors are
usually written with only subscripts.

Example 2.17
Consider the vector v whose Cartesian representation is
 
4
v= . (2.308)
2

Consider also the set of non-orthogonal basis vectors

   
2 1
e1 = , e2 = . (2.309)
0 1

Find the contravariant and covariant components, v i and vi , of v.

Here the Jacobian matrix is

 . .. 
.. .  
  2 1
J =  e1 e2  = . (2.310)
.. .. 0 1
. .

© 05 December 2025, Joseph M. Powers. All rights reserved.

64 CHAPTER 2. GEOMETRY

The contravariant components v i are given by solving

 1
v
v = J· , (2.311)
v2
    1 
4 2 1 v
= . (2.312)
2 0 1 v2

Inverting, we find
   
v1 1
= . (2.313)
v2 2

Thus, we have

v = 1e1 + 2e2 , (2.314)

     
4 2 1
= 1 +2 . (2.315)
2 0 1

The covariant basis vectors are given by

 T  1 
−1 · · · e1 ··· 2 − 12
J = T = . (2.316)
· · · e2 ··· 0 1

Thus, we have
 1
  
0
e1 = 2 , e2 = . (2.317)
− 21 1

The metric tensor is given by

    
T 2 0 2 1 4 2
gij = G = J · J = = . (2.318)
1 1 0 1 2 2

We can get the covariant components in many ways. Let us choose

    
j 4 2 1 8
vi = gij v = = . (2.319)
2 2 2 6

Thus, we have the covariant representation of v as

v = 8e1 + 6e2 , (2.320)

   1   
4 2 0
= 8 +6 . (2.321)
2 − 12 1

In Cartesian coordinates we have

 
4
vT · v = ( 4 2) = 20. (2.322)
2

This is invariant under coordinate transformation as in our non-orthogonal coordinate system, we have
 
i 1
vi v = ( 8 6 ) = 20. (2.323)
2

The vectors represented in Fig. 2.6 are proportional to those of this problem.

© 05 December 2025, Joseph M. Powers. All rights reserved.

2.5. GENERAL COORDINATE TRANSFORMATIONS 65

The Christoffel18 symbols are given by

 
m 1 mk ∂gjk ∂gki ∂gij
Γij = g + − k . (2.324)
2 ∂xi ∂xj ∂x

As discussed in detail by Powers and Sen (2015), such notions as covariant derivatives of
contravariant vectors can be defined and can be shown to take the form
∂v i
∇j v i = + Γijl v l . (2.325)
∂xj
Here the notation ∇j is that for a generalized covariant derivative. It can be shown that
rotational transformations from Cartesian systems have zero Christoffel symbols. However,
more general transformations, such as Cartesian to cylindrical coordinates, have non-trivial
Christoffel symbols. Their physical manifestation are terms such as centripetal and Coriolis19
accelerations, as will be demonstrated in detail in Ch. 7.1.1. These terms are not based on
the derivative of the vector, but are related to the vector itself.

18
Elwin Bruno Christoffel, 1829-1900, German mathematician and physicist.
19
Gaspard Gustave de Coriolis, 1792-1843, Paris-born mathematician, taught with Navier, introduced the
terms “work” and “kinetic energy” with modern scientific meaning, wrote on the mathematical theory of
billiards.

© 05 December 2025, Joseph M. Powers. All rights reserved.

66 CHAPTER 2. GEOMETRY

© 05 December 2025, Joseph M. Powers. All rights reserved.

Chapter 3

Kinematics

see Powers, Chapter 3,

see Panton, Chapter 4,
see Whitaker, Chapter 3,
see Aris, Chapter 4.

The previous chapter was in many ways a discussion of geometry or place. Here we will
consider kinematics, the study of motion in space. Here we will pay no regard to the forces
that cause the motion. If we knew the position of every fluid particle as a function of time,
then we could in principle also describe the velocity and acceleration of each particle. We
could also make statements about how groups of particles translate, rotate, and deform.
This is the essence of kinematics. Fluid motion is generally a highly nonlinear phenomenon.
In this chapter, we will develop tools, using a local linear analysis, to break down the most
complex fluid flows to a summation of fundamental motions.

3.1 Lagrangian description

A Lagrangian1 description is similar to a classical description of motion in that each fluid
particle is effectively labeled and tracked in terms of its initial position xoj and time t̂. We
take the position vector of a particle ri to be

ri = r̃i (xoj , t̂). (3.1)

The velocity vi of a particular particle is the time derivative of its position, holding xoj fixed:

∂r̃i
vi = . (3.2)
∂ t̂ xoj

1
Joseph-Louis Lagrange (originally Giuseppe Luigi Lagrangia), 1736-1813, Italian born, Italian-French
mathematician. Worked on celestial mechanics and the three body problem. Worked in Berlin and Paris.
Part of the committee that formulated the metric system.

67
68 CHAPTER 3. KINEMATICS

The acceleration ai of a particular particle is the second time derivative of its position,
holding xoj fixed:
∂ 2 r̃i
ai = . (3.3)
∂ t̂2 xoj
We can also write other variables as functions of time and initial position, for example, we
could have for pressure p(xoj , t̂).
The Lagrangian description has important pedagogical value, but is only occasionally
used in practice, except maybe where it can be useful to illustrate a particular point. In
solid mechanics, it is often critically important to know the location of each solid element,
and it is the method of choice.

3.2 Eulerian description

It is more common in fluid mechanics to use the Eulerian description of fluid motion. In
this description, all variables are taken to be functions of time and local position, rather
than initial position. Here, we will take the local position to be given by the position vector
xi = ri . A general transformation from one coordinate system (xi , t) to another (xoi , t̂) can
take the general form

t = t(t̂, xoj ), (3.4)

xi = xi (t̂, xoj ). (3.5)

At this point, we can extend and adapt the analysis introduced in Ch. 2.5. While that
discussion was focused on spatial coordinate transformations, there is no reason it cannot
be extended to so-called space-time systems such as we have here. The chain rule tells us
 
  ∂t ∂t  
dt ∂ t̂ xoj ∂xjo
d t̂
= t̂  . (3.6)
dxi ∂xi ∂xi
o
dxoj
∂ t̂ xoj ∂x
| {z j t̂ }
J

This has the Jacobian matrix J of

 
∂t ∂t
∂ t̂ xoj ∂xoj
J= t̂ . (3.7)
∂xi ∂xi
∂ t̂ xoj ∂xoj
t̂

We will consider the transformation from Lagrangian coordinates to Eulerian coordinates

is given by the more specific form

t = t̂, (3.8)
xi = r̃i (t̂, xoj ). (3.9)

© 05 December 2025, Joseph M. Powers. All rights reserved.

3.3. MATERIAL DERIVATIVES 69

This has the more specific Jacobian matrix

 
1 0
J = ∂ r̃i ∂ r̃i . (3.10)
∂ t̂ xoj ∂xoj
t̂
We have
∂r̃i
J = det J = det . (3.11)
∂xoj t̂

As long as J > 0, the transformation is non-singular and orientation-preserving. At t̂ = 0,

we have xi = r̃i = xoi , Thus at t̂ = 0, we get J = det δij = 1. So at the initial state, the
transformation is volume- and orientation-preserving.

3.3 Material derivatives

The material derivative is the derivative following a fluid particle. It is also known as the
substantial derivative or the total derivative. It is trivial in Lagrangian coordinates, because
by definition, a Lagrangian description tracks a fluid particle. It is not as straightforward
in the Eulerian viewpoint. Let us first solve an example using a simplistic, but effective
approach, to determine the material derivative. We will consider a scaler fluid property such
as temperature T , that is a function of position and time.

Example 3.1
Find the material derivative of T (t, x1 , x2 , x3 ) using a simplistic approach based on the definition
of the derivative.

If T = T (t, x1 , x2 , x3 ), then by definition of the total derivative, we have

∂T ∂T ∂T ∂T
dT = dt + dx1 + dx2 + dx3 . (3.12)
∂t ∂x1 ∂x2 ∂x3
Let us scale by dt and then say
dT ∂T ∂T dx1 ∂T dx2 ∂T dx3
= + + + . (3.13)
dt ∂t ∂x1 dt ∂x2 dt ∂x3 dt
On fluid particle path, we have dx1 /dt = v1 , dx2 /dt = v2 , dx3 /dt = v3 , so we get
dT ∂T ∂T ∂T ∂T
= + v1 + v2 + v3 , (3.14)
dt ∂t ∂x1 ∂x2 ∂x3
∂T ∂T ∂T ∂T
= + v1 + v2 + v3 , (3.15)
∂t ∂x1 ∂x2 ∂x3
= ∂o T + vi ∂i T, (3.16)
∂T
= + vT · ∇T, (3.17)
∂t 
∂
= + vT · ∇ T. (3.18)
∂t
| {z }
d/dt

© 05 December 2025, Joseph M. Powers. All rights reserved.

70 CHAPTER 3. KINEMATICS

Here d/dt is the material derivative.

Let us find the material derivative in a more formal fashion. We can characterize the
position and time in either an Eulerian or Lagrangian fashion. Let the Lagrangian repre-
sentation be T = TL (xoi , t̂) and the Eulerian representation be T = TE (xi , t). Now both
formulations must give the same result at the same time and position; applying our trans-
formation between the two systems thus yields

T = TL (xoi , t̂) = TE (xi = r̃i (xoj , t̂), t = t̂). (3.19)

Now from basic calculus we have

∂r̃i ∂r̃i
dxi = dt̂ + o
dxoj . (3.20)
∂ t̂ xoj ∂xj t̂

From basic calculus, we also have

∂TL ∂TL
dTL = dt̂ + dxo , (3.21)
∂ t̂ xoj ∂xoj t̂ j
∂TE ∂TE
dTE = dt + dxi . (3.22)
∂t xi ∂xi t

Now, we must have dT = dTL = dTE for the same fluid particle, so making appropriate
substitutions, we get
!
∂TL ∂TL ∂T E ∂TE ∂r̃ i ∂r̃ i
dt̂ + o
dxoj = dt + dt̂ + o
dxoj . (3.23)
∂ t̂ xoj ∂xj t̂ ∂t xi ∂x i t ∂ t̂ xoj ∂x j t̂
| {z } | {z }
dTL dxi
| {z }
dTE

For the variation of T of a particular particle, we hold xoj fixed, so that dxoj = 0. Using also
the fact that t̂ = t, so dt̂ = dt, and dividing by dt̂, we get

∂TL ∂TE ∂TE ∂r̃i

= + , (3.24)
∂ t̂ xoj ∂t xi ∂xi t ∂ t̂ xoj

and using the definition of fluid particle velocity, Eq. (3.2), we get

∂TL ∂TE ∂TE

= + vi . (3.25)
∂ t̂ xoj ∂t xi ∂xi t

© 05 December 2025, Joseph M. Powers. All rights reserved.

3.3. MATERIAL DERIVATIVES 71

Removing the operands T , TL , and TE , and recognizing that holding xoj fixed is the same as
holding xoi fixed, we can write the derivative following a particle in the following manner as
an operator

∂ ∂ ∂ ∂ ∂ D d
= + vi = + vT · ∇ = + vT · grad ≡ ≡ . (3.26)
∂ t̂ xoi ∂t xi ∂xi t ∂t x ∂t x Dt dt

We will generally use the following shorthand for d/dt, the derivative following a particle:

d
= ∂o + vi ∂i . (3.27)
dt
Here we have invoked our shorthand for the spatial gradient operator, ∂i , Eq. (2.226, and
for the partial derivative with respect to time, ∂o , Eq. (2.250). The notation D/Dt is more
common for the derivative following a fluid particle, but the advantage of this extra notation
is not clear. We will instead use d/dt.
We can achieve the same result in a different fashion involving the chain rule. The chain
rule gives
∂T ! ∂t ∂xi ! ∂T !
∂ t̂ xoj ∂ t̂ xoj ∂ t̂ xoj ∂t xi
∂T
= ∂t ∂xi ∂T . (3.28)
∂xoj ∂xoj ∂xoj ∂xi
t̂ t̂ t̂ t
| {z }
=JT

We recognize JT as the transpose of the Jacobian matrix of the transformation. Invoking

our transformation, Eqs. (3.9,3.8), we get
∂T ! ∂ r̃i ! ∂T ! ! ∂T !
∂ t̂ xoj
1 ∂ t̂ xoj ∂t xi 1 vi ∂t xi
= = ∂ r̃i . (3.29)
∂T
0 ∂ r̃i ∂T 0 ∂xoj
∂T
∂xoj ∂xoj ∂xi t̂ ∂xi
t̂ t̂ t t

Thus, we get the equivalent, after again recognizing that holding xoj fixed is the same as
holding xoi fixed:

∂T ∂T ∂T
= + vi . (3.30)
∂ t̂ xoi ∂t xi ∂xi t

We also find from inverting Eq. (3.28), that one gets the relationship
∂ ! ∂ !
∂t xi

T −1
∂ t̂ xoj
∂ = J ∂
. (3.31)
∂xi ∂xoj
t t̂

This extends the earlier-developed Eq. (2.286).

© 05 December 2025, Joseph M. Powers. All rights reserved.

72 CHAPTER 3. KINEMATICS

Example 3.2
The one-dimensional unsteady motion of a set of fluid particles is given by
 2 !
o o t̂
r̃(x , t̂) = x 1 + , (3.32)
τ

where τ is a constant. Heat is transferred to the particles in such a way that its temperature evolution
is governed by the equation
 o 
∂T To xo o x xo
= , T (0, x ) = To 1− . (3.33)
∂ t̂ xo Lτ L L
Here To and L are constant reference temperature and length, respectively. Analyze the fluid motion
and temperature evolution in an Eulerian frame.

First note that when t̂ = 0 that r̃ = xo , as required. The transformation to Eulerian coordinates is
given by
 2 !
t̂
x = xo 1 + , (3.34)
τ
t = t̂. (3.35)

The velocity and acceleration of a fluid particle are

∂r̃ t̂
v = = 2xo 2 , (3.36)
∂ t̂ xo τ
∂ 2 r̃ 2xo
a = = 2 . (3.37)
∂ t̂2 xo τ
We can integrate the temperature evolution equation to get

To xo t̂
T (t̂, xo ) = + f (xo ). (3.38)
Lτ
Here f (xo ) is an arbitrary function of xo , that can be evaluated with the initial condition so that
 o 
xo t̂ x xo
T (t̂, xo ) = To + To 1− , (3.39)
L τ L L
 
xo t̂ xo
= To +1− . (3.40)
L τ L
Because of the simple nature of the flow field, the transformation to Eulerian coordinates is easy and
seen by inspection to give
!
x/L t x/L
T (t, x) = To
2 +1−
2 . (3.41)
1+ t τ 1+ t
τ τ

For τ = 1, L = 1, To = 300, plots of T (x, t) and T (xo , t̂) are shown in Fig. 3.1.

© 05 December 2025, Joseph M. Powers. All rights reserved.

3.4. STREAMLINES 73

T
T

xo
x

Figure 3.1: Eulerian plot T (x, t) and Lagrangian plot of T (xo , t̂) for example problem
originally described in Lagrangian coordinates, with τ = 1, L = 1, To = 300.

3.4 Streamlines
Streamlines are curves that are everywhere instantaneously parallel to velocity vectors. If a
differential vector dxk is parallel to a velocity vector vj , then the cross product of the two
vectors must be zero; hence, for a streamline we must have

ǫijk vj dxk = 0. (3.42)

In Gibbs notation, we would say
v × dx = 0. (3.43)
Recalling that the cross product can be interpreted as a determinant, we get this condition
to reduce to
e1 e2 e3
v1 v2 v3 = 0. (3.44)
dx1 dx2 dx3
Expanding the determinant gives

e1 (v2 dx3 − v3 dx2 ) + e2 (v3 dx1 − v1 dx3 ) + e3 (v1 dx2 − v2 dx1 ) = 0. (3.45)

Because the basis vectors e1 , e2 , and e3 are linearly independent, the coefficient on each
must be zero, giving rise to
dx3 dx2
v2 dx3 = v3 dx2 , ⇒ = , (3.46)
v3 v2
dx1 dx3
v3 dx1 = v1 dx3 , ⇒ = , (3.47)
v1 v3
dx2 dx1
v1 dx2 = v2 dx1 , ⇒ = . (3.48)
v2 v1
Combining, we get
dx1 dx2 dx3
= = . (3.49)
v1 v2 v3

© 05 December 2025, Joseph M. Powers. All rights reserved.

74 CHAPTER 3. KINEMATICS

At a fixed instant in time, t = to , we set the terms in Eq. (3.49) all equal to an arbitrary
differential parameter dτ to obtain
dx1 dx2 dx3
= = = dτ. (3.50)
v1 (x1 , x2 , x3 ; t = to ) v2 (x1 , x2 , x3 ; t = to ) v3 (x1 , x2 , x3 ; t = to )
Here τ should not be thought of as time, but just as a dummy parameter. Streamlines are
only defined at a fixed time. While they will generally look different at different times, in
the process of actually integrating to obtain them, time does not enter into the calculation.
We then divide each equation by dτ and find they are equivalent to a system of differential
equations of the autonomous form
dx1
= v1 (x1 , x2 , x3 ; t = to ), x1 (τ = 0) = x1o , (3.51)
dτ
dx2
= v2 (x1 , x2 , x3 ; t = to ), x2 (τ = 0) = x2o , (3.52)
dτ
dx3
= v3 (x1 , x2 , x3 ; t = to ), x3 (τ = 0) = x3o . (3.53)
dτ
After integration, that in general must be done numerically, we find
x1 (τ ; to , x1o ), (3.54)
x2 (τ ; to , x2o ), (3.55)
x3 (τ ; to , x3o ), (3.56)
where we let the parameter τ vary over whatever domain we choose.

3.5 Pathlines
The pathlines are the locus of points traversed by a particular fluid particle. For an Eulerian
description of motion where the velocity field is known as a function of space and time
vj (xi , t), we can get the pathlines by integrating the following set of three non-autonomous
ordinary differential equations, with the associated initial conditions:
dx1
= v1 (x1 , x2 , x3 , t), x1 (t = to ) = x1o , (3.57)
dt
dx2
= v2 (x1 , x2 , x3 , t), x2 (t = to ) = x2o , (3.58)
dt
dx3
= v3 (x1 , x2 , x3 , t), x3 (t = to ) = x3o . (3.59)
dt
In general these are non-linear equations, and often require full numerical solution, that gives
us
x1 (t; x1o ), (3.60)
x2 (t; x2o ), (3.61)
x3 (t; x3o ). (3.62)

© 05 December 2025, Joseph M. Powers. All rights reserved.

3.6. STREAKLINES 75

3.6 Streaklines
A streakline is the locus of points that have passed through a particular point at some past
time t = t̂. Streaklines can be found by integrating a similar set of equations to those for
pathlines.
dx1
= v1 (x1 , x2 , x3 , t), x1 (t = t̂) = x1o , (3.63)
dt
dx2
= v2 (x1 , x2 , x3 , t), x2 (t = t̂) = x2o , (3.64)
dt
dx3
= v3 (x1 , x2 , x3 , t), x3 (t = t̂) = x3o . (3.65)
dt
After integration, that is generally done numerically, we get
x1 (t; x1o , t̂), (3.66)
x2 (t; x2o , t̂), (3.67)
x3 (t; x3o , t̂). (3.68)
Then, if we fix time t and the particular point in which we are interested (x1o , x2o , x3o )T , we
get a parametric representation of a streakline
x1 (t̂), (3.69)
x2 (t̂), (3.70)
x3 (t̂). (3.71)

Example 3.3
If v1 = 2x1 + t, v2 = x2 − 2t, find a) the streamline through the point (1, 1)T at t = 1, b) the
pathline for the fluid particle that is at the point (1, 1)T at t = 1, and c) the streakline through the
point (1, 1)T at t = 1.

a) streamline

For the streamline, we have the following set of differential equations,

dx1
= 2x1 + t|t=1 , x1 (τ = 0) = 1, (3.72)
dτ
dx2
= x2 − 2t|t=1 , x2 (τ = 0) = 1. (3.73)
dτ
Here it is inconsequential where the parameter τ has its origin, as long as some value of τ corresponds
to a streamline through (1, 1)T , so we have taken the origin for τ = 0 to be the point (1, 1)T . These
equations at t = 1 are
dx1
= 2x1 + 1, x1 (τ = 0) = 1, (3.74)
dτ
dx2
= x2 − 2, x2 (τ = 0) = 1. (3.75)
dτ

© 05 December 2025, Joseph M. Powers. All rights reserved.

76 CHAPTER 3. KINEMATICS

Solving, we get
3 2τ 1
x1 = e − , (3.76)
2 2
x2 = −eτ + 2. (3.77)

Solving for τ , we find

  
1 2 1
τ= ln x1 + . (3.78)
2 3 2

So, eliminating τ and writing x2 (x1 ), we get the streamline to be

s  
2 1
x2 = 2 − x1 + . (3.79)
3 2

b) pathline

For the pathline we have the following equations

dx1
= 2x1 + t, x1 (t = 1) = 1, (3.80)
dt
dx2
= x2 − 2t, x2 (t = 1) = 1. (3.81)
dt
These have solution
7 2(t−1) t 1
x1 = e − − , (3.82)
4 2 4
x2 = −3et−1 + 2t + 2. (3.83)

It is algebraically difficult to eliminate t so as to write x2 (x1 ) explicitly. However, the analysis certainly
gives a parametric representation of the pathline, that can be plotted in x1 , x2 space.

c) streakline

For the streakline we have the following equations

dx1
= 2x1 + t, x1 (t = t̂) = 1, (3.84)
dt
dx2
= x2 − 2t, x2 (t = t̂) = 1. (3.85)
dt
These have solution
5 + 2t̂ 2(t−t̂) t 1
x1 = e − − , (3.86)
4 2 4
x2 = −(1 + 2t̂)et−t̂ + 2t + 2. (3.87)

We evaluate the streakline at t = 1 and get

5 + 2t̂ 2(1−t̂) 3
x1 = e − , (3.88)
4 4
x2 = −(1 + 2t̂)e1−t̂ + 4. (3.89)

© 05 December 2025, Joseph M. Powers. All rights reserved.

3.7. KINEMATIC DECOMPOSITION OF MOTION 77

1.5

1.0
streakline

0.5

pathline
streamline

2 4 6 8

Figure 3.2: Streamline, pathlines, and streaklines for unsteady flow of example problem.

Once again, it is algebraically difficult to eliminate t̂ so as to write x2 (x1 ) explicitly. However, the
analysis gives a parametric representation of the streakline, that can be plotted in x1 , x2 space.
A plot of the streamline, pathline, and streakline for this problem is shown in Fig. 3.2. At the point
(1, 1)T , all three intersect with the same slope. This can also be deduced from the equations governing
streamlines, pathlines, and streaklines.

3.7 Kinematic decomposition of motion

In general the motion of a fluid is non-linear in nearly all respects. Certainly, it is common
for particle pathlines to have a non-linear path; however, this is not actually a hallmark of
non-linearity in that linear theories of fluid motion routinely predict pathlines with finite
curvature. More to the point, we cannot in general use the method of superposition to add
one flow to another to generate a third. One fundamental source of non-linearity is the
non-linear operator vi ∂i , that we will see appears in most of our governing equations.
However, the local behavior of fluids is nearly always dominated by linear effects. By
analyzing only the linear effects induced by small changes in velocity, that we will associate
with the velocity gradient, we will learn about the richness of fluid motion. In the linear
analysis, we will see that a fluid particle’s motion can be described as a summation of

• linear translation,

• rotation as a solid body, and

• straining:

© 05 December 2025, Joseph M. Powers. All rights reserved.

78 CHAPTER 3. KINEMATICS

Figure 3.3: Sketch of fluid particle P in motion with velocity vi and nearby neighbor particle
P ′ with velocity vi + dvi .

– extensional, and
– shear.
Both types of straining can be thought of as deformation rates. We use the word “straining”
in contrast to “strain” to distinguish fluid and flexible solid behavior. Generally it is the
rate of change of strain (that is the “straining”) that has most relevance for a fluid, while it
is the actual strain that has the most relevance for a flexible solid. This is because the stress
in a flexible solid responds to strain, while the stress in a fluid responds to a strain rate.
Nevertheless, while strain itself is associated with equilibrium configurations of a flexible
solid, when its motion is decomposed, strain rate is relevant. In contrast, a rigid solid can
be described by only a sum of linear translation and rotation. A point mass only translates;
it cannot rotate or strain.

fluid motion = translation + rotation + extensional straining + shear straining,

| {z }
straining
flexible solid motion = translation + rotation + extensional straining + shear straining,
| {z }
straining
rigid solid motion = translation + rotation,
point mass motion = translation.

Let us consider in detail the configuration shown in Fig. 3.3. Here we have a fluid particle
at point P with coordinates xi and velocity vi . A small distance dri = dxi away is the fluid
particle at point P ′ , with coordinates xi + dxi . This particle moves with velocity vi + dvi .
We can describe the difference in location by the product of a unit tangent vector αi and a
scalar differential distance magnitude ds:

dri = dxi = αi ds. (3.90)

Note that αi is in general not aligned with the velocity vector, and the differential distance
ds is not associated with the arc length along a particle path. Later in Sec. 3.13, we will

© 05 December 2025, Joseph M. Powers. All rights reserved.

3.7. KINEMATIC DECOMPOSITION OF MOTION 79

select an alignment with the particle path, and thus choose αi = αti and ds = ds, where αti
is the unit tangent to the particle path and ds is the arc length.

3.7.1 Translation
We have the motion at P ′ to be vi + dvi . Obviously, the first term vi represents translation.

3.7.2 Solid body rotation and straining

What remains is dvi , and we shall see that it is appropriate to characterize this term as a solid
body rotation combined with straining. We have from the definition of a total derivative
that

dvj = dxi ∂i vj , (3.91)

dvT = dxT · ∇vT , (3.92)

T
dv = ∇vT · dx, (3.93)
T
= L · dx. (3.94)

Here

∂i vj = ∇vT ≡ L, (3.95)

is the velocity gradient tensor. Let us expand this for clarity. Our convention allows ordering
consistent with Lij = ∂i vj . The literature is not uniform in this convection, and many
sources use LT to denote what we define as L. It is easy to reconcile one with the other by
exchanging L for LT . We can expand Eq. (3.92) as
 
∂1 v1 ∂1 v2 ∂1 v3
( dv1 dv2 dv3 ) = ( dx1 dx2 dx3 )  ∂2 v1 ∂2 v2 ∂2 v3  . (3.96)
| {z } | {z }
dvT dxT
∂ v
3 1 ∂ v
3 2 ∂ v
3 3
| {z }
L

We see here that L = ∇vT is the same as presented in Eq. (2.238). If we transpose both
sides of Eq. (3.96), we recover Eq. (3.94):
    
dv1 ∂1 v1 ∂2 v1 ∂3 v1 dx1
 dv2  =  ∂1 v2 ∂2 v2 ∂3 v2   dx2  . (3.97)
dv3 ∂1 v3 ∂2 v3 ∂3 v3 dx3
| {z } | {z } | {z }
dv LT dx

Comparing the form of Eq. (3.97) with that of Eq. (2.282), we see that LT is analogous to the
Jacobian J; however, the Jacobian maps one differential distance vector to another, while
the velocity gradient maps a differential distance vector to a differential velocity vector.

© 05 December 2025, Joseph M. Powers. All rights reserved.

80 CHAPTER 3. KINEMATICS

We can break ∂i vj = L into a symmetric and anti-symmetric part and say then

dvj = dxi ∂(i vj) + dxi ∂[i vj] . (3.98)

| {z } | {z }
shear and extensional straining rotation

We also will find it useful to decompose the velocity gradient tensor L into a deformation
tensor, D:

D = Dij ≡ ∂(i vj) , (3.99)

and a rotation tensor R:

R = Rij ≡ ∂[i vj] . (3.100)

This yields

L = D + R. (3.101)

Thus,

dvj = dxi Dij + dxi Rij = (αi Dij + αi Rij ) ds, (3.102)


dvT = dxT · D + dxT · R = αT · D + αT · R ds, (3.103)


dv = D · dx + RT · dx = D · α + RT · α ds. (3.104)

Let

(s)
dvj = dxi ∂(i vj) = αi ∂(i vj) ds, (3.105)
T
dv(s) = dxT · D = αT · D ds, (3.106)
dv(s) = D · dx = D · α ds. (3.107)

We will see this is associated with straining, both by shear and extension. We will call the
symmetric tensor ∂(i vj) = D the strain rate or deformation tensor. Further, let

(r)
dvj = dxi ∂[i vj] = αi ∂[i vj] ds, (3.108)
T
dv(r) = dxT · R = αT · R ds, (3.109)
dv(r) = RT · dx = RT · α ds. (3.110)

We will see this is associated with rotation as a solid body, with ∂[i vj] = R as the rotation
tensor.

© 05 December 2025, Joseph M. Powers. All rights reserved.

3.7. KINEMATIC DECOMPOSITION OF MOTION 81

[Link] Solid body rotation

(r)
Let us examine dvj . First, we define the vorticity vector ωk as the curl of the velocity field

ωk = ǫkij ∂i vj , (3.111)
ω = ∇ × v. (3.112)

Let us now split the velocity gradient ∂i vj into its symmetric and anti-symmetric parts and
recast the vorticity vector as
ωk = ǫkij ∂(i vj) +ǫkij ∂[i vj] . (3.113)
| {z }
=0

The first term on the right side is zero because it is the tensor inner product of an anti-
symmetric and symmetric tensor. In what remains, we see that half of the vorticity ωk is
actually the dual vector, Ωk , associated with the anti-symmetric ∂[i vj] . See Ch. [Link].

ωk = ǫkij ∂[i vj] = ∇ × v, (3.114)

1 1 1
Ωk = ωk = ǫkij ∂[i vj] = ∇ × v. (3.115)
2 2 2
Using Eq. (2.108) to invert Eq. (3.115), we find

1
∂[i vj] = ǫkij Ωk = ǫkij ωk . (3.116)
2
Thus, we have

(r) 1
dvj = dxi ǫkij ωk , (3.117)
2ω 
k
= ǫkij dxi , (3.118)
2
ω 
k
= ǫjki dxi , (3.119)
2
1 ω
= ω × dr and if Ω= , (3.120)
2 2
= Ω × dr
| {z } . (3.121)
Solid body rotation of one point about another

By introducing this definition for Ω, we see this term takes on the exact form for the
differential velocity due to solid body rotation of P ′ about P from classical rigid body
kinematics. Hence, we give it the same interpretation.

[Link] Straining
Next we consider the remaining term, that we will associate with straining. First, let us
further decompose this into what will be seen to be an extensional (es) straining and a shear

© 05 December 2025, Joseph M. Powers. All rights reserved.

82 CHAPTER 3. KINEMATICS

straining (ss):
(s) (es) (ss)
dvk = dvk + dvk , (3.122)
| {z } | {z }
extension shear
(s) (es) (ss)
dv = dv + dv . (3.123)

[Link].1 Extensional straining Let us define the extensional straining to be the com-
(s)
ponent of straining in the direction of dxj . To do this, we need to project dvj onto the unit
vector αj , then point the result in the direction of that same unit vector;
 
(es) (s)
dvk = αj dvj αk . (3.124)
| {z }
projection of straining

(s)
Now using the definition of dvj , Eq. (3.105), we get
 

(es)


dvk =  vj) ds 
αj |αi ∂(i{z αk , (3.125)
}
(s)
=dvj


= αi ∂(i vj) αj αk ds, (3.126)
(es)


dv = αT · D · α α ds. (3.127)

Now, because αi αj is symmetric, we can be led to a useful result. Consider the series of
operations involving the velocity gradient, in general asymmetric, and a scalar quantity, D:

D = αT · L · α, (3.128)
= αT · (D + R) · α, (3.129)
= αT · D · α + |αT ·{z
R · α}, (3.130)
=0
T
= α · D · α. (3.131)

Thus, we can recast Eq. (3.127) as

dv(es) = αT · L · α α ds. (3.132)

[Link].2 Shear straining What straining that is not aligned with the axis connecting
P and P ′ must then be normal to that axis, and is easily visualized to represent a shearing
between the two points. Hence the shear straining is
(ss) (s) (es)
dvj = dvj − dvj , (3.133)


= ∂(j vi) αi − αi ∂(i vk) αk αj ds, (3.134)

© 05 December 2025, Joseph M. Powers. All rights reserved.

3.7. KINEMATIC DECOMPOSITION OF MOTION 83
 
 
= ∂(j vi) αi − αp ∂(p vk) αk δji αi  ds, (3.135)
|{z}
αj



= ∂(j vi) − αp ∂(p vk) αk δji αi ds, (3.136)
(ss)



dv = D − αT · D · α I · α ds, (3.137)
= (α × (D · α)) × α ds. (3.138)

[Link].3 Principal axes of strain rate We recall from our earlier discussion of Ch. 2.3
that the principal axes of stress are those axes for which the force associated with a given
axis points in the same direction as that axis. We can extend this idea to straining, but
develop it in a slightly different, but ultimately equivalent fashion based on notions from
linear algebra. We first recall that most2 arbitrary asymmetric square matrices L can be
decomposed into a diagonal form as follows:

L = P · Λ · P−1 . (3.139)

Here P is a matrix of the same dimension as L that has in its columns the right eigenvectors
of L. When L is symmetric, it can be shown that its eigenvalues are guaranteed to be real,
and its eigenvectors are guaranteed to be orthogonal. Further, because the eigenvectors can
always be scaled by a constant and remain eigenvectors, we can choose to scale them in such
a way that they are all normalized. In such a case in which the matrix P has orthonormal
columns, the matrix is orthogonal, and we call it Q, as discussed on p. 27. So, when L is
symmetric, such as when L = D, the symmetric part of the velocity gradient, we also have
the following decomposition

D = Q · Λ · Q−1 . (3.140)

Orthogonal matrices have the property that their transpose is equal to their inverse, Eq. (2.15),
and so we also have the even more useful

D = Q · Λ · QT . (3.141)

Geometrically Q can be constructed to be equivalent to a matrix of direction cosines; as we

have seen before when so done, its transpose QT is a rotation matrix that rotates but does
not stretch a vector when it operates on the vector.
Now let us consider the straining component of the velocity difference; taking the sym-
metric ∂(i vj) = D, that we further assume to be a constant for this analysis, we rewrite
2
Some matrices, that often do not have enough linearly independent eigenvectors, cannot be diagonalized;
however, the argument can be extended through use of the singular value decomposition. The singular value
decomposition can also be used to effectively diagonalize asymmetric matrices; however, in that case it can
be shown there is no equivalent interpretation of the principal axes. Consequently, we will quickly focus the
discussion on symmetric matrices.

© 05 December 2025, Joseph M. Powers. All rights reserved.

84 CHAPTER 3. KINEMATICS

Eq. (3.107) using Gibbs notation as

T
dv(s) = dxT · D, (3.142)
dv(s) = DT · dx, (3.143)
= D · dx, because D is symmetric, (3.144)
T
= Q · Λ · Q · dx. (3.145)

Now let us select what amounts to a special axes rotation via matrix multiplication by the
orthogonal matrix QT :

QT · dv(s) = QT · Q · Λ · QT · dx, (3.146)

= Q−1 · Q · Λ · QT · dx, (3.147)
= Λ · QT · dx, (3.148)
T (s)



d Q ·v = Λ · d QT · x , because D and QT are assumed constant. (3.149)
| {z } | {z }
=v′(s) =x′

Recall from the definition of vectors, Eq. (2.45), that QT · v(s) = v′ (s) and QT · x = x′ . That
is, these are the representations of the vectors in a specially rotated coordinate system, so
we have
(s)
dv′ = Λ · dx′ . (3.150)

Now because Λ is diagonal, we see that a perturbation in x′ confined to any one of the
rotated coordinate axes induces a change in velocity that lies in the same direction as that
(s)
coordinate axis. For instance on the 1′ axis, we have dv ′ 1 = Λ11 dx′ 1 . That is to say that in
this specially rotated frame, all straining is extensional; there is no shear straining.

[Link].4 Extensional strain rate quadric Let us study in some more detail the scalar
that is the magnitude of the velocity difference due to extensional strain rate, given by
Eq. (3.131):

D = αT · D · α. (3.151)

For a given D, this is a quadratic equation for the components of α. However, it is subject
to the constraint

αT · α = 1. (3.152)

For a two-dimensional system, for example, this becomes

D = D11 α12 + 2D12 α1 α2 + D22 α22 . (3.153)

© 05 December 2025, Joseph M. Powers. All rights reserved.

3.7. KINEMATIC DECOMPOSITION OF MOTION 85

In (α1 , α2 ) space and fixed D, this may form an ellipse, hyperbola, or circle, depending on
numerical values of Dij . However, we have the constraint

α12 + α22 = 1. (3.154)

One may imagine that as α1 is varied for a given D that α2 will also vary, as will D, and
that there could be extreme values of D, depending on α1 .
Because D is symmetric, we can decompose it into a diagonal form and then say

D = αT · Q · Λ · QT ·α, (3.155)
| {z }
D
= α · Q · Λ · QT · α,
T
(3.156)
= (QT · α)T · Λ · QT · α. (3.157)

Then, defining a rotated coordinate system by α′ = QT · α, we see

D = α′T · Λ · α′ . (3.158)

Because of the diagonal form of Λ, this is easily written in full as

D = λ(1) α1′2 + λ(2) α2′2 + λ(3) α3′2 . (3.159)

These are subject to the constraint that α′ is a unit vector; thus,

1 = α1′2 + α2′2 + α3′2 . (3.160)

One can formally show through techniques of calculus of variations that D has a maximum
given by the maximum eigenvalue and a minimum given by the minimum eigenvalue.

Example 3.4
Analyze D, the magnitude of the velocity difference attributable to extensional strain in a selected
direction α for the deformation tensor
 
1 1
D= . (3.161)
1 5

We first note the eigenvalues of D are given by the roots of the characteristic polynomial

(1 − λ)(5 − λ) − 1 = 0. (3.162)

This yields √ √
λ(1) = 3 + 5 = 5.23607, λ(2) = 3 − 5 = 0.76932. (3.163)
We have

D = D11 α21 + 2D12 α1 α2 + D22 α22 , (3.164)

= α21 + 2α1 α2 + 5α22 . (3.165)

© 05 December 2025, Joseph M. Powers. All rights reserved.

86 CHAPTER 3. KINEMATICS

-1

-2

-3
-3 -2 -1 0 1 2 3

Figure 3.4: Two special contours and one generic contour of D along with the unit circle
illustrating the extreme values of D as well as the orientation of the axes along with the
extreme values of extensional strain are realized.

We can plot contours for which D is constant in the (α1 , α2 ) plane and get an infinite family of curves.
However, we also have a constraint, namely α21 + α22 = 1.
Our Eq. (3.159) suggests that the eigenvalues may well be special values of the contours of D, and
so we examine those two contours:
5.23607 = α21 + 2α1 α2 + 5α22 , (3.166)
0.76932 = α21 + 2α1 α2 + 5α22 . (3.167)
These two contours, the contour for D = 3, and the unit circle α21 + α22 = 1 are plotted in Fig. 3.4. An
infinite family contours of D exist. Many of them will also intersect the unit circle, and so are candidate
solutions. However the special contours we selected are extreme values. For intersection with the unit
circle, we require
D ∈ [λmin , λmax ], (3.168)
∈ [0.76392, 5.23607]. (3.169)
It is easily shown by computing the eigenvectors of D and casting their normalized values into the
columns of the orthogonal matrix Q that the diagonal decomposition of D is
D = Q · Λ · QT , (3.170)
     
1 1 0.229753 −0.973249 5.23607 0 0.229753 0.973249
= . (3.171)
1 5 0.973249 0.229753 0. 0.763932 −0.973249 0.229753
We also have
α′ = QT · α, (3.172)
  
0.229753 0.973249 α1
= , (3.173)
−0.973249 0.229753 α2
 
0.229753α1 + 0.973249α2
= . (3.174)
−0.973249α1 + 0.229753α2

© 05 December 2025, Joseph M. Powers. All rights reserved.

3.8. EXPANSION RATE 87

Our equation for D in rotated and then unrotated coordinates becomes

D = α′T · Λ · α′ , (3.175)
= λ(1) α′2
1+ λ(2) α′2
2 , (3.176)
= 5.23607α1 + 0.763932α′2
′2
2, (3.177)
= 5.23607(0.229753α1 + 0.973249α2)2 + 0.763932(−0.973249α1 + 0.763932α2)2 . (3.178)
Full expansion recovers our original
D = α21 + 2α1 α2 + 5α22 . (3.179)
One should prefer Eq. (3.178) over the original, equivalent, and easy to obtain form of Eq. (3.179).
That is because Eq. (3.178) is obviously an ellipse because of the positive coefficients on the quadratic
terms. Moreover, the form gives the alignment of the major and minor axes via the rotation matrix
QT . Simple trigonometric analysis shows the principal axes are rotated clockwise by θ = 13◦ . This can
be found by considering one of the orthonormal eigenvectors in Q and computing
0.229753
θ = arctan = −13.2825◦. (3.180)
−0.973249

3.8 Expansion rate

Consider a small material region of fluid, also called a particle of fluid. As introduced in
Ch. [Link].1, we define a material region as a region enclosed by a surface across which there
is no flux of mass. We shall later see in Ch. 4.1 by invoking the mass conservation axiom for a
non-relativistic system, that the implication is that the mass of a material region is constant,
but we need not yet consider this. In general the volume containing this particle can increase
or decrease. It is useful to quantify the rate of this increase or decrease. Additionally, this
will give a flavor of the analysis to come for the conservation axioms.
Taking Vm (t) to denote the time-dependent finite material region in space, we must have
Z
Vm = dV. (3.181)
Vm (t)

Using the Reynolds transport theorem, Eq. (2.273), we take the time derivative of both sides
and obtain
Z Z
dVm
= ∂o (1) dV + ni vi dA, (3.182)
dt Vm (t) | {z } Am (t)
=0
Z
= ni vi dA, (3.183)
Am (t)
Z
= ∂i vi dV, by Gauss’s theorem, Eq. (2.252), (3.184)
Vm (t)
= (∂i vi )∗ Vm , by the mean value theorem. (3.185)

© 05 December 2025, Joseph M. Powers. All rights reserved.

88 CHAPTER 3. KINEMATICS

10

0 1 2 3 4 x

Figure 3.5: Plot illustrating the mean value theorem applied to the function f (x) = 5+x2 /5+
sin 5x within the domain x ∈ [1, 3]. The mean value theorem holds that the mean value of
the function is the area under the curve, i.e. the integral,
R 3 scaled by the domain length, here
3 − 1 = 2. So the mean value theorem gives f∗ = ( 1 (5 + x2 /5 + sin 5x) dx)/(3 − 1) =
(88 + 3(sin 5)(sin 10))/15 = 5.97.

We recall from calculus the mean value theorem that states

Rb that for any integral, a mean
value can be defined, denoted by a ∗, as for example a f (x) dx = f∗ (b − a). We give a
simple example illustrating the mean value theorem in Fig. 3.5. As we shrink the size of the
material volume to zero, the mean value approaches the local value, so we get
1 dVm
= (∂i vi )∗ , (3.186)
Vm dt
1 dVm
lim = ∂i vi = ∇T · v = div v = tr D = ϑ. (3.187)
Vm →0 Vm dt

Equation (3.187) describes the relative expansion rate, ϑ, also known as the dilatation rate of
a material fluid particle. A fluid particle for which ∂i vi = 0 must have a relative expansion
rate of zero, and satisfies conditions to be an incompressible fluid. The velocity field for an
incompressible fluid is solenoidal.

3.9 Invariants of the strain rate tensor

The tensor associated with straining (also called the deformation rate tensor or strain rate
(i)
tensor) ∂(i vj) is symmetric. Consequently, it has three real eigenvalues, λǫ̇ , and an orienta-
tion for which the strain rate is aligned with the eigenvectors. As with stress, there are also
three principal invariants of strain rate, the analog to Eqs. (2.158-2.160):
(1) (1) (2) (3)
Iǫ̇ = ∂(i vi) = ∂i vi = λǫ̇ + λǫ̇ + λǫ̇ , (3.188)
(2) 1 (1) (2) (2) (3) (3) (1)
Iǫ̇ = (∂(i vi) ∂(j vj) − ∂(i vj) ∂(j vi) ) = λǫ̇ λǫ̇ + λǫ̇ λǫ̇ + λǫ̇ λǫ̇ , (3.189)
2

© 05 December 2025, Joseph M. Powers. All rights reserved.

3.10. INVARIANTS OF THE VELOCITY GRADIENT TENSOR 89

(3) (1) (2) (3)

Iǫ̇ = ǫijk ∂(1 vi) ∂(2 vj) ∂(3 vk) = λǫ̇ λǫ̇ λǫ̇ . (3.190)
(1)
The physical interpretation for Iǫ̇ is obvious in that it is equal to the relative rate of
(2)
volume change for a material element, (1/V )dV /dt. Aris (1962) discusses how Iǫ̇ is related
(3)
to (1/V )d2 V /dt2 and Iǫ̇ is related to (1/V )d3 V /dt3.

3.10 Invariants of the velocity gradient tensor

For completeness, the invariants of the more general velocity gradient tensor are included.
They are also analogous to Eqs. (2.158-2.160):
(1) (1) (2) (3)
I∇v = ∂i vi = λ∇v + λ∇v + λ∇v , (3.191)
(2) 1 (1) (2) (2) (3) (3) (1)
I∇v = ((∂i vi )(∂j vj ) − (∂i vj )(∂j vi )) = λ∇v λ∇v + λ∇v λ∇v + λ∇v λ∇v , (3.192)
2
1

= (∂i vi )(∂j vj ) + ∂[i vj] ∂[i vj] − ∂(i vj) ∂(i vj) , (3.193)
2 
1 1
= (∂i vi )(∂j vj ) + ωi ωi − ∂(i vj) ∂(i vj) , (3.194)
2 2
(3) (1) (2) (3)
I∇v = ǫijk ∂1 vi ∂2 vj ∂3 vk = λ∇v λ∇v λ∇v . (3.195)

3.11 Two-dimensional kinematics

Next, consider some important two-dimensional cases, first for general two-dimensional flows,
and then for specific examples.

3.11.1 General two-dimensional flows

For two-dimensional motion, we have the velocity vector as (v1 , v2 , v3 = 0), and for the unit
tangent of the vector separating two nearby particles (α1 , α2 , α3 = 0).

[Link] Rotation
Recalling that dxi = αi ds, for rotation, we have from Eq. (3.108)
(r)
dvj = ∂[i vj] dxi = αi ∂[i vj] ds, (3.196)


= α1 ∂[1 vj] + α2 ∂[2 vj] ds, (3.197)
 
(r)
dv1 = α1 ∂[1 v1] +α2 ∂[2 v1]  ds, (3.198)
| {z }
=0
= α2 ∂[2 v1] ds, (3.199)

© 05 December 2025, Joseph M. Powers. All rights reserved.

90 CHAPTER 3. KINEMATICS
 
(r)
dv2 = α1 ∂[1 v2] + α2 ∂[2 v2]  ds, (3.200)
| {z }
=0
= α1 ∂[1 v2] ds. (3.201)

Rewriting in terms of the actual derivatives, we get

(r) 1
dv1 = α2 (∂2 v1 − ∂1 v2 ) ds, (3.202)
2
(r) 1
dv2 = α1 (∂1 v2 − ∂2 v1 ) ds. (3.203)
2
Also for the vorticity vector, we get

ωk = ǫkij ∂i vj . (3.204)

The only non-zero component is ω3 , that comes to

ω3 = ǫ311 ∂1 v1 + ǫ312 ∂1 v2 + ǫ321 ∂2 v1 + ǫ322 ∂2 v2 , (3.205)

|{z} |{z} |{z} |{z}
=0 =1 =−1 =0
= ∂1 v2 − ∂2 v1 . (3.206)

[Link] Extension

(es)
dvk = αk αi αj ∂(i vj) ds, (3.207)


= αk α1 α1 ∂(1 v1) + α1 α2 ∂(1 v2) + α2 α1 ∂(2 v1) + α2 α2 ∂(2 v2) ds (3.208)


= αk α12 ∂1 v1 + α1 α2 (∂1 v2 + ∂2 v1 ) + α22 ∂2 v2 ds, (3.209)
(es)

dv1 = α1 α12 ∂1 v1 + α1 α2 (∂1 v2 + ∂2 v1 ) + α22 ∂2 v2 ds, (3.210)
(es)

dv2 = α2 α12 ∂1 v1 + α1 α2 (∂1 v2 + ∂2 v1 ) + α22 ∂2 v2 ds. (3.211)

[Link] Shear

(ss) (s) (es)

dvj = dvj − dvj , (3.212)


= αi ∂(i vj) − αj αi αk ∂(i vk) ds, (3.213)
  
(ss) ∂2 v1 + ∂1 v2
dv1 = α1 ∂1 v1 + α2 −
2



α1 α12 ∂1 v1 + α1 α2 (∂1 v2 + ∂2 v1 ) + α22 ∂2 v2 ds, (3.214)
  
(ss) ∂1 v2 + ∂2 v1
dv2 = α2 ∂2 v2 + α1 −
2



α2 α12 ∂1 v1 + α1 α2 (∂1 v2 + ∂2 v1 ) + α22 ∂2 v2 ds. (3.215)

© 05 December 2025, Joseph M. Powers. All rights reserved.

3.11. TWO-DIMENSIONAL KINEMATICS 91

P ds

Figure 3.6: Sketch of fluid particle at P and P ′ in motion.

[Link] Expansion

1 dV
= ∂1 v1 + ∂2 v2 . (3.216)
V dt

3.11.2 Relative motion along 1 axis

Let us consider in detail the configuration shown in Fig. 3.6 in which the particle separation
is along the 1 axis. Hence α1 = 1, α2 = 0, and α3 = 0.
• Rotation

(r)
dv1 = 0, (3.217)
(r) 1 ω3
dv2 = (∂1 v2 − ∂2 v1 ) ds = ds. (3.218)
2 2
• Extension

(es)
dv1 = ∂1 v1 ds, (3.219)
(es)
dv2 = 0. (3.220)

• Shear
(ss)
dv1 = 0, (3.221)
(ss) 1
dv2 = (∂1 v2 + ∂2 v1 ) ds = ∂(1 v2) ds. (3.222)
2
• Expansion:
1 dV
= ∂1 v1 + ∂2 v2 . (3.223)
V dt

© 05 December 2025, Joseph M. Powers. All rights reserved.

92 CHAPTER 3. KINEMATICS

ds
P

Figure 3.7: Sketch of fluid particle at P and P ′ in motion.

3.11.3 Relative motion along 2 axis

Let us consider in detail the configuration shown in Fig. 3.7 in which the particle separation
is aligned with the 2 axis. Hence α1 = 0, α2 = 1, and α3 = 0.

• Rotation

(r) 1 ω3
dv1 = (∂2 v1 − ∂1 v2 ) ds = − ds, (3.224)
2 2
(r)
dv2 = 0. (3.225)

• Extension
(es)
dv1 = 0, (3.226)
(es)
dv2 = ∂2 v2 ds. (3.227)

• Shear
(ss) 1
dv1 = (∂2 v1 + ∂1 v2 ) ds = ∂(1 v2) ds, (3.228)
2
(ss)
dv2 = 0. (3.229)

• Expansion:
1 dV
= ∂1 v1 + ∂2 v2 . (3.230)
V dt

© 05 December 2025, Joseph M. Powers. All rights reserved.

3.11. TWO-DIMENSIONAL KINEMATICS 93

Figure 3.8: Sketch of uniform flow.

3.11.4 Uniform flow

Consider the kinematics of a uniform two-dimensional flow in which

v1 = k1 , v2 = k2 , v3 = 0, (3.231)

as sketched in Fig. 3.8.

 
dx1 dx2 dx1 dx2 k1
• Streamlines: v1
= v2
, k1
= k2
, x1 = k2
x2 + C.

• Rotation: ω3 = ∂1 v2 − ∂2 v1 = ∂1 (k1 ) − ∂2 (k2 ) = 0.

• Extension

– on 1-axis: ∂1 v1 = 0.
– on 2-axis: ∂2 v2 = 0.
1
• Shear for unrotated element: 2
(∂1 v2 + ∂2 v1 ) = 0.

• Expansion: ∂1 v1 + ∂2 v2 = 0.

• Acceleration:
dv1
dt
= ∂o v1 + v1 ∂1 v1 + v2 ∂2 v1 = 0 + k1 ∂1 (k1 ) + k2 ∂2 (k1 ) = 0,
dv2
dt
= ∂o v2 + v1 ∂1 v2 + v2 ∂2 v2 = 0 + k1 ∂1 (k2 ) + k2 ∂2 (k2 ) = 0.

For this simple flow, the streamlines are straight lines, there is no rotation, no extension, no
shear, no expansion, and no acceleration.

© 05 December 2025, Joseph M. Powers. All rights reserved.

94 CHAPTER 3. KINEMATICS

Figure 3.9: Sketch of pure rigid body rotation.

3.11.5 Pure rigid body rotation

Consider the kinematics of a two-dimensional flow in which

v1 = −kx2 , v2 = kx1 , v3 = 0, (3.232)

as sketched in Fig. 3.9.

dx1 dx2 dx1 dx2
• Streamlines: v1
= v2
, −kx2
= kx1
, x1 dx1 = −x2 dx2 , x21 + x22 = C.

• Rotation: ω3 = ∂1 v2 − ∂2 v1 = ∂1 (kx1 ) − ∂2 (−kx2 ) = 2k.

• Extension

– on 1-axis: ∂1 v1 = 0,
– on 2-axis: ∂2 v2 = 0.
1
• Shear for unrotated element: 2
(∂1 (kx1 ) + ∂2 (−kx2 ) = k − k = 0.

• Expansion: ∂1 v1 + ∂2 v2 = 0 + 0 = 0.

• Acceleration:
dv1
dt
= ∂o v1 + v1 ∂1 v1 + v2 ∂2 v1 = 0 − kx2 ∂1 (−kx2 ) + kx1 ∂2 (−kx2 ) = −k 2 x1 ,
dv2
dt
= ∂o v2 + v1 ∂1 v2 + v2 ∂2 v2 = 0 − kx2 ∂1 (kx1 ) + kx1 ∂2 (kx1 ) = −k 2 x2 .

In this flow, the velocity magnitude grows linearly with distance from the origin. This is
precisely how a rotating rigid body behaves. The streamlines are circles. The rotation is
positive for positive k, hence counterclockwise, there is no deformation in extension or shear,
and there is no expansion. The acceleration is pointed towards the origin.

© 05 December 2025, Joseph M. Powers. All rights reserved.

3.11. TWO-DIMENSIONAL KINEMATICS 95

Figure 3.10: Sketch of extensional flow (1-D compressible).

3.11.6 Pure extensional motion (a compressible flow)

Consider the kinematics of a two-dimensional flow in which

v1 = kx1 , v2 = 0, v3 = 0, (3.233)

as sketched in Fig. 3.10.

dx1 dx2
• Streamlines: v1
= v2
, v2 dx1 = v1 dx2 , 0 = kx1 dx2 , x2 = C.

• Rotation: ω3 = ∂1 v2 − ∂2 v1 = ∂1 (0) − ∂2 (kx1 ) = 0.

• Extension

– on 1-axis: ∂1 v1 = k,
– on 2-axis: ∂2 v2 = 0.
1 1
• Shear for unrotated element: 2
(∂1 v2 + ∂2 v1 ) = 2
(∂1 (0) + ∂2 (kx1 )) = 0.

• Expansion: ∂1 v1 + ∂2 v2 = k.

• Acceleration:
dv1
dt
= ∂o v1 + v1 ∂1 v1 + v2 ∂2 v1 = 0 + kx1 ∂1 (kx1 ) + 0∂2 (kx1 ) = k 2 x1 ,
dv2
dt
= ∂o v2 + v1 ∂1 v2 + v2 ∂2 v2 = 0 + kx1 ∂1 (0) + 0∂2 (0) = 0.

In this flow, the streamlines are straight lines; there is no fluid rotation; there is extension
(stretching) deformation along the 1-axis, but no shear deformation along this axis. The rela-
tive expansion rate is positive for positive k, indicating a compressible flow. The acceleration
is confined to the x1 direction.

© 05 December 2025, Joseph M. Powers. All rights reserved.

96 CHAPTER 3. KINEMATICS

Figure 3.11: Sketch of pure shearing flow.

3.11.7 Pure shear straining

Consider the kinematics of a two-dimensional flow in which

v1 = kx2 , v2 = kx1 , v3 = 0, (3.234)

as sketched in Fig. 3.11.

dx1 dx2 dx1 dx2
• Streamlines: v1
= v2
, kx2
= kx1
, x1 dx1 = x2 dx2 , x21 = x22 + C.

• Rotation: ω3 = ∂1 v2 − ∂2 v1 = ∂1 (kx1 ) − ∂2 (kx2 ) = k − k = 0.

• Extension

– on 1-axis: ∂1 v1 = ∂1 (kx2 ) = 0,
– on 2-axis: ∂2 v2 = ∂2 (kx1 ) = 0.
1 1
• Shear for unrotated element: 2
(∂1 v2 + ∂2 v1 ) = 2
(∂1 (kx1 ) + ∂2 (kx2 )) = k.

• Expansion: ∂1 v1 + ∂2 v2 = 0.

• Acceleration:
dv1
dt
= ∂o v1 + v1 ∂1 v1 + v2 ∂2 v1 = 0 + kx2 ∂1 (kx2 ) + kx1 ∂2 (kx2 ) = k 2 x1 ,
dv2
dt
= ∂o v2 + v1 ∂1 v2 + v2 ∂2 v2 = 0 + kx2 ∂1 (kx1 ) + kx1 ∂2 (kx1 ) = k 2 x2 .

In this flow, the streamlines are hyperbolas; there is no rotation or axial extension along
the coordinate axes; there is positive shear deformation for an element aligned with the
coordinate axes, and no expansion. So, the pure shear deformation preserves volume. The
fluid is accelerating away from the origin.

© 05 December 2025, Joseph M. Powers. All rights reserved.

3.11. TWO-DIMENSIONAL KINEMATICS 97

Figure 3.12: Sketch of an ideal corner flow.

3.11.8 Ideal corner flow

Consider the kinematics of a two-dimensional flow in which

v1 = kx1 , v2 = −kx2 , v3 = 0, (3.235)

as sketched in Fig. 3.12.

dx1 dx2 dx1 dx2
• Streamlines: v1
= v2
, kx1
= −kx2
, ln x1 = − ln x2 + C ′ , x1 x2 = C.

• Rotation: ω3 = ∂1 v2 − ∂2 v1 = ∂1 (−kx2 ) − ∂2 (kx1 ) = 0.

• Extension

– on 1-axis: ∂1 v1 = ∂1 (kx1 ) = k,
– on 2-axis: ∂2 v2 = ∂2 (−kx2 ) = −k.
1 1
• Shear for unrotated element: 2
(∂1 v2 + ∂2 v1 ) = 2
(∂1 (−kx2 ) + ∂2 (kx1 )) = 0.

• Expansion: ∂1 v1 + ∂2 v2 = k − k = 0.

• Acceleration:
dv1
dt
= ∂o v1 + v1 ∂1 v1 + v2 ∂2 v1 = 0 + kx1 ∂1 (kx1 ) − kx2 ∂2 (kx1 ) = k 2 x1 ,
dv2
dt
= ∂o v2 + v1 ∂1 v2 + v2 ∂2 v2 = 0 + kx1 ∂1 (−kx2 ) − kx2 ∂2 (−kx2 ) = k 2 x2 .

In this flow, the streamlines are hyperbolas; there is no rotation or shear along the coordinate
axes; there is extensional strain for an element aligned with the coordinate axes, but no net
expansion. So, the ideal corner flow preserves volume. The fluid is accelerating away from
the origin.

© 05 December 2025, Joseph M. Powers. All rights reserved.

98 CHAPTER 3. KINEMATICS

Figure 3.13: Sketch of Couette flow.

3.11.9 Couette flow: shear + rotation

Consider the kinematics of a two-dimensional flow in which

v1 = kx2 , v2 = 0, v3 = 0, (3.236)

as sketched in Fig. 3.13. This is known as a Couette3 flow.

dx1 dx2 dx1 dx2
• Streamlines: v1
= v2
, kx2
= 0
, 0 = kx2 dx2 , x2 = C.

• Rotation: ω3 = ∂1 v2 − ∂2 v1 = ∂1 (0) − ∂2 (kx2 ) = −k.

• Extension

– on 1-axis: ∂1 v1 = ∂1 (kx2 ) = 0,
– on 2-axis: ∂2 v2 = ∂2 (0) = 0.
1 1
• Shear for unrotated element: 2
(∂1 v2 + ∂2 v1 ) = 2
(∂1 (0) + ∂2 (kx2 )) = k2 .

• Expansion: ∂1 v1 + ∂2 v2 = 0.

• Acceleration:
dv1
dt
= ∂o v1 + v1 ∂1 v1 + v2 ∂2 v1 = 0 + kx2 ∂1 (kx2 ) + 0∂2 (kx2 ) = 0,
dv2
dt
= ∂o v2 + v1 ∂1 v2 + v2 ∂2 v2 = 0 + kx2 ∂1 (0) + 0∂2 (0) = 0.

Here the streamlines are straight lines, and the flow is rotational (clockwise because ω < 0 for
k > 0)! The constant volume rotation is combined with a constant volume shear deformation
for the element aligned with the coordinate axes. The fluid is not accelerating.

3
Maurice Marie Alfred Couette, 1858-1943, French fluid mechanician, rheologist and teacher; student of
Joseph Valentin Boussinesq, and faculty member at Catholic University of Angers.

© 05 December 2025, Joseph M. Powers. All rights reserved.

3.12. THREE-DIMENSIONAL KINEMATICS: SUMMARY 99

Figure 3.14: Sketch of ideal irrotational vortex.

3.11.10 Ideal irrotational vortex: extension + shear

Consider the kinematics of a two-dimensional flow sketched in Fig. 3.14.
x2 x1
v1 = −k , v2 = k , v3 = 0. (3.237)
x21 + x22 x21 + x22

dx1 dx2 dx1 dx2

• Streamlines: v1
= v2
, −k
x2 = k
x1 , − dx
x2
1
= dx2
x1
, x21 + x22 = C.
x2 2
1 +x2 x2 2
1 +x2

   
• Rotation: ω3 = ∂1 v2 − ∂2 v1 = ∂1 k x2x+x
1
2 − ∂2 −k x2
x2 +x2
= 0.
1 2 1 2

• Extension
 
x2
– on 1-axis: ∂1 v1 = ∂1 −k x2 +x2 = 2k (x2x+x 1 x2
2 )2 ,
1 2 1 2
 
– on 2-axis: ∂2 v2 = ∂2 k x2x+x
1
2 = −2k (x2x+x
1 x2
2 )2 .
1 2 1 2

1 x2 −x2
• Shear for unrotated element: 2
(∂1 v2 + ∂2 v1 ) = k (x22+x21)2 .
1 2

• Expansion: ∂1 v1 + ∂2 v2 = 0.

• Acceleration:
dv1 2
dt
= ∂o v1 + v1 ∂1 v1 + v2 ∂2 v1 = − (x2k+x
x1
2 )2 ,
1 2
dv2 2
dt
= ∂o v2 + v1 ∂1 v2 + v2 ∂2 v2 = − (x2k+x
x2
2 )2 .
1 2

The streamlines are circles, and the fluid element does not rotate about its own axis! It does
rotate about the origin. It deforms by extension and shear in such a way that overall the
volume is constant.

© 05 December 2025, Joseph M. Powers. All rights reserved.

100 CHAPTER 3. KINEMATICS

P
x
D

x
a

Figure 3.15: Sketch of three-dimensional kinematics.

3.12 Three-dimensional kinematics: summary

We graphically summarize three-dimensional kinematics in Fig. 3.15. Here we depict the
motion of a fluid particle P , that has Eulerian position vector x, velocity vector v, and
acceleration vector a. The particle P is also shown undergoing a solid-body rotation about
an axis aligned with the vorticity vector ω. It also undergoes an extensional deformation
aligned with the principal axes of the deformation tensor D. As the position x is varied
continuously, all field quantities, v, a, ω, and D, vary continuously as well.

3.13 Kinematics as a dynamical system

Let us apply some standard notions from dynamical systems theory (see Powers and Sen
(2015), Ch. 9.6; Powers (2016), Ch. 6.1; or Paolucci (2016), Ch. 3.2.6) to fluid kinematics.

3.13.1 General three-dimensional analysis

Let us imagine that we are given a time-independent flow field, where the fluid velocity is
known and is a function of position only. Then the motion of an individual fluid particle is
governed by the following autonomous system of non-linear ordinary differential equations:
dx
= v(x(t)), x(0) = xo . (3.238)
dt
Here, the initial position of the fluid particle is given by the constant vector xo . The solution
of Eq. (3.238) can be expressed in general form
x = x(t; xo ), (3.239)
a function of time parameterized by the initial condition of the fluid particle. Such a solution
is certainly a pathline, streamline, and streakline. It is also known as a trajectory in the
dynamical systems literature.

© 05 December 2025, Joseph M. Powers. All rights reserved.

3.13. KINEMATICS AS A DYNAMICAL SYSTEM 101

Let us analyze Eq. (3.238) in some more detail. From the definition of the total derivative,
see Eq. (3.94), we have

dv = (∇vT )T ·dx, (3.240)

| {z }
LT
T
= L · dx. (3.241)

This gives the acceleration vector as

dv dx
= LT · , (3.242)
dt dt
= LT · v. (3.243)

Example 3.5
Study the following non-linear autonomous system, that could describe the steady three-dimensional
kinematics of a fluid:
dx1
= v1 (x1 , x2 , x3 ) = 1 + x1 x2 x3 , xo1 = 0, (3.244)
dt
dx2
= v2 (x1 , x2 , x3 ) = x1 + x22 + x1 x32 , xo2 = 0, (3.245)
dt
dx3
= v3 (x1 , x2 , x3 ) = 2 − x1 + x2 x3 , xo3 = 0. (3.246)
dt

When considering dynamic systems, one should always consider equilibrium points. Such points
exist when v1 = v2 = v3 = 0; in fluid mechanics, they are known as stagnation points. They are found
by solving the nonlinear algebra problem:

0 = v1 (x1 , x2 , x3 ) = 1 + x1 x2 x3 , (3.247)
0 = v2 (x1 , x2 , x3 ) = x1 + + x22 x1 x32 , (3.248)
0 = v3 (x1 , x2 , x3 ) = 2 − x1 + x2 x3 . (3.249)

Numerical solution reveals three roots. Two of them are complex, and a third real. As we are generally
only concerned with real solutions, we focus only on that root, which is

x1o = 1, x2o = −1.46557, x3o = 0.682328. (3.250)

The “o” as a subscript denotes a stagnation condition, in contrast to “o” as a superscript, that denotes
an initial condition. Direct substitution into the equations for velocity confirm this is a stagnation
point. Flow in the neighborhood of a stagnation point can be understood by considering the locally
linear behavior. Taylor4 series of the velocity in the neighborhood of the stagnation point reveals the
4
Brook Taylor, 1685-1731, English mathematician and artist, educated at Cambridge, published on cap-
illary action, magnetism, and thermometers, adjudicated the dispute between Newton and Leibniz over
priority in developing calculus, contributed to the method of finite differences, invented integration by parts,
has his name ascribed to Taylor series of which variants were earlier discovered by Gregory, Newton, Leibniz,
Johann Bernoulli, and de Moivre.

© 05 December 2025, Joseph M. Powers. All rights reserved.

102 CHAPTER 3. KINEMATICS

local kinematics are given by the locally linear system

 dx1   ∂v1 ∂v1 ∂v1   

dt ∂x1 ∂x2 ∂x3 x1 − x1o
 dx2  =  ∂v2 ∂v2 ∂v2   x2 − x2o  , (3.251)
dt ∂x1 ∂x2 ∂x3
dx3 ∂v3 ∂v3 ∂v3 x3 − x3o
dt ∂x1 ∂x ∂x3
| {z2 }
=LT xio

   
x2 x3 x1 x3 x1 x2 x1 − x1o
=  1 + x32 2x2 + 3x1 x22 0   x2 − x2o  , (3.252)
−1 x3 x2 x3 − x3o
| {z }
=LT xio

   
−1. 0.682328 −1.46557 x1 − x1o
=  −2.1479 3.51255 0   x2 − x2o  . (3.253)
−1 0.682328 −1.46557 x3 − x3o
| {z }
=LT xio

As discussed in standard texts on applied mathematics, e.g. Powers and Sen (2015), Ch. 9, the local
dynamics in the neighborhood of the stagnation point are dictated by the eigenvalues of the coefficient
matrix. Those are easily numerically evaluated as λ = 3.25643, −2.20944, and 0. The positive, negative,
and zero eigenvalues are associated with unstable, stable, and neutrally stable modes, respectively.
Because of the presence of both stable and unstable modes, this stagnation point is a so-called saddle
node. Had all eigenvalues been real and positive, it would have been an unstable source node. Had
all eigenvalues been real and negative, it would have been a stable sink node. Had any eigenvalues
contained an imaginary component, the solution could take on a locally oscillatory behavior.
Numerical solution of this nonlinear system of ordinary differential equations yields x1 (t), x2 (t),
x3 (t), that for this time-independent velocity field induces the particle pathlines, streamlines, and
streaklines. All are plotted in Fig. 3.16. We could also apply the complete mathematical theory of
dynamic systems to understand the system better.
We can use Eq. (3.243) to calculate the acceleration vector field:
 dv1   ∂v1 ∂v1 ∂v1   dx1 
dt ∂x1 ∂x2 ∂x3 dt
 dv2  =  ∂v2 ∂v2 ∂v2   dx2  , (3.254)
dt ∂x1 ∂x2 ∂x3 dt
dv3 ∂v3 ∂v3 ∂v3 dx3
dt ∂x1 ∂x2 ∂x3 dt
  
x2 x3 x1 x3 x1 x2 1 + x1 x2 x3
=  1 + x32 2x2 + 3x1 x22 0   x1 + x22 + x1 x32  , (3.255)
−1 x3 x2 2 − x1 + x2 x3
 
2x1 x2 − x1 x2 + x1 x3 + x2 x3 + 2x1 x22 x3 + x21 x32 x3 + x1 x22 x23
2 2

=  1 + 2x1 x2 + 3x21 x22 + 3x32 + 5x1 x42 + 3x21 x52 + x1 x2 x3 + x1 x42 x3  . (3.256)
−1 + 2x2 − x1 x2 + x1 x3 − x1 x2 x3 + 2x22 x3 + x1 x32 x3

If we know the kinematics of a fluid particle, we know everything about its motion, including its
acceleration. We shall soon, Ch. 4.2, discuss things like Newton’s second law of motion that relates
acceleration to forces. If we know the acceleration, it is possible to induce what the force was that
generated it by simply multiplying the acceleration by the mass. Rarely is this the case however. It is
more common to know something about the forces and to use this to deduce what the motion is.

© 05 December 2025, Joseph M. Powers. All rights reserved.

3.13. KINEMATICS AS A DYNAMICAL SYSTEM 103

1.2
1.5
1.0

0.8
1.0
0.6

0.4
0.5
0.2

t t
0.2 0.4 0.6 0.8 1.0 0.2 0.4 0.6 0.8 1.0

0.6

0.5

0.4

0.3

0.2

0.1

t
0.2 0.4 0.6 0.8 1.0

Figure 3.16: Plot of x1 (t), x2 (t), x3 (t), along with the coincident pathline, streamline, and
streakline for a steady three-dimensional fluid particle that commences at the origin.

Now, we seek to analyze a particular pathline. Note that the velocity vector is tangent
to the fluid particle trajectory. Let us study a unit vector that happens to be tangent to the
velocity field:
v
αt = . (3.257)
||v||
Next, use the quotient rule to examine how the unit tangent vector evolves with time:
dαt 1 dv v d||v||
= − . (3.258)
dt ||v|| dt ||v||2 dt

We can scale Eq. (3.243) by ||v|| to get (1/||v||)dv/dt = LT · v/||v|| = LT · αt . Thus

Eq. (3.258) can be rewritten as
dαt v d||v||
= LT · αt − , (3.259)
dt ||v||2 dt
1 d||v||
= LT · αt − αt . (3.260)
||v|| dt
Next consider the following series of operations starting with Eq. (3.243):
dv
= LT · v, (3.261)
dt

© 05 December 2025, Joseph M. Powers. All rights reserved.

104 CHAPTER 3. KINEMATICS

dv
vT · = vT · LT · v, (3.262)
 dt

d vT · v
= vT · LT · v, (3.263)
dt 2
 
d ||v||2
= vT · LT · v, (3.264)
dt 2
d
||v|| (||v||) = vT · LT · v, (3.265)
dt
1 d vT v
(||v||) = · LT · , (3.266)
||v|| dt ||v|| ||v||
= αTt · LT · αt . (3.267)

Now substitute Eq. (3.267) into Eq. (3.260) to get

dαt

= LT · αt − αTt · LT · αt αt . (3.268)
dt
As an aside, take the dot product of Eq. (3.268) with αt to get
dαt

αTt · = αTt · LT · αt − αTt · LT · αt αTt · αt , (3.269)
dt | {z }
=1
= αTt T
· L · αt − αTt T
· L · αt , (3.270)
= 0. (3.271)

This must be an identity, because αTt · αt = 1, and its time derivative gives αTt · dα/dt = 0.
Now recalling Eq. (3.101), and employing αTt · RT · αt = 0, because of the anti-symmetry of
R, and DT = D, because of the symmetry of D, Eq. (3.268) can be rewritten as
dαt

= LT · αt − αTt · D · αt αt . (3.272)
dt
Let us consider how a volume stretches in a direction aligned with the velocity vector.
We first specialize the general differential arc length to that found along the particle path:
ds = ds. Now, recall from geometry that the square of the differential arc length must be

ds2 = dxT · dx, (3.273)

where dx is also confined to the particle path. Consider now how this quantity changes with
time when we move with the particle:
d d

(ds)2 = dxT · dx , (3.274)
dt dt
 T
T d d
= dx · (dx) + (dx) · dx, (3.275)
dt dt

© 05 December 2025, Joseph M. Powers. All rights reserved.

3.13. KINEMATICS AS A DYNAMICAL SYSTEM 105

   T
T dx dx
= dx · d + d · dx, (3.276)
dt dt
= dxT · dv + dvT · dx, (3.277)
= 2 dxT · dv, (3.278)
= 2 dxT · LT · dx, (3.279)
d
2 ds (ds) = 2 dxT · LT · dx, (3.280)
dt
1 d dx T dx
(ds) = · LT · . (3.281)
ds dt ds ds
Recall now that
dx
v dt dx
αt = = ds
= . (3.282)
||v|| dt
ds

So, Eq. (3.281) can be rewritten as

1 d
(ds) = αTt · LT · αt , (3.283)
ds dt
d
(ln ds) = αTt · (D + R)T · αt , (3.284)
dt
= αTt · D · αt , (3.285)
= D : αt αTt . (3.286)

This relative tangential stretching rate is closely related to the result of Eq. (3.127) for
extensional strain rate. Specializing Eq. (3.127) for a particle pathline, and combining, we
can say

dv(es) = (αTt · D · αt )αt ds, (3.287)

dv(es)
= (αTt · D · αt )αt , (3.288)
ds
(es)
dv
αTt · = (αTt · D · αt ) αTt · αt , (3.289)
ds | {z }
=1
1 d
= αTt · D · αt = (ds), (3.290)
ds dt 
1 ds
= αTt · D · αt = d , (3.291)
ds dt
d||v||
= αTt · D · αt = , (3.292)
ds
d||v||
= D : αt αTt = . (3.293)
ds

© 05 December 2025, Joseph M. Powers. All rights reserved.

106 CHAPTER 3. KINEMATICS

Here, we invoked Eq. (3.285) to obtain Eq. (3.291). The quantity αTt · D · αt = D : αt αTt
is a measure of how the magnitude of the velocity changes with respect to arc length along
the particle path.
We can gain further insight into how velocity magnitude changes by a diagonal decom-
position of D = Q · Λ · QT , where Q is an orthogonal rotation matrix with the normalized
eigenvectors of D in its columns, and Λ is the diagonal matrix with the eigenvalues of D in
its diagonal. Thus

d||v||
= αTt · Q · Λ · QT ·αt = (QT · αt )T · Λ · (QT · αt ). (3.294)
ds | {z }
D

The operation QT · αt ≡ αs generates a new rotated unit vector αs = (αs1 , αs2 , αs3 )T . Thus
we can state
d||v|| 2 (1) 2 (2) 2 (3)
= αs1 λ + αs2 λ + αs3 λ , (3.295)
ds
2 2 2
1 = αs1 + αs2 + αs3 . (3.296)

The rate of change of the velocity magnitude along a particle pathline can be understood to
be a weighted average of the eigenvalues of the deformation tensor D. In the special case
in which αt is the ith eigenvector of D, we simply get d||v||/ds = λ(i) , where λ(i) is the
corresponding eigenvalue.
If we extend Eq. (3.187) to differential material volumes, we could say the relative ex-
pansion rate is

1 d d
(dV ) = (ln V ) = tr D = ϑ. (3.297)
dV dt dt
Now our differential volume can be formed by

dV = dA ds, (3.298)

where dA is the cross-sectional area normal to the flow direction. Thus

ln dV = ln dA + ln ds, (3.299)
ln dA = ln dV − ln ds, (3.300)
d d d
(ln dA) = (ln dV ) − (ln ds) . (3.301)
dt dt dt
Substitute from Eqs. (3.285,3.297) to get the relative rate of change of the differential area
normal to the flow direction:
d
(ln dA) = tr D − αTt · D · αt . (3.302)
dt

© 05 December 2025, Joseph M. Powers. All rights reserved.

3.13. KINEMATICS AS A DYNAMICAL SYSTEM 107

This relation, while not identical, is similar to the expression for shear strain rate, Eq. (3.137).
We can also use Eq. (2.100) to rewrite Eq. (3.302) as
d

(ln dA) = D : I − D : αt αTt = D : I − αt αTt . (3.303)
dt
Now the matrix I − αt αTt has some surprising properties. It is singular and has rank two.
Because it is symmetric, it has a set of three orthogonal eigenvectors that can be normalized
to form an orthonormal set. Its three eigenvalues are 1, 1, and 0. Remarkably, the eigenvector
associated with the zero eigenvalue must be parallel to and can be selected as αt , the unit
tangent to the curve. Thus the other two eigenvectors can be thought of as unit normals to
the curve, that we label αn1 and αn2 . These eigenvectors are not unique; however, a set can
always be found. We can summarize the decomposition in the following steps:

I − αt αTt = Q · Λ · QT , (3.304)
 . .. ..    
.. . . 1 0 0 · · · αTn1 ···
 
=  αn1 αn2 αt  0 1 0   · · · αTn2 ···, (3.305)
.. .. .. 0 0 0 · · · αTt ···
. . .
= αn1 αTn1 + αn2 αTn2 . (3.306)

The two unit normals are orthogonal to each other, αTn1 · αn2 = 0. Thus, we have
d

(ln dA) = D : αn1 αTn1 + αn2 αTn2 , (3.307)
dt
= D : αn1 αTn1 + D : αn2 αTn2 , (3.308)
= αTn1 · D · αn1 + αTn2 · D · αn2 . (3.309)

Comparing to Eq. (3.285) that has one mode associated with αt available for stretching of
the one-dimensional arc length in the streamwise direction, there are two modes associated
with αn1 , αn2 available for stretching the two-dimensional area.
The form αTn1 · D · αn1 suggests it determines the relative normal stretching rate in the
direction of αn1 ; a similar rate exists for the other normal direction. One might imagine
that there exists a normal direction that yields extreme values for relative normal stretching
rates. It is easily shown this achieved by the following. First, define a rectangular matrix,
b whose columns are populated by αn1 and αn2 :
Q,
 . .. 
.. .
b  
Q =  αn1 αn2  . (3.310)
.. ..
. .
b associated with
Then project the 3 × 3 matrix D onto this basis to form the 2 × 2 matrix D
stretching in the directions normal to the motion:
b =Q
D b T · D · Q.
b (3.311)

© 05 December 2025, Joseph M. Powers. All rights reserved.

108 CHAPTER 3. KINEMATICS

The eigenvalues of Db give the maximum and minimum values of the relative normal stretch-
ing rates, and the eigenvectors give the associated directions of extremal normal stretching.
Looked at another way and motivated by standard results from differential geometry,
we can make special choices, αn1 = αnp , αn2 = αnb , where αnp is the so-called “principal
normal unit vector” and αnb is the so-called “binormal unit vector,”

αnb = αt × αnp . (3.312)

The following results are described in more detail in many sources, e.g. Powers and Sen
(2015), p. 89. We have the so-called “Frenet-Serret”5 relations:

dαt dαnp dαnb

= καnp , = −καt + τ αnb , = −τ αnp . (3.313)
ds ds ds
Here κ is the so-called “curvature,” of the curve and τ is the so-called “torsion” of the curve.
One can, with effort show that κ and τ are given by
r  
2 2 2
d2 x dx dx T d2 x d2 x
dt2 dt
− dt
· dt2
dx
×
dt dt2
κ = 3 = 3 , (3.314)
dx dx
dt dt
 T  
dx d2 x d3 x dx d2 x d3 x
dt
× dt2
· dt3
det , ,
dt dt2 dt3
τ =  2 = 2 . (3.315)
2 2 dx T dx d2 x
d2 x dx
− · d2 x
dt
× dt2
dt2 dt dt dt2

The sign convention for τ here is the most common and is opposite that of Powers and Sen
(2015). Note κ and τ are expressed here as functions of time. This is certainly the case for a
particle moving along a path in time. But just as the intrinsic curvature of a mountain road
is independent of the speed of the vehicle traveling on the road, despite the traveling vehicle
experiencing a time-dependency of curvature, the curvature and torsion can be considered
more fundamentally to be functions of position only, given that the velocity field is known
as a function of position. Analysis reveals in fact that
q
(vT · L · LT · v) (vT · v) − (vT · LT · v)2
κ = . (3.316)
(vT · v)3/2

One could also develop an expression for torsion that is explicitly dependent on position.
The expression is complicated and requires the use of third order tensors to capture the
higher order spatial variations. Note from Eq. (3.313)

1 dαt 1 dα
dt
t
1 dαdt
t

αnp = = = . (3.317)
κ ds κ ds
dt
κ ||v||
5
Jean Frédeŕic Frenet, 1816-1900, and Joseph Alfred Serret, 1819-1885, French mathematicians.

© 05 December 2025, Joseph M. Powers. All rights reserved.

3.13. KINEMATICS AS A DYNAMICAL SYSTEM 109

Then one can use Eq. (3.312) to get αnb .

We can also use this intrinsic orthonormal basis to get
d

(ln dA) = D : αnp αTnp + αnb αTnb , (3.318)
dt
= D : αnp αTnp + D : αnb αTnb , (3.319)
= αTnp · D · αnp + αTnb · D · αnb . (3.320)

One can also show that the binormal vector is orthogonal to both the velocity and
acceleration. Because of this
dx d2 x
dt
× dt2 v×a
αnb = dx d2 x
= . (3.321)
|| dt × dt2
|| ||v × a||

So given v, it is easy to get both αt and a. And given v and a, it is easy to get αnb . Then
one can also get the principal normal unit vector by

αnp = αnb × αt . (3.322)

The principal normal is in the plane defined by v and a and is orthogonal to v. If one
constructs a circle that is locally aligned with the trajectory, the circle will have a radius
that is 1/κ and is known as the radius of curvature. The principle normal vector αnp will
point towards the center of this circle.

Example 3.6
Consider the fluid particle trajectory defined by
 
2 cos t
x(t) =  2 sin t  . (3.323)
0
Find v, a, κ, τ , αt , αnb , and αnp .

We see by inspection that the trajectory is a circle of radius 2. At t = 0, the fluid particle is at the
point (2, 0, 0)T . We get
   
−2 sin t −2 cos t
dx  dv 
v= = 2 cos t  , a= = −2 sin t  . (3.324)
dt dt
0 0
Then the unit tangent to the curve, that is the direction of the velocity vector, is
 
− sin t
v
αt = =  cos t  . (3.325)
||v||
0
The binormal is
v×a
αnb = . (3.326)
||v × a||

© 05 December 2025, Joseph M. Powers. All rights reserved.

110 CHAPTER 3. KINEMATICS

The relevant cross product is

 
i j k 0
v × a = −2 sin t 2 cos t 0 = 0. (3.327)
−2 cos t −2 sin t 0 4

Obviously ||v × a|| = 4, so

 
0
αnb = 0. (3.328)
1

Then the principal normal is

 
i j k − cos t
αnp = αnb × αt = 0 0 1 =  − sin t  . (3.329)
− sin t cos t 0 0

The curvature is
||v × a|| 4 1
κ= 3
= 3 = . (3.330)
||v|| 2 2
Thus, the radius of curvature is 1/κ = 2, that is the radius of the circle. The torsion is τ = 0 because
d3 x/dt3 = 0. This is illustrated in Fig. 3.17, where we focus on the configuration at t = π/4.

-2 -1 1 2

-1

-2

Figure 3.17: Kinematic features of a fluid particle with position vector x(t) =
(2 cos t, 2 sin t, 0)T at t = π/4 rotating in a circle of radius 2 about the origin.

The following example, adapted from Powers and Sen (2015), illustrates how kinematics
illuminates the general field of nonlinear dynamical systems.

© 05 December 2025, Joseph M. Powers. All rights reserved.

3.13. KINEMATICS AS A DYNAMICAL SYSTEM 111

Example 3.7
Consider the example of Mengers6

dx1 1
= (1 − x21 ), (3.331)
dt 20
dx2 35
= −2x2 − x3 + 2(1 − x21 )x3 , (3.332)
dt 16
dx3
= x2 + x3 , (3.333)
dt
and identify so-called heteroclinic trajectories and their attractiveness.

There are only two finite equilibria for this system, a saddle at (−1, 0, 0)T and a sink at (1, 0, 0).
Because the first equation is uncoupled from the second two and is sufficiently simple, it can be inte-
grated exactly to form x1 = tanh(t/20). This, coupled with x2 = 0 and x3 = 0, satisfies all differential
equations and connects the equilibria, so the x1 axis for x1 ∈ [−1, 1] is what is known as the heteroclinic
trajectory. One then asks if nearby trajectories are attracted to it. This can be answered by a local
geometry-based analysis. Our system is of the form dx/dt = v(x). Let us consider its behavior in the
neighborhood of a generic point x0 that is on the heteroclinic trajectory, but is far from equilibrium.
We then locally linearize our system as

d
(x − x0 ) = v(x0 ) + L|x0 · (x − x0 ) + . . . , (3.334)
dt | {z } | {z }
translation deformation+rotation
= v(x0 ) + D|x0 · (x − x0 ) + R|x0 · (x − x0 ) + . . . . (3.335)
| {z } | {z } | {z }
translation deformation rotation

Here, we have employed the local velocity gradient L as well as its symmetric (D) and anti-symmetric
(R) parts:

∂v L + LT L − LT
L= = D + R, D= , R= . (3.336)
∂x 2 2
The symmetry of D allows definition of a real orthonormal basis. For this three-dimensional system,
the dual vector ω of the anti-symmetric R defines the axis of rotation, and its magnitude ω describes
the rotation rate. Now the relative volumetric stretching rate is given by tr L = tr D = div v. And it
is not difficult to show that the linear stretching rate D associated with any direction with unit normal
α is D = αT · D · α.
For our system, we have
 
− x101 0 0
L =  −4x1 x3 −2 − 16 35
+ 2(1 − x21 )  . (3.337)
0 1 1

We see that the relative volumetric expansion rate is

x1
ϑ = tr L = −1 − . (3.338)
10
6
J. D. Mengers, 2012, “Slow invariant manifolds for reaction-diffusion systems,” Ph.D. Dissertation,
University of Notre Dame, Notre Dame, Indiana.

© 05 December 2025, Joseph M. Powers. All rights reserved.

112 CHAPTER 3. KINEMATICS

Because on the heteroclinic trajectory x1 ∈ [−1, 1], we always have a locally shrinking volume on that
trajectory. Now by inspection the unit tangent vector to the heteroclinic trajectory is αt = (1, 0, 0)T .
So the tangential stretching rate on the heteroclinic trajectory is
x1
Dt = αTt · L · αt = − . (3.339)
10
So near the saddle we have Dt = 1/10, and near the sink we have Dt = −1/10. Now we are concerned
with stretching in directions normal to the heteroclinic trajectory. Certainly two unit normal vectors
are αn1 = (0, 1, 0)T and αn2 = (0, 0, 1)T . But there are also infinitely many other unit normals. A
detailed optimization calculation reveals however that if we 1) form the 3 × 2 matrix Qn with αn1 and
αn2 in its columns:
 
0 0
Qn =  1 0  , (3.340)
0 1

and 2) form the 2×2 matrices Dn and Rn associated with the plane normal to the heteroclinic trajectory

Dn = QTn · D · Qn , Rn = QTn · R · Qn , (3.341)

that a) the eigenvalues of Dn give the extreme values of the normal stretching rates Dn1 and Dn2 ,
and the normalized eigenvectors give the associated directions for extreme normal stretching and b)
the magnitude of extremal rotation in the hyperplane normal to αt is given by ω = ||Rn ||2 . On the
heteroclinic trajectory, we find
 x 
− 101 0 0
D= 0 −2 19
− 32 + 1 − x21  . (3.342)
19 2
0 − 32 + 1 − x1 1

The reduced deformation tensor associated with motion in the normal plane is
 
−2 − 19 2
32 + 1 − x1
Dn = QTn · D · Qn = 19 . (3.343)
− 32 + 1 − x21 1

Its eigenvalues give the extremal normal stretching rates that are
p
1 2473 − 832x21 + 1024x41
Dn,1,2 = − ± . (3.344)
2 32
For x1 ∈ [−1, 1], we have Dn,1 ≈ 1 and Dn,2 ≈ −2. Because of the presence of a positive normal
stretching rate, one cannot guarantee trajectories are attracted to the heteroclinic trajectory, even though
volume of nearby points is shrinking. Positive normal stretching does not guarantee divergence from the
heteroclinic trajectory; it permits it. Rotation can orient a collection of nearby points into regions where
there is either positive or negative normal stretching. There are two possibilities for the heteroclinic
trajectory to be attracting: either 1) all normal stretching rates are negative, or 2) the rotation rate is
sufficiently fast and the overall system is volume-decreasing,7 so that the integrated effect is relaxation
to the heteroclinic trajectory. For the heteroclinic trajectory to have the additional property of being
restricted to the slow dynamics, we must additionally require that the smallest normal stretching rate
be larger than the tangential stretching rate.
We illustrate these notions in the sketch of Fig. 3.18. Here we imagine a sphere of points as initial
7
Such systems have ∇T · v < 0. In the dynamic systems literature, this is known as a dissipative
system; however, in fluid mechanics we reserve the word “dissipative” for systems that have thermodynamic
irreversibilities.

© 05 December 2025, Joseph M. Powers. All rights reserved.

3.13. KINEMATICS AS A DYNAMICAL SYSTEM 113

sink

saddle

Figure 3.18: Sketch of a phase volume with ∇T · v < 0 showing heteroclinic connection
between a saddle and sink equilibria along with the evolution of a set of points initially
configured as a sphere as they move into regions with some positive normal stretching rates.

conditions near the saddle. We imagine that the system is such that the volume shrinks as the sphere
moves. While the overall volume shrinks, one of the normal stretching rates is positive, admitting
divergence of nearby trajectories from the heteroclinic trajectory. Rotation orients the volume into a
region where negative normal stretching brings all points ultimately to the sink.
For our system, families of trajectories are shown in Fig. 3.19a, and it is seen that there is divergence
from the heteroclinic trajectory. This must be attributed to some points experiencing positive normal

0 1
-1
-1
-1

equilibrium
saddle with one unstable
mode
0
0
sink

sink

1 1
1
-1
-1
0 0
0
a) 1 -1
1 b)

Figure 3.19: Plots of trajectories near the heteroclinic connection between equilibria with
one unstable mode and a sink illustrating a) divergence of nearby trajectories due to positive
normal stretching with insufficiently rapid rotation and b) convergence of nearby trajectories
in the presence of positive normal stretching with sufficiently rapid rotation.

stretching away from the heteroclinic trajectory. For this case, the rotation rate is ω = −51/32 + 1 − x21.
Thus, the local rotation has a magnitude of near unity near the heteroclinic orbit, and the time scales
of rotation are close to the time scales of normal stretching.
We can modify the system to include more rotation. For instance, replacing Eq. (3.333) by dx3 /dt =
10x2 +x3 introduces a sufficient amount of rotation to render the heteroclinic trajectory to be attractive

© 05 December 2025, Joseph M. Powers. All rights reserved.

114 CHAPTER 3. KINEMATICS

to nearby trajectories. Detailed analysis reveals that this small change 1) does not change the location
of the two equilibria, 2) does not change the heteroclinic trajectory connecting the two equilibria, 3)
modifies the dynamics near each equilibrium such that both have two stable oscillatory modes, with
the equilibrium at (−1, 0, 0)T also containing a third unstable mode and that at (1, 0, 0)T containing
a third stable mode, 4) does not change that the system has a negative volumetric stretch rate on
the heteroclinic trajectory, 5) does not change that a positive normal stretching mode exists on the
heteroclinic trajectory, and 6) enhances the rotation such that the heteroclinic trajectory is locally
attractive. This is illustrated in Fig. 3.19b.
Had the local velocity gradient been purely symmetric, interpretation would be much easier. It is
the effect of a non-zero anti-symmetric part of L that induces the geometrical complexities of rotation.
Such systems are often known as non-normal dynamical systems.

3.13.2 Two-dimensional limit

Here let us study the mathematics of general two-dimensional linear systems. They can represent the
kinematics of a fluid particle. Let us study two-dimensional homogeneous linear systems of the form
dx
= A · x, x(0) = x0 , (3.345)
dt
with
   
x cos θ
x= , x0 = , θ ∈ [0, 2π], (3.346)
y sin θ

for a variety of constant 2 × 2 matrices A. We consider a family of initial conditions that lie on the
unit circle. We study how the geometry of this unit circle evolves with time. We will omit details of
analysis, but demonstrate a few rules of thumb:
• If tr A = 0, the area will not grow or decay.
• If tr A > 0, the area will grow.
• If tr A < 0, the area will decay.
• If A is diagonal with equal positive eigenvalues, the circle will grow uniformly.
• If A is diagonal with equal negative eigenvalues, the circle will decay uniformly.
• If A is symmetric with unequal eigenvalues, the circle will evolve into an ellipse whose major and
minor axes are associated with the eigenvectors of A, and whose major and minor axes grow or decay
as dictated by the eigenvalues of A.
• The anti-symmetric part of A is associated with rotation about the origin. A purely anti-symmetric
A will cause points on the unit circle to rotate about the origin, but stay on the unit circle.
• If the real parts of the eigenvalues of A are negative, all trajectories will approach the origin (x, y)T = 0
as t → ∞; the circle reduces to a point.
We present some behaviors of the evolution of a family of initial conditions that lie on the unit
circle for a variety of matrices A in Fig. 3.20. The upper left plot has no evolution and so yields the
circle of initial conditions. Eight specific points on that unit circle are marked. All plots are at t = 1/2.
The behavior as t → ∞ is shown in Fig. 3.21

© 05 December 2025, Joseph M. Powers. All rights reserved.

3.13. KINEMATICS AS A DYNAMICAL SYSTEM 115

y y y
2 2 2

no isotropic
motion 1 expansion 1 isotropic 1
compression

x x x
-2 -1 1 2 -2 -1 1 2 -2 -1 1 2

-1 -1 -1

-2 -2 -2
y y y
2 2 2

area-preserving
anisotropic area-preserving area-preserving
1 anisotropic 1 anisotropic 1
deformation; no
rotation deformation; no deformation; no
rotation rotation
x x x
-2 -1 1 2 -2 -1 1 2 -2 -1 1 2

-1 -1 -1

-2 -2 -2

y y y
2 2 2

1 1 1

x x x
-2 -1 1 2 -2 -1 1 2 -2 -1 1 2

area-preserving
anisotropic -1 area-preserving -1
area-preserving - 1
deformation; no counter-clockwise
clockwise rotation;
rotation rotation; no
no deformation
deformation
-2 -2 -2

y y y
2 2 2

1 1 1

x x x
-2 -1 1 2 -2 -1 1 2 -2 -1 1 2

area-preserving
-1 -1 -1
anisotropic
anisotropic anisotropic deformation with
expansion; compression; counter-clockwise
-2 no rotation -2 no rotation -2 rotation

Figure 3.20: Evolution of a unit circle of initial conditions for the system dx/dt = A · x for
many different values of A. All plots are for t = 1/2.

© 05 December 2025, Joseph M. Powers. All rights reserved.

116 CHAPTER 3. KINEMATICS

y y y
2 2 2

no isotropic
motion 1 expansion 1 isotropic 1
compression

x x x
-2 -1 1 2 -2 -1 1 2 -2 -1 1 2

-1 -1 -1

-2 -2 -2
y y y
2 2 2

area-preserving
anisotropic area-preserving area-preserving
1 anisotropic 1 anisotropic 1
deformation; no
rotation deformation; no deformation; no
rotation rotation
x x x
-2 -1 1 2 -2 -1 1 2 -2 -1 1 2

-1 -1 -1

-2 -2 -2

y y y
2 2 2

1 1 1

x x x
-2 -1 1 2 -2 -1 1 2 -2 -1 1 2

area-preserving
anisotropic -1 area-preserving -1
area-preserving - 1
deformation; no counter-clockwise
clockwise rotation;
rotation rotation; no
no deformation
deformation
-2 -2 -2

y y y
2 2 2

1 1 1

x x x
-2 -1 1 2 -2 -1 1 2 -2 -1 1 2

area-preserving
-1 -1 -1
anisotropic
anisotropic anisotropic deformation with
expansion; compression; counter-clockwise
-2 no rotation -2 no rotation -2 rotation

Figure 3.21: Evolution of a unit circle of initial conditions for the system dx/dt = A · x for
many different values of A. All plots are for t → ∞.

© 05 December 2025, Joseph M. Powers. All rights reserved.

3.13. KINEMATICS AS A DYNAMICAL SYSTEM 117

The concept of “area-preserving” has seemingly contrasting results for linear algebraic operators
and linear differential systems of equations. For 2 × 2 matrices, it is area that should be considered. If
we were to consider 3 × 3 matrices, we would consider volume instead. Consider a constant matrix 2 × 2
matrix A. A linear algebraic operation takes x to a new x via x = A · x. But the linear differential
operator gives us a continuous evolution of x via dx/dt = A · x. But the previous example considered
an area-preserving linear differential system. We can ask the question for either algebraic or differential
equations of whether or not the area of a unit circle is preserved after advancement of the system via
the linear operation. The result is, in short, that for two-dimensional linear algebraic systems, the area
is preserved if | det A| = 1. And the orientation of the area after linear algebraic transformation is
preserved if det A = 1. However, for linear differential systems, the area is preserved if tr A = 0.

Example 3.8
Examine the effect of a linear algebraic operator and linear differential operator on a unit circle if
 
1 0
A= . (3.347)
0 1

The unit circle is parameterized by

x = cos θ, y = sin θ, θ = [0, 2π]. (3.348)

We have

det A = 1, tr A = 2. (3.349)

For the algebraic operator, the area and orientation are preserved because det A = 1. The mapping
x = A · x is simply the identity mapping x = I · x = x:
         
x 1 0 x 1 0 cos θ cos θ
= = = . (3.350)
y 0 1 y 0 1 sin θ sin θ

The unit circle transforms to the unit circle, and the points that comprise the unit circle retain their
original positions. The magnitude of a vector x is preserved, and it does not rotate.
For the differential operator, the area increases because tr A > 0. There is no rotation because A
is symmetric. The solution is
   t 
x e cos θ
= , (3.351)
y et sin θ
so

x2 + y 2 = e2t . (3.352)

The radius of the circle is growing, and is given by r = et . The area of the circle is growing and is given
by A = πe2t . Relevant plots are given in Fig. 3.22.
If we consider a simple finite difference approximation, the advance through one time step can be
approximated by an algebraic transformation via
xn+1 − xn
= A · xn , (3.353)
∆t
xn+1 = xn + ∆tA · xn , (3.354)
xn+1 = (I + ∆tA) ·xn . (3.355)
| {z }
σ

© 05 December 2025, Joseph M. Powers. All rights reserved.

118 CHAPTER 3. KINEMATICS

y algebraic y y
2 differential
2
2 2

1 1 1 1

x x x
-2 -1 1 2 -2 -1 1 2 -2 -1 1 2 -2 -1 1 2

-1 -1 -1 -1

-2 -2 -2 -2

Figure 3.22: Mapping of a unit circle x = A · x and evolution of a unit circle of initial
conditions for dx/dt = A · x, t = 1/2; det A = 1, tr A = 2.

Here the relevant matrix is not A but the matrix σ:

σ = I + ∆tA. (3.356)

Thus the algebraic transformation through one time step is

xn+1 = σ · xn . (3.357)

And for our A, this yields

 
1 + ∆t 0
σ= . (3.358)
0 1 + ∆t

We see then that

det σ = (1 + ∆t)2 . (3.359)

Thus the algebraic approximation to the differential map confirms it is area-increasing, consistent with
tr A > 0.

Example 3.9
Examine the effect of a linear algebraic operator and linear differential operator on a unit circle if
 
1 0
A= . (3.360)
0 −1

We have

det A = −1, tr A = 0. (3.361)

© 05 December 2025, Joseph M. Powers. All rights reserved.

3.13. KINEMATICS AS A DYNAMICAL SYSTEM 119

For the algebraic operator, the area is preserved and the orientation is reversed because det A = −1.
The mapping is a reflection mapping about the x axis:
         
x 1 0 x 1 0 cos θ cos θ
= = = . (3.362)
y 0 −1 y 0 −1 sin θ − sin θ
The unit circle transforms to the unit circle, and the points that comprise the unit circle reflect about
the x axis. The magnitude of a vector x is preserved.
For the differential operator, the area is preserved because tr A = 0. There is no rotation because
A is symmetric. But there is anisotropic deformation of the circle into an ellipse, with compression
along the y axis and extension along the x axis. The x axis is aligned with the major axis of the ellipse,
and the y axis is aligned with the minor axis. The solution is
   t 
x e cos θ
= , (3.363)
y e−t sin θ
that can be recast as an ellipse
 x 2  y 2
+ = 1. (3.364)
et e−t
The semi-major and semi-minor
p axes have lengths a = et and b = e−t . With the eccentricity e of the
ellipse defined as e = 1 − b2 /a2 , we have
r
e−2t p
e = 1 − 2t = 1 − e−4t . (3.365)
e
The eccentricity is initially zero and is nearly unity when t = 1/2. Thus the circle is deforming into a
long narrow ellipse confined to the region near the x axis. The area is A = πab = πet e−t = π and is
constant. Relevant plots are given in Fig. 3.23.

y y y
2 2 differential
algebraic 2 2

1 1 1 1

x x x
-2 -1 1 2 -2 -1 1 2 -2 -1 1 2 -2 -1 1 2

-1 -1 -1 -1

-2 -2 -2 -2

Figure 3.23: Mapping of a unit circle x = A · x and evolution of a unit circle of initial
conditions for dx/dt = A · x, t = 1/2; det A = −1, tr A = 0.

Example 3.10
Examine the effect of a linear algebraic operator and linear differential operator on a unit circle if
 
1 0
A= . (3.366)
0 0

© 05 December 2025, Joseph M. Powers. All rights reserved.

120 CHAPTER 3. KINEMATICS

We have

det A = 0, tr A = 1. (3.367)

For the algebraic operator, the area is collapsed to zero because det A = 0. Because the rank of A is
1, the matrix maps two-dimensional geometries onto one-dimensional geometries. Here, the mapping
takes the entire unit circle onto a line segment on the x axis, x ∈ [−1, 1]:
         
x 1 0 x 1 0 cos θ cos θ
= = = . (3.368)
y 0 0 y 0 0 sin θ 0

Except for θ = 0, π, the magnitude of a vector x on the unit circle is attenuated by the transformation.
For the differential operator, the area increases because tr A = 1. There is no rotation because A
is symmetric. But there is anisotropic deformation of the circle into an ellipse, with no change along
the y axis and extension along the x axis. The x axis is aligned with the major axis of the ellipse, and
the y axis is aligned with the minor axis. The solution is
   t 
x e cos θ
= , (3.369)
y sin θ

that can be recast as an ellipse

 x 2
+ y 2 = 1. (3.370)
et
The semi-major and semi-minor axes have lengths a = et and b = 1. The eccentricity is
r
1 p
e = 1 − 2t = 1 − e−2t . (3.371)
e

The eccentricity is initially zero and is nearly unity when t = 1. Thus the circle is deforming into a
long narrow ellipse confined to the region near the x axis. The area is A = πab = πet and is increasing.
Relevant plots are given in Fig. 3.24.

y y y
2 algebraic 2 2 differential 2

1 1 1 1

x x x
-2 -1 1 2 -2 -1 1 2 -2 -1 1 2 -2 -1 1 2

-1 -1 -1 -1

-2 -2 -2 -2

Figure 3.24: Mapping of a unit circle x = A · x and evolution of a unit circle of initial
conditions for dx/dt = A · x, t = 1/2; det A = 0, tr A = 1.

© 05 December 2025, Joseph M. Powers. All rights reserved.

3.13. KINEMATICS AS A DYNAMICAL SYSTEM 121

Example 3.11
Examine the effect of a linear algebraic operator and linear differential operator on a unit circle if
 
0 0
A= . (3.372)
0 0

We have

det A = 0, tr A = 0. (3.373)

For the algebraic operator, the area is collapsed to zero because det A = 0. Because the rank of A is
0, the matrix maps two-dimensional geometries onto a zero-dimensional geometry. Here, the mapping
takes the entire unit circle onto the the point (0, 0):
         
x 0 0 x 0 0 cos θ 0
= = = . (3.374)
y 0 0 y 0 0 sin θ 0
For the differential operator, the area remains constant because tr A = 0. There is no rotation
because A is symmetric. The time-independent solution is
   
x cos θ
= , (3.375)
y sin θ
that can be recast as a circle

x2 + y 2 = 1. (3.376)

The area is and remains A = π. Relevant plots are given in Fig. 3.25.

y y y
2 2
2 2
differential
algebraic

1 1 1 1

x x x
-2 -1 1 2 -2 -1 1 2 -2 -1 1 2 -2 -1 1 2

-1 -1 -1 -1

-2 -2 -2 -2

Figure 3.25: Mapping of a unit circle x = A · x and evolution of a unit circle of initial
conditions for dx/dt = A · x, t = 1/2; det A = 0, tr A = 0.

Example 3.12
Examine the effect of a linear algebraic operator and linear differential operator on a unit circle if
 
1 −1
A= . (3.377)
0 1

© 05 December 2025, Joseph M. Powers. All rights reserved.

122 CHAPTER 3. KINEMATICS

We have

det A = 1, tr A = 2. (3.378)

For the algebraic operator, the area and orientation are preserved and because det A = 1. Because A
has asymmetry, there is rotation. And there is area-conserving deformation of the circle into an ellipse.
T
The major and minor axes of the ellipse are aligned with the eigenvectors of A−1 ·A−1 . And the length
T
of these axes is associated with the square root of the reciprocals of the eigenvalues of A−1 · A−1 . The
algebraic mapping is
         
x 1 −1 x 1 −1 cos θ cos θ − sin θ
= = = . (3.379)
y 0 1 y 0 1 sin θ sin θ

We can use the singular value decomposition (SVD) of the matrix A to better interpret our results.
Any matrix A can be decomposed as

A = Q2 · Σ · QT1 . (3.380)

Here Σ is a diagonal square matrix with real positive semi-definite numbers on its diagonal. Of these
r are nonzero, where r is the rank of σ. The columns of Q2 are the eigenvectors of A · AT . The
columns of Q1 are the eigenvectors of AT · A. Both Q1 and Q2 are square orthogonal matrices, and
they are rotation or reflection matrices. The action of rotation or reflection matrices on vectors is to
rotate or reflect them and leave their magnitude unchanged. Detailed analysis of the singular value
decomposition of A shows its action on a vector is composed of a rotation, a non-uniform stretching,
and a second rotation. The decomposition is
     
1 −1 −0.850651 −0.525731 1.61803 0 −0.525731 0.850651
= . (3.381)
0 1 0.525731 −0.850651 0 0.618034 −0.850651 −0.525731
| {z } | {z }| {z }| {z }
A Q2 Σ QT
1

Detailed analysis of the algebraic system shows the unit circle of initial conditions transforms to
an ellipse described by
 2  2
0.525731x + 0.850651y −0.850651x + 0.525731y
+ = 1. (3.382)
0.618034 1.61803

The semi-major and semi-minor axes have lengths

s s
2 2
a= √ = 1.61803, b= √ = 0.618034. (3.383)
3− 5 3+ 5

The eccentricity is
s √
3− 5
e= 1− √ = 0.924176. (3.384)
3+ 5

The area of the ellipse is

s s
2 2
A = πab = π √ √ = π. (3.385)
3− 5 3+ 5

© 05 December 2025, Joseph M. Powers. All rights reserved.

3.13. KINEMATICS AS A DYNAMICAL SYSTEM 123

y y y
2 2 differential
algebraic 2 2

1 1 1 1

x x x
-2 -1 1 2 -2 -1 1 2 -2 -1 1 2 -2 -1 1 2

-1 -1 -1 -1

-2 -2 -2 -2

Figure 3.26: Mapping of a unit circle x = A · x and evolution of a unit circle of initial
conditions for dx/dt = A · x, t = 1/2; det A = 1, tr A = 2.

This is identical to the area of the unit circle. The area is preserved for the algebraic mapping.
For the differential operator, the area increases because tr A = 2. There is rotation because A is
asymmetric. And there is anisotropic deformation. Relevant plots are given in Fig. 3.26. With the set
of initial conditions given by x(0) = x0 = cos θ, y(0) = y0 = sin θ, θ = [0, 2π], detailed analysis of the
differential system shows the solution to be given parametrically as
   t    t  
x e −tet x0 e −tet cos θ
= = , θ ∈ [0, 2π]. (3.386)
y 0 et y0 0 et sin θ

So

x = et cos θ − tet sin θ, y = et sin θ. (3.387)

Note
 
et −tet
det = e2t . (3.388)
0 et

Specializing for t = 1/2, we see the ratio of the new area to the original area should be e1 = 2.71828.
For t = 1/2, the unit circle of initial conditions has evolved to an ellipse described by
 2  2
−0.615412x − 0.788205y −0.788205x + 0.615412y
+ = 1. (3.389)
1.28728 2.11164

The semi-major and semi-minor axes grow with time and at t = 1/2 have lengths
s s
2e 2e
a=2 √ = 2.11164, b=2 √ = 1.28728. (3.390)
9 − 17 9 + 17

The eccentricity at t = 1/2 is

s √
9 − 17
e= 1− √ = 0.792700. (3.391)
9 + 17

The area of the ellipse at t = 1/2 then is

A = πab = π(2.11164)(1.28728) = π(2.71828) = πe. (3.392)

The original area of the unit circle is π, so the ratio of the new to the original is e, as predicted.

© 05 December 2025, Joseph M. Powers. All rights reserved.

124 CHAPTER 3. KINEMATICS

Further analysis gives the time dependencies, that are

√ t √ t s √
2e 2e t2 − t2 + 4t + 2
a= p √ , b = p √ , e = 1 − √ , A = πe2t . (3.393)
t2 − t2 + 4t + 2 t2 + t2 + 4t + 2 t2 + t2 + 4t + 2

And the inclination of the major axis to the horizontal starts at −π/4 and as time progresses, goes to
zero via
 
−2
θ = tan−1 √ . (3.394)
t + t2 + 4
As time progresses, the circle first evolves into a weakly eccentric ellipse whose major axis is inclined
at −π/4 to the horizontal and whose minor axis is inclined at π/4 to the horizontal. As time grows,
the area of the ellipse grows, its major and minor axes rotate counterclockwise, and it becomes more
eccentric. At long time, the major axis is horizontal, and the minor axis is vertical.
We can draw upon the presentation of Powers and Sen (2015, Sec. 7.9.5) to better understand the
transformation of the unit circle in terms of eigenvalues and eigenvectors. Considering our algebraic
operator from x to x, we get

x = A · x, (3.395)
−1
A ·x = x, (3.396)
−1 T −1 −1 T
A ·A ·x = A ·x (3.397)
T
Because A−1 · A−1 is symmetric, it has the diagonal decomposition
T
A−1 · A−1 = Q · Λ · QT . (3.398)
T
Here Λ is diagonal with its diagonal populated by the eigenvalues of A−1 ·A−1 , and Q is an orthogonal
T
matrix whose columns are populated by the normalized eigenvectors of A−1 · A−1 . We then get
T
Q · Λ · QT · x = A−1 · x, (3.399)
T T T −1 T
x ·Q·Λ·Q ·x = x ·A · x, (3.400)
(QT · x)T · Λ · (QT · x) = (A−1 · x)T · x, (3.401)
(QT · x)T · Λ · (QT · x) = xT · x. (3.402)

Now for the unit circle xT · x = 1, so we have

(QT · x)T · Λ · (QT · x) = 1. (3.403)

This is the equation of an ellipse in standard form.

As an aside, we note that linear transformations of the type x = A · x are known as linear
homogeneous transformations. The origin x = 0 maps to the origin in the transformed space
x = 0. In contrast, the linear transformation x = A · x + b is known as a linear affine
transformation. The origin x = 0 maps to x = b.

© 05 December 2025, Joseph M. Powers. All rights reserved.

Chapter 4

Evolution axioms

see Powers, Chapter 4,

see Panton, Chapter 4,
see Hughes and Gaylord, Chapter 1,
see Yih, Chapters 1 and 2,
see Whitaker, Chapters 4 and 5,
see Aris, Chapters 5 and 6.

A fundamental goal of this chapter is to convert the verbal notions that embody the basic
axioms of non-relativistic continuum mechanics into usable mathematical expressions. First,
we must list those axioms. The axioms themselves are simply principles that have been
observed to have wide validity as long as the particle velocity is small relative to the speed
of light and length scales are sufficiently large to contain many molecules. Many of these
axioms can be applied to molecules as well. The axioms cannot be proven. They are simply
statements that have been useful in describing the universe.
A summary of the axioms in words is as follows:

• Mass evolution principle: The time rate of change of mass of a material region is zero.

• Linear momenta evolution principle: The time rate of change of the linear momenta
of a material region is equal to the sum of forces acting on the region. This is Euler’s
generalization of Newton’s second law of motion.

• Angular momenta evolution principle: The time rate of change of the angular momenta
of a material region is equal to the sum of the torques acting on the region. This was
first formulated by Euler.

• Energy evolution principle: The time rate of change of energy within a material region
is equal to the rate that energy is received by heat and work interactions. This is the
first law of thermodynamics.

125
126 CHAPTER 4. EVOLUTION AXIOMS

• Entropy evolution inequality: The time rate of change of entropy within a material
region is greater than or equal to the ratio of the rate of heat transferred to the region
and the absolute temperature of the region. This is the second law of thermodynamics.

Some secondary concepts related to these axioms are as follows:

• The local stress on one side of a surface is identically opposite that stress on the
opposite side.

• Stress can be separated into thermodynamic and viscous stress.

• Forces can be separated into surface and body forces.

• In the absence of body couples, the angular momenta principle reduces to a nearly
trivial statement.

• The energy equation can be separated into mechanical and thermal components. The
mechanical energy is associated with ordered kinetic energy at the macroscale, and the
thermal energy is associated with random kinetic energy of molecular motion at the
microscale.

Next we shall systematically convert these words into mathematical form.

4.1 Mass
The mass evolution axiom is simple to state mathematically. It is

d

mVm (t) = 0. (4.1)
dt
Thus mass of of a fluid particle is conserved; it does not evolve. As introduced in Ch. [Link].1,
Vm (t) stands for a material region that can evolve in time, and mVm (t) is the mass in the
material region. A relevant material region is sketched in Fig. 4.1. We can define the mass
of the material region based upon the local value of density:
Z
mVm (t) = ρ dV. (4.2)
Vm (t)

So, the mass conservation axiom is

Z
d
ρ dV = 0. (4.3)
dt Vm (t)

R R R
Recalling Leibniz’s rule, Eq. (2.270), dtd Va (t) [ ] dV = Va (t) ∂o [ ] dV + Aa (t) ni wi [ ] dA, we
specialize the arbitrary velocity to the fluid velocity so that wi = vi . This is because we are

© 05 December 2025, Joseph M. Powers. All rights reserved.

4.1. MASS 127

dA

Figure 4.1: Sketch of finite material region Vm , infinitesimal mass element ρ dV , and
infinitesimal surface element dA with unit normal ni , and general velocity wi equal to fluid
velocity vi .

considering a material region, and thus the Reynolds transport theorem, Eq. (2.273). So we
get Z Z Z
d
ρ dV = ∂o ρ dV + ni vi ρ dA = 0. (4.4)
dt Vm (t) Vm (t) Am (t)

Now, Eq. (4.4) is in fact the most fundamental representation of the mass conservation prin-
ciple. It applies for both continuous flows as well as for flows with embedded discontinuities
such as the shock waves we will study in Ch. 9.4.1.
For this chapter, we will assume that there are no embedded R discontinuities,
R and proceed
forward. Now we invoke Gauss’s theorem, Eq. (2.252) Vm (t) ∂i [ ] dV = Am (t) ni [ ] dA, to
convert a surface integral to a volume integral to get the mass conservation axiom to read
as
Z Z
∂o ρ dV + ∂i (ρvi ) dV = 0, (4.5)
Vm (t) Vm (t)
Z
(∂o ρ + ∂i (ρvi )) dV = 0. (4.6)
Vm (t)

Now, in an important step, we realize that the only way for this integral, that has arbitrary
limits of integration, to always be zero, is for the integrand itself to always be zero. Hence,
we have
∂o ρ + ∂i (ρvi ) = 0. (4.7)
This step requires all state variables be continuous, and so cannot be done if discontinuities,
such as shock waves, are embedded within Vm (t), as will be discussed in Ch. 9.4.1. We write
this in expanded Cartesian and Gibbs notation as
∂o ρ + ∂1 (ρv1 ) + ∂2 (ρv2 ) + ∂3 (ρv3 ) = 0, (4.8)
∂ρ
+ ∇T · (ρv) = 0. (4.9)
∂t

© 05 December 2025, Joseph M. Powers. All rights reserved.

128 CHAPTER 4. EVOLUTION AXIOMS

These equations, along with Eq. (4.7), are all in what is known as conservative or divergence
form. The conservative form shows mass (equivalently ρ) is conserved when mass fluxes, ρv,
are in balance. There are several alternative forms for this axiom. Using the product rule,
we can say also
∂o ρ + vi ∂i ρ +ρ∂i vi = 0, (4.10)
| {z }
material derivative of density

or, writing in what is called the non-conservative form,

dρ
+ ρ∂i vi = 0, (4.11)
dt
dρ
+ ρ∇T · v = 0, (4.12)
dt
(∂o ρ + v1 ∂1 ρ + v2 ∂2 ρ + v3 ∂3 ρ) + ρ (∂1 v1 + ∂2 v2 + ∂3 v3 ) = 0. (4.13)

For flows with no embedded discontinuities, the conservative and non-conservative forms
give identical information. So, we can also say
1 dρ
=− ∂i vi . (4.14)
ρ dt |{z}
|{z} relative rate of particle volume expansion
relative rate of density increase

Recalling Eq. (3.187), we see the relative rate of density increase of a fluid particle is the
negative of its relative rate of expansion, as expected. So, we also have
1 dρ 1 dVm
= − , (4.15)
ρ dt Vm dt
dVm dρ
ρ + Vm = 0, (4.16)
dt dt
d
(ρVm ) = 0, (4.17)
dt
d
(mVm ) = 0. (4.18)
dt
This returns us to our original mass conservation statement, Eq. (4.1). We note that in a
relativistic system, in which mass-energy is conserved, but not mass, that we can have a
material region, that is a region bounded by a surface across which there is no flux of mass,
for which the mass can indeed change, thus violating our non-relativistic mass conservation
axiom.
Let us consider a special case of the Reynolds transport theorem, Eq. (2.273) for a fluid
that obeys mass conservation. The general tensor in Eq. (2.273) can be recast as

Tjk.... = ρTjk.... (4.19)

This is useful when Tjk... as some intensive property that has units of some quantity per
unit mass. Then Tjl... is the same quantity per unit volume. Then the Reynolds transport

© 05 December 2025, Joseph M. Powers. All rights reserved.

4.2. LINEAR MOMENTA 129

dA

Figure 4.2: Sketch of finite material region Vm , infinitesimal linear momenta element ρvi dV ,
infinitesimal body force element ρfi dV , and infinitesimal surface element dA with unit
normal ni , surface traction ti and general velocity wi equal to fluid velocity vi .

theorem becomes
Z Z Z
d
ρTjk....(xi , t) dV = ∂o (ρTjk....) dV + nl ρvl Tjk.... dA, (4.20)
dt Vm (t) Vm (t) Am (t)
Z
= (∂o (ρTjk....) + ∂l (ρvl Tjk....)) dV, (4.21)
Vm (t)
    
Z
  
= Tjk... ∂o ρ + ∂l (ρvl ) + ρ ∂o Tjk.... + vl ∂l Tjk.... dV,
Vm (t) | {z } | {z }
=0 =dTjk... /dt

(4.22)
Z
dTjk...
= ρ dV. (4.23)
Vm (t) dt

4.2 Linear momenta

4.2.1 Statement of the principle
The linear momenta evolution axiom is simple to state mathematically. It is
Z Z Z
d
ρvi dV = ρfi dV + ti dA . (4.24)
dt Vm (t) Vm (t) Am (t)
| {z } | {z } | {z }
rate of change of linear momenta body forces surface forces

Again Vm (t) stands for a material region that can evolve in time. A relevant material region
is sketched in Fig. 4.2. The term fi represents a body force per unit mass. An example of
such a force would be the gravitational force acting on a body, that when scaled by mass,

© 05 December 2025, Joseph M. Powers. All rights reserved.

130 CHAPTER 4. EVOLUTION AXIOMS

yields gi . The term ti is a traction, that is a vector representing force per unit area. A major
challenge of this section will be to express the traction vector in terms of what is known as
the stress tensor.
Consider first the left hand side, LHS, of the linear momenta principle
Z Z
LHS = ∂o (ρvi ) dV + nj ρvi vj dA, from Reynolds, (4.25)
Vm (t) Am (t)
Z
= (∂o (ρvi ) + ∂j (ρvj vi )) dV, from Gauss. (4.26)
Vm (t)

So, the linear momenta principle is

Z Z Z
(∂o (ρvi ) + ∂j (ρvj vi )) dV = ρfi dV + ti dA. (4.27)
Vm (t) Vm (t) Am (t)

These are all expressed in terms of volume integrals except for the term involving surface
forces.

4.2.2 Surface forces

The surface force per unit area is a vector we call the traction tj . It has the units of stress,
but it is not formally a stress, that is a tensor. The traction is a function of both position
xi and surface orientation nk : tj = tj (xi , nk ).
We intend to demonstrate the following: The traction can be stated in terms of a stress
tensor Tij as written next:
tj = ni Tij ,
tT = nT · T,
t = TT · n. (4.28)
The following excursions are necessary to show this.
• Show force on one side of surface equal and opposite to that on the opposite side
Let us apply the principle of linear momenta to the material region as sketched in
Fig. 4.3. Here
we indicate the dependency of the traction on orientation by notation
such as ti nIIi . This does not indicate multiplication, nor that i is a dummy index
here. In Fig. 4.3, the thin pillbox has width ∆l, circumference s, and a surface area
for the circular region of ∆A. Surface I is a circular region; surface II is the opposite
circular region, and surface III is the cylindrical side.
We apply the mean value theorem to the linear momenta principle for this region and
get
(∂o (ρvi ) + ∂j (ρvj vi ))∗ (∆A)(∆l) = (ρfi )∗ (∆A)(∆l)
+t∗i (nIi )∆A
+t∗i (nII
i )∆A
+ti (nIII
∗
i )s(∆l). (4.29)

© 05 December 2025, Joseph M. Powers. All rights reserved.

4.2. LINEAR MOMENTA 131

Figure 4.3: Sketch of pillbox element for stress analysis.

Now we let ∆l → 0, holding for now s and ∆A fixed to obtain

0 = t∗i (nIi ) + t∗i (nII
i ) ∆A. (4.30)

Now letting ∆A → 0, so that the mean value approaches the local value, and taking
nIi = −nII
i ≡ ni , we get a useful result

ti (ni ) = −ti (−ni ). (4.31)

At an infinitesimal length scale, the traction on one side of a surface is equal and
opposite to that on the other. That is, there is a local force balance. This applies even
if there is velocity and acceleration of the material on a macroscale. On the microscale,
surface forces dominate inertia and body forces. This is a useful general principle to
remember. It the fundamental reason why microorganisms have different propulsion
systems than macroorganisms: they are fighting different forces.
• Study stress on arbitrary plane and relate to stress on coordinate planes
Now let us consider a rectangular parallelepiped aligned with the Cartesian axes that
has been sliced at an oblique angle to form a tetrahedron. We will apply the linear
momenta principle to this geometry and make a statement about the existence of a
stress tensor. The described material region is sketched in Fig. 4.4. Let ∆L be a
characteristic length scale of the tetrahedron. Also let four unit normals nj exist, one
for each surface. They will be −n1 , −n2 , −n3 for the surfaces associated with each
coordinate direction. They are negative because the outer normal points opposite to
the direction of the axes. Let ni be the normal associated with the oblique face. Let
∆A denote the surface area of each face.
Now the volume of the tetrahedron must be of order L3 and the surface area of order
L2 . Thus applying the mean value theorem to the linear momenta principle, we obtain

© 05 December 2025, Joseph M. Powers. All rights reserved.

132 CHAPTER 4. EVOLUTION AXIOMS

Figure 4.4: Sketch of tetrahedral element for stress analysis on an arbitrary plane.

the form

(inertia) × (∆L)3 = (body forces) × (∆L)3 + (surface forces) × (∆L)2 . (4.32)

As before, for small volumes, ∆L → 0, and the linear momenta principle reduces to
X
surface forces = 0. (4.33)

Applying this to the configuration of Fig. 4.4, we get

0 = t∗i (ni )∆A + t∗i (−n1 )∆A1 + t∗i (−n2 )∆A2 + t∗i (−n3 )∆A3 . (4.34)

But we know that tj (nj ) = −tj (−nj ), so

t∗i (ni )∆A = t∗i (n1 )∆A1 + t∗i (n2 )∆A2 + t∗i (n3 )∆A3 . (4.35)

Now it is not a difficult geometry problem to show that ni ∆A = ∆Ai , so we get

t∗i (ni )∆A = n1 t∗i (n1 )∆A + n2 t∗i (n2 )∆A + n3 t∗i (n3 )∆A, (4.36)
t∗i (ni ) = n1 t∗i (n1 ) + n2 t∗i (n2 ) + n3 t∗i (n3 ). (4.37)

Now we can consider terms like ti to obviously be a vector, and the indicator, for
example (n1 ), tells us with which surface the vector is associated. This is precisely
what a tensor does, and in fact we can say

ti (ni ) = n1 T1i + n2 T2i + n3 T3i . (4.38)

© 05 December 2025, Joseph M. Powers. All rights reserved.

4.2. LINEAR MOMENTA 133

In shorthand, we can say the same thing with

ti = nj Tji , or equivalently tj = ni Tij , QED. (4.39)

Here Tij is the component of stress in the j direction associated with the surface whose
normal is in the i direction.

• Consider pressure and the viscous stress tensor

Pressure is a familiar concept from thermodynamics and fluid statics. It is often
tempting and sometimes correct to think of the pressure as the force per unit area
normal to a surface and the force tangential to a surface being somehow related to
frictional forces. We shall see that in general, this view is too simplistic.
First recall from thermodynamics that what we will call p, the thermodynamic pressure,
is for a simple compressible substance a function of at most two intensive thermody-
namic variables, say p = f (ρ, e), where e is the specific internal energy. Also recall
that the thermodynamic pressure must be a normal stress, as thermodynamics con-
siders formally only materials at rest, and viscous stresses are associated with moving
fluids.
To distinguish between thermodynamic stresses and other stresses, let us define the
viscous stress tensor τij as follows

τij = Tij + pδij . (4.40)

Recall that Tij is the total stress tensor. We obviously also have

Tij = −pδij + τij . (4.41)

With this definition, pressure is positive in compression, while Tij and τij are positive
in tension. Let us also define the mechanical pressure, p(m) , as the negative of the
average normal surface stress
1 1
p(m) ≡ − Tii = − (T11 + T22 + T33 ). (4.42)
3 3

The often invoked Stokes’ assumption, that remains a subject of widespread misunder-
standing since it was first made in 1845,1 is often adopted for lack of a good alternative
in answer to a question that will be addressed later in Ch. 5.4.3. It asserts that the
thermodynamic pressure is equal to the mechanical pressure:
1
p = p(m) = − Tii . (4.43)
3
1
Stokes, G. G., 1845, “On the theories of internal friction of fluids in motion,” Transactions of the
Cambridge Philosophical Society, 8: 287-305.

© 05 December 2025, Joseph M. Powers. All rights reserved.

134 CHAPTER 4. EVOLUTION AXIOMS

Presumably a pressure measuring device in a moving flow field would actually measure
the mechanical pressure, and not necessarily the thermodynamic pressure, so it is im-
portant to have this issue clarified for proper reconciliation of theory and measurement.
It will be seen that Stokes’ assumption gives some minor æsthetic pleasure in certain
limits, but it is not well-established, and is more a convenience than a requirement
for most materials. It is the case that various incarnations of more fundamental ki-
netic theory under the assumption of a dilute gas composed of inert hard spheres give
rise to the conclusion that Stokes’ assumption is valid. At moderate densities, these
hard sphere kinetic theory models predict that Stokes’ assumption is invalid. However,
none of the common kinetic theory models is able to predict results from experiments,
that nevertheless also give indication, albeit indirect, that Stokes’ assumption is in-
valid. Kinetic theories and experiments that consider polyatomic molecules, that can
suffer vibrational and rotational effects as well, show further deviation from Stokes’
assumption. It is often plausibly argued that these so-called non-equilibrium effects,
that is molecular vibration and rotation, that are only important in high speed flow
applications in which the flow velocity is on the order of the fluid sound speed, are the
mechanisms that cause Stokes’ assumption to be violated. Because they only are im-
portant in high speed applications, they are difficult to measure, though measurement
of the decay of acoustic waves has provided some data. For liquids, there is little to no
theory, and the limited data indicate that Stokes’ assumption is invalid.
Now contracting Eq. (4.41), we get
Tii = −pδii + τii . (4.44)
Using the fact that δii = 3 and inserting Eq. (4.43) in Eq. (4.44), we find for a fluid
that obeys Stokes’ assumption that
1
Tii = Tii (3) +τii , (4.45)
3
|{z} |{z}
δii
−p
0 = τii . (4.46)
That is to say, the trace of the viscous stress tensor is zero. Moreover, for a fluid that
obeys Stokes’ assumption, we can interpret the viscous stress as the deviation from the
mean stress; that is, the viscous stress is a deviatoric stress:
1
Tij = Tkk δij + τij , valid only if Stokes’ assumption holds.
|{z} 3
| {z } |{z}
total stress deviatoric stress
mean stress
(4.47)
If Stokes’ assumption does not hold, then a portion of τij will also contribute to the
mean stress; that is, the viscous stress is not then entirely deviatoric.
Finally, let us note what the traction vector is when the fluid is static. For a static
fluid, there is no viscous stress, so τij = 0, and we have
Tij = −pδij , static fluid. (4.48)

© 05 December 2025, Joseph M. Powers. All rights reserved.

4.2. LINEAR MOMENTA 135

We get the traction vector for a static fluid on any surface with normal ni by

tj = ni Tij = −pni δij = −pnj , static fluid. (4.49)

Changing indices, we see ti = −pni , that is the traction vector must be oriented in the
same direction as the surface normal for a static fluid; all stresses are normal to any
arbitrarily oriented surface.

4.2.3 Final form of linear momenta equation

We are now prepared to write the linear momenta equation in final form. Substituting our
expression for the traction vector, Eq. (4.39) into the linear momenta expression, Eq. (4.27),
we get Z Z Z
(∂o (ρvi ) + ∂j (ρvj vi )) dV = ρfi dV + nj Tji dA. (4.50)
Vm (t) Vm (t) Am (t)

Using Gauss’s theorem, Eq. (2.252), to convert the surface integral into a volume integral,
and combining all under one integral sign, we get
Z
(∂o (ρvi ) + ∂j (ρvj vi ) − ρfi − ∂j Tji ) dV = 0. (4.51)
Vm (t)

Making the same argument as before regarding arbitrary material volumes, this must then
require that the integrand be zero (we actually must require all variables be continuous to
make this work), so we obtain

∂o (ρvi ) + ∂j (ρvj vi ) − ρfi − ∂j Tji = 0. (4.52)

Using then Tij = −pδij + τij , we get in Cartesian index, Gibbs2 , and full notation
∂o (ρvi ) + ∂j (ρvj vi ) = ρfi − ∂i p + ∂j τji , (4.53)
∂
T
T
(ρv) + ∇T · (ρvvT ) = ρf − ∇p + ∇T · τ , (4.54)
∂t
∂o (ρv1 ) + ∂1 (ρv1 v1 ) + ∂2 (ρv2 v1 ) + ∂3 (ρv3 v1 ) = ρf1 − ∂1 p + ∂1 τ11 + ∂2 τ21 + ∂3 τ31 , (4.55)
∂o (ρv2 ) + ∂1 (ρv1 v2 ) + ∂2 (ρv2 v2 ) + ∂3 (ρv3 v2 ) = ρf2 − ∂2 p + ∂1 τ12 + ∂2 τ22 + ∂3 τ32 , (4.56)
∂o (ρv3 ) + ∂1 (ρv1 v3 ) + ∂2 (ρv2 v3 ) + ∂3 (ρv3 v3 ) = ρf3 − ∂3 p + ∂1 τ13 + ∂2 τ23 + ∂3 τ33 . (4.57)

The form is known as the linear momenta principle cast in conservative or divergence form.
It is the first choice of forms for many numerical simulations, as discretizations of this form
of the equation naturally preserve the correct values of global linear momenta, up to roundoff
error.
However, there is a commonly used reduced, non-conservative form that makes some
analysis and physical interpretation easier. Let us use the product rule to expand the linear
2
Here the transpose notation is particularly cumbersome and unfamiliar, though necessary for full con-
∂
sistency. One will more commonly see this equation written simply as ∂t (ρv) + ∇ · (ρvv) = ρf − ∇p + ∇ · τ .

© 05 December 2025, Joseph M. Powers. All rights reserved.

136 CHAPTER 4. EVOLUTION AXIOMS

momenta principle, then rearrange it, and use mass conservation, Eq. (4.7), and the definition
of material derivative to rewrite the expression:

ρ∂o vi + vi ∂o ρ + vi ∂j (ρvj ) + ρvj ∂j vi = ρfi − ∂i p + ∂j τji , (4.58)

ρ(∂o vi + vj ∂j vi ) + vi ( ∂o ρ + ∂j (ρvj ) ) = ρfi − ∂i p + ∂j τji , (4.59)
| {z }
=0 by mass conservation
ρ (∂o vi + vj ∂j vi ) = ρfi − ∂i p + ∂j τji , (4.60)
| {z }
dvi
= dt

dvi
ρ = ρfi − ∂i p + ∂j τji , (4.61)
dt
dv
T
ρ = ρf − ∇p + ∇T · τ . (4.62)
dt
Written in full, this becomes

ρ(∂o v1 + v1 ∂1 v1 + v2 ∂2 v1 + v3 ∂3 v1 ) = ρf1 − ∂1 p + ∂1 τ11 + ∂2 τ21 + ∂3 τ31 , (4.63)

ρ(∂o v2 + v1 ∂1 v2 + v2 ∂2 v2 + v3 ∂3 v2 ) = ρf2 − ∂2 p + ∂1 τ12 + ∂2 τ22 + ∂3 τ32 , (4.64)
ρ(∂o v3 + v1 ∂1 v3 + v2 ∂2 v3 + v3 ∂3 v3 ) = ρf3 − ∂3 p + ∂1 τ13 + ∂2 τ23 + ∂3 τ33 . (4.65)

So, we see that particles accelerate due to body forces and unbalanced surface forces. If the
surface forces are non-zero but uniform, they will have no gradient or divergence, and hence
not contribute to accelerating a particle.

Example 4.1
Show Newton’s linear momenta principle in the limit of no viscous stress or body force,
dv 1
= − ∇p, (4.66)
dt ρ
is invariant under the Galilean transformation of Ch. 1.4, written here as

x′1 = x1 − v1o t, (4.67)

x′2 = x2 − v2o t, (4.68)
x′3 = x3 − v3o t, (4.69)
′
t = t. (4.70)

We see dt′ = dt. The spatial coordinates of fluid particles have time derivatives, with respect to
the equivalent t or t′ , of
dx′1 dx1
= − v1o , (4.71)
dt′ dt
dx′2 dx2
= − v2o , (4.72)
dt′ dt
dx′3 dx3
= − v3o . (4.73)
dt′ dt

© 05 December 2025, Joseph M. Powers. All rights reserved.

4.2. LINEAR MOMENTA 137

Defining the fluid particle velocities as usual, dxi /dt = vi , dx′i /dt′ = vi′ , we see
v1′ = v1 − v1o , (4.74)
v2′ = v2 − v2o , (4.75)
v3′ = v3 − v3o . (4.76)
Let us expand the equation set to be considered for transformation:
∂v1 ∂v1 ∂v1 ∂v1 1 ∂p
+ v1 + v2 + v3 = − , (4.77)
∂t ∂x1 ∂x2 ∂x3 ρ ∂x1
∂v2 ∂v2 ∂v2 ∂v2 1 ∂p
+ v1 + v2 + v3 = − , (4.78)
∂t ∂x1 ∂x2 ∂x3 ρ ∂x2
∂v3 ∂v3 ∂v3 ∂v3 1 ∂p
+ v1 + v2 + v3 = − . (4.79)
∂t ∂x1 ∂x2 ∂x3 ρ ∂x3
We need representations of the partial derivatives in the transformed coordinate system. Here it is
advantageous to consider a so-called space-time formulation. Our original Cartesian system is obviously
found by inverting the given transformation:
x1 = x′1 + v1o t′ , (4.80)
x2 = x′2 ′
+ v2o t , (4.81)
x3 = x′3 + v3o t′ , (4.82)
′
t = t. (4.83)
For a general space-time transformation, we have
 ∂x1 ∂x1 ∂x1 ∂x1 
  ∂x′1 ∂x′2 ∂x′3 ∂t′  ′
dx1 dx1
 ∂x′2 ∂x2′ ∂x′2 ∂x2 
 dx2   ∂x1 ∂x2 ∂x3 ∂t′   dx′2 
  =  ∂x3 ∂x3 ∂x3 ∂x3  ′ . (4.84)
dx3  ∂x′ ∂x′ ∂x′ ∂t′  dx3
1 2 3
dt ∂t ∂t ∂t ∂t dt′
∂x′1 ∂x′2 ∂x′3 ∂t′
| {z }
J

We have the Jacobian matrix J, specialized for our Galilean transformation, as

 ∂x1 ∂x1 ∂x1 ∂x1 
∂x′1 ∂x′2 ∂x′3 ∂t′  
1 0 0 v1o
 ∂x2′ ∂x′2 ∂x2′ ∂x′2 
 ∂x ∂x ∂x ∂t   0 1 0 v2o 
J =  ∂x31 ∂x23 ∂x33 ∂x3  =  . (4.85)
 ∂x′ ∂x′ ∂x′ ∂t′  0 0 1 v3o
1 2 3
∂t ∂t ∂t ∂t 0 0 0 1
∂x′1 ∂x′2 ∂x′3 ∂t′

Here J = det J = 1, so the transformation is “volume”- and “orientation”-preserving, in the sense of

space-time, and never singular. We also can easily show
 
1 0 0 0

−1  0 1 0 0
JT = , (4.86)
0 0 1 0
−v1o −v2o −v3o 1
and thus have from Eq. (2.286) that
   ∂ 
∂   ∂x
∂x1 1 0 0 0 ′
1
 ∂   0  ∂′ 
 ∂x2 = 1 0 0   ∂x2 
 ∂   . (4.87)
∂x3
0 0 1 0  ∂′ 
∂x3
∂ −v1o −v2o −v3o 1 ∂
∂t ∂t′

© 05 December 2025, Joseph M. Powers. All rights reserved.

138 CHAPTER 4. EVOLUTION AXIOMS

This gives then the simple

∂ ∂ ∂ ∂ ∂ ∂
= , = , = , (4.88)
∂x1 ∂x′1 ∂x2 ∂x′2 ∂x3 ∂x′3
and the slightly more complicated
∂ ∂ ∂ ∂ ∂
= ′ − v1o ′ − v2o ′ − v3o ′ . (4.89)
∂t ∂t ∂x1 ∂x2 ∂x3
Let us apply the full Galilean transformation to one of the linear momenta equations, Eq. (4.77).
∂v1 ∂v1 ∂v1 ∂v1 1 ∂p
+ v1 + v2 + v3 =− , (4.90)
∂t ∂x1 ∂x2 ∂x3 ρ ∂x1
First, we replace all the velocities with their respective relative velocities:
∂ ′ ∂ ∂ ∂
(v + v1o ) + (v1′ + v1o ) (v ′ + v1o ) + (v2′ + v2o ) (v ′ + v1o ) + (v3′ + v3o ) (v ′ + v1o )
∂t 1 ∂x1 1 ∂x2 1 ∂x3 1
1 ∂p
=− . (4.91)
ρ ∂x1
Because vio are all constant, the equation simplifies to
∂v1′ ∂v ′ ∂v ′ ∂v ′ 1 ∂p
+ (v1′ + v1o ) 1 + (v2′ + v2o ) 1 + (v3′ + v3o ) 1 = − . (4.92)
∂t ∂x1 ∂x2 ∂x3 ρ ∂x1
Because of how spatial derivatives transform, Eq. (4.88), we can say
∂v1′ ∂v ′ ∂v ′ ∂v ′ 1 ∂p
+ (v1′ + v1o ) 1′ + (v2′ + v2o ) 1′ + (v3′ + v3o ) 1′ = − . (4.93)
∂t ∂x1 ∂x2 ∂x3 ρ ∂x′1
We next transform the time derivative via Eq. (4.89) to get
∂v1′ ∂v ′ ∂v ′ ∂v ′ ∂v ′ ∂v ′ ∂v ′
′
− v1o 1′ − v2o 1′ − v3o 1′ + (v1′ + v1o ) 1′ + (v2′ + v2o ) 1′ + (v3′ + v3o ) 1′
∂t ∂x1 ∂x2 ∂x3 ∂x1 ∂x2 ∂x3
1 ∂p
=− . (4.94)
ρ ∂x′1
This simplifies significantly to yield an equation that is invariant in form from the untransformed
Eq. (4.77):
∂v1′ ∂v ′ ∂v ′ ∂v ′ 1 ∂p
+ v1′ 1′ + v2′ 1′ + v3′ 1′ = − . (4.95)
∂t ′ ∂x1 ∂x2 ∂x3 ρ ∂x′1
This extends to the 2 and 3 linear momentum equations, yielding the general transformed linear mo-
menta equation to be represented as
∂vi′ ∂v ′ 1 ∂p
+ vj′ i′ = − . (4.96)
∂t ′ ∂xj ρ ∂x′i
In terms of the material derivative, we could say
dvi′ 1 ∂p
=− . (4.97)
dt′ ρ ∂x′i
The invariance of the linear momemta principle under Galilean transformation is the linchpin of New-
tonian mechanics.

© 05 December 2025, Joseph M. Powers. All rights reserved.

4.3. ANGULAR MOMENTA 139

v
m

Figure 4.5: Sketch of particle of mass m velocity v rotating about an axis centered at point
O, with radial distance vector r.

4.3 Angular momenta

It is often easy to overlook the angular momenta principle, and its consequence is so simple
that, it is often just asserted without proof. In fact in classical rigid body mechanics, it
is redundant with the linear momenta principle. It is, however, an independent axiom for
continuous deformable media.
Let us first recall some notions from classical rigid body mechanics, while referring to the
sketch of Fig. 4.5. We have the angular momenta vector L̂ for the particle of Fig. 4.5

L̂ = r × (mv). (4.98)

Any force F that acts on m with lever arm r induces a torque T̂ that is

T̂ = r × F. (4.99)

Now let us apply these notions for an infinitesimal fluid particle with differential mass ρ dV .

Angular momenta = r × (ρ dV )v = ρǫijk rj vk dV, (4.100)

Torque of body force = r × f(ρ dV ) = ρǫijk rj fk dV, (4.101)
Torque of surface force = r × t dA = ǫijk rj tk dA,
= r × (nT · T)T dA = ǫijk rj np Tpk dA,
(4.102)
T
Angular momenta from surface couples = n · H dA = nk Hki dA. (4.103)

Now the principle, that in words says the time rate of change of angular momenta of a
material region is equal to the sum of external couples (or torques) on the system becomes
mathematically,
Z Z Z
d
ρǫijk rj vk dV = ρǫijk rj fk dV + (ǫijk rj np Tpk + nk Hki ) dA . (4.104)
dt Vm (t) Vm (t) Am (t)
| {z } | {z }
Apply Reynolds then Gauss apply Gauss

© 05 December 2025, Joseph M. Powers. All rights reserved.

140 CHAPTER 4. EVOLUTION AXIOMS

We apply Reynolds transport theorem and Gauss’s theorem to the indicated terms and let
the volume of the material region shrink to zero now. First with Reynolds, we get
Z Z
∂o ρǫijk rj vk dV + ǫijk ρrj vk np vp dA =
Vm (t) Am (t)
Z Z
ρǫijk rj fk dV + (ǫijk rj np Tpk + nk Hki ) dA. (4.105)
Vm (t) Am (t)

Next with Gauss we get

Z Z
∂o ρǫijk rj vk dV + ǫijk ∂p (ρrj vk vp ) dV =
Vm (t) Vm (t)
Z Z Z
ρǫijk rj fk dV + ǫijk ∂p (rj Tpk ) dV + ∂k Hki dV. (4.106)
Vm (t) Vm (t) Vm (t)

As the region is arbitrary, the integrand formed by placing all terms under the same integral
must be zero, that yields

ǫijk (∂o (ρrj vk ) + ∂p (ρrj vp vk ) − ρrj fk − ∂p (rj Tpk )) = ∂k Hki . (4.107)

Using the product rule to expand some of the derivatives, we get

 
 
ǫijk 
 r ∂ (ρv ) + ρv ∂ r +r ∂ (ρv v ) + ρv v ∂ r −r ρf − r ∂ T − T ∂ r  = ∂k Hki .
|{z}
j o k k o j j p p k p k p j j k j p pk pk p j
|{z} |{z}
=0 δpj δpj
(4.108)
Applying the simplifications indicated and rearranging, we get

ǫijk rj (∂o (ρvk ) + ∂p (ρvp vk ) − ρfk − ∂p Tpk ) = ∂k Hki − ρǫijk vj vk + ǫijk Tjk . (4.109)
| {z }
=0 by linear momenta

So, we can say,

 
 
∂k Hki = ǫijk (ρvj vk − Tjk ) = ǫijk ρvj vk − T(jk) − T[jk]  , (4.110)
|{z} | {z } |{z} |{z}
anti−sym. sym. sym. anti−sym.

= −ǫijk T[jk] . (4.111)

We have utilized the fact that the tensor inner product of any anti-symmetric tensor with
any symmetric tensor must be zero. Now, if we have the case where there are no externally
imposed angular momenta fields, such as could be the case when electromagnetic forces are

© 05 December 2025, Joseph M. Powers. All rights reserved.

4.4. ENERGY 141

important, we have the common condition of Hki = 0, and the angular momenta principle
reduces to the simple statement that
T[ij] = 0. (4.112)

That is, the anti-symmetric part of the stress tensor must be zero. Hence, the stress tensor,
absent any surface couples, must be symmetric, and we get in Cartesian index and Gibbs
notation:

Tij = Tji , (4.113)

T = TT . (4.114)

4.4 Energy
We recall the first law of thermodynamics, that states the time rate of change of a material
region’s internal and kinetic energy is equal to the rate of heat transferred to the material
region less the rate of work done by the material region. Here we have adopted the common
engineering sign convention for heat and work, motivated by steam engine analysis, for which
thermal energy came “in” and work came “out.” Mathematically, this is stated as

dE dQ dW
= − . (4.115)
dt dt dt

In this case (though this is not uniformly enforced in these notes), the upper case letters de-
note extensive thermodynamic properties. For example, E is extensive total energy, inclusive
of internal and kinetic,
 
1
E = ρV e + vj vj , (4.116)
2

with SI units of J.3 We could have included potential energy in E, but will instead absorb it
into the work term W . The corresponding intensive total energy with SI units of J/kg is

ε = e + vj vj /2. (4.117)

Let us consider each term in the first law of thermodynamics in detail and then write the
equation in final form.
3
The computational fluid dynamics literature often makes the unfortunate choice of defining the “total
energy” as E = e + vj vj /2 with units of J/kg. Thus, it is really a specific energy and violates the ther-
modynamics convention that lower case variables are used for intensive properties. We will not use this
nomenclature, and will generally reserve upper case variables for extensive properties. We will consider
“total” to imply the sum of internal and kinetic, that could either be extensive or intensive. We take the
extensive internal energy to be E = ρV e.

© 05 December 2025, Joseph M. Powers. All rights reserved.

142 CHAPTER 4. EVOLUTION AXIOMS

4.4.1 Total energy term

For a fluid particle, the differential amount of extensive total energy is
 
1
dE = ρ e + vj vj dV, (4.118)
2
 
1
= ρ dV e + vj vj . (4.119)
| {z } 2
mass | {z }
specific internal +kinetic energy

4.4.2 Work term

Recall that work is done when a force acts through a distance, and a work rate arises when
a force acts through a distance at a particular rate in time (hence, a velocity is involved).
Recall also that work is the dot product (inner product) of the force vector with the position
or velocity that gives the true work or work rate. In shorthand, we could say
dW = dxT · F, (4.120)
T
dW dx
= · F = vT · F. (4.121)
dt dt
Here W has the SI units of J, and F has the SI units of N. We contrast this with our
expression for body force per unit mass f, that has SI units of N/kg = m/s2 . Now for the
materials we consider, we must describe work done by two types of forces: 1) body, and 2)
surface.
• Work rate done by a body force
Work rate done by force on fluid = (ρ dV )(fi )vi , (4.122)
Work rate done by fluid = −ρvi fi dV. (4.123)

• Work rate done by a surface force

Work rate done by force on fluid = (ti dA)vi = ((nj Tji ) dA)vi , (4.124)
Work rate done by fluid = −nj Tji vi dA. (4.125)

4.4.3 Heat transfer term

The only thing confusing about the heat transfer rate is the sign convention. We recall that
heat transfer to a body is associated with an increase in that body’s energy. Now following
the scenario sketched in the material region of Fig. 4.6, we define the heat flux vector qi as a
vector that points in the direction of thermal energy flow that has units of energy per area
per time; in SI this would be W/m2 . So, we have
heat transfer rate from body through dA : ni qi dA, (4.126)
heat transfer rate to body through dA : −ni qi dA. (4.127)

© 05 December 2025, Joseph M. Powers. All rights reserved.

4.4. ENERGY 143

dA

Figure 4.6: Sketch of finite material region Vm , infinitesimal mass element ρ dV , and
infinitesimal surface element dA with unit normal ni , and heat flux vector qi .

4.4.4 Conservative form of the energy equation

Putting the words of the first law into equation form, we get
Z   Z Z Z 
d 1
ρ e + vj vj dV = (−ni qi ) dA − (−ni Tij vj ) dA + (−ρfi vi ) dV ,
dt Vm (t) 2 Am (t) Am (t) Vm (t)
| {z } | {z } | {z }
dE dQ dW
dt dt dt

(4.128)
Z   Z Z Z
d 1
ρ e + vj vj dV = − ni qi dA + ni Tij vj dA + ρfi vi dV.
dt Vm (t) 2 Am (t) Am (t) Vm (t)
(4.129)

Skipping the details of an identical application of the Reynolds transport theorem and
Gauss’s theorem, and shrinking the volume to approach zero, we obtain the differential
equation of energy in conservative or divergence form (in first Cartesian index then Gibbs
notation):
     
1 1
∂o ρ e + vj vj + ∂i ρvi e + vj vj =
2 2
| {z } | {z }
rate of change of total energy advection of total energy
− ∂i qi + ∂ (T v ) + ρvi fi , (4.130)
|{z} | i {zij j} | {z }
diffusive heat flux surface force work rate body force work rate
     
∂ 1 1
ρ e + vT · v + ∇T · ρv e + vT · v =
∂t 2 2
−∇T · q + ∇T · (T · v) + ρvT · f. (4.131)

This is a scalar equation as there are no free indices.

© 05 December 2025, Joseph M. Powers. All rights reserved.

144 CHAPTER 4. EVOLUTION AXIOMS

We can segregate the work done by the surface forces into that done by pressure forces
and that done by viscous forces by rewriting this in terms of p and τij as follows
     
1 1
∂o ρ e + vj vj + ∂i ρvi e + vj vj =
2 2
−∂i qi − ∂i (pvi ) + ∂i (τij vj ) + ρvi fi , (4.132)
     
∂ 1 T T 1 T
ρ e+ v ·v + ∇ · ρv e + v · v =
∂t 2 2
−∇T · q − ∇T · (pv) + ∇T · (τ · v) + ρvT · f. (4.133)

4.4.5 Secondary forms of the energy equation

While the energy equation just derived is perfectly valid for all continuous materials, it is
common to see other forms. Many will be described here.

[Link] Enthalpy-based conservative formulation

It is common to bring the pressure-volume work term to the left side to rewrite the conser-
vative energy equation, Eq. (4.132), as
     
1 1 p
∂o ρ e + vj vj + ∂i ρvi e + vj vj + = −∂i qi + ∂i (τij vj )
2 2 ρ
+ρvi fi , (4.134)
     
∂ 1 1 p
ρ e + vT · v + ∇T · ρv e + vT · v + = −∇T · q + ∇T · (τ · v)
∂t 2 2 ρ
+ρvT · f. (4.135)

Recall from elementary thermodynamics the specific enthalpy h is defined as

p
h=e+ . (4.136)
ρ

Using this definition, the first law in conservative form can be rewritten as
     
1 1 ∂p
∂o ρ h + vj vj + ∂i ρvi h + vj vj = − ∂i qi + ∂i (τij vj )
2 2 ∂t
+ρvi fi , (4.137)
     
∂ 1 1 ∂p
ρ h + vT · v + ∇T · ρv h + vT · v = − ∇T · q + ∇T · (τ · v)
∂t 2 2 ∂t
+ρvT · f. (4.138)

© 05 December 2025, Joseph M. Powers. All rights reserved.

4.4. ENERGY 145

[Link] Mechanical energy equation

The mechanical energy equation has no foundation in the first law of thermodynamics;
instead, it is entirely a consequence of the linear momenta principle. It is the type of energy
that is often considered in classical Newtonian particle mechanics, a world in which energy
is either potential or kinetic but not thermal. We include it here because one needs to be
able to distinguish mechanical from thermal energy and it will be useful in later analyses.
The mechanical energy equation, a pure consequence of the linear momenta principle, is
obtained by taking the dot product (inner product) of the velocity vector with the linear
momenta principle:
vT · linear momenta.
In detail, we get

vj (ρ∂o vj + ρvi ∂i vj ) = ρvj fj − vj ∂j p + (∂i τij )vj , (4.139)

v v  v v 
j j j j
ρ∂o + ρvi ∂i = ρvj fj − vj ∂j p + (∂i τij )vj , (4.140)
2 2
vj vj vj vj vj vj
2
mass : ∂o ρ + ∂i (ρvi ) = 0. (4.141)
2 2
We add Eqs. (4.140) and (4.141) and use the product rule to get
 vv   vv 
j j j j
∂o ρ + ∂i ρvi = ρvj fj − vj ∂j p + (∂i τij )vj . (4.142)
 T 2  2 
∂ v ·v T vT · v

ρ + ∇ · ρv = ρvT · f − vT · ∇p + ∇T · τ · v. (4.143)
∂t 2 2

The term ρvj vj /2 represents the volume-averaged kinetic energy, with SI units J/m3 . The
mechanical energy equation, Eq. (4.142), predicts the kinetic energy increases due to three
effects:

• fluid motion in the direction of a body force,

• fluid motion in the direction of decreasing pressure, or

• fluid motion in the direction of increasing viscous stress.

Body forces themselves affect mechanical energy, while it is imbalances in surface forces that
affect mechanical energy.
We could also summarize the non-conservative form of the mechanical energy equation,
Eq. (4.140), as

d  vj vj 
ρ = ρvj fj − vj ∂j p + (∂i τij )vj , (4.144)
dt
 T 2 
d v ·v

ρ = ρvT · f − vT · ∇p + ∇T · τ · v. (4.145)
dt 2

© 05 December 2025, Joseph M. Powers. All rights reserved.

146 CHAPTER 4. EVOLUTION AXIOMS

[Link] Thermal energy equation

If we take the conservative form of the energy equation (4.132) and subtract from it the
mechanical energy equation (4.142), we get an equation for the evolution of thermal energy:

∂o (ρe) + ∂i (ρvi e) = −∂i qi − p∂i vi + τij ∂i vj , (4.146)

∂
(ρe) + ∇T · (ρve) = −∇T · q − p∇T · v + τ : ∇vT . (4.147)
∂t
Here ρe is the volume-averaged internal energy with SI units J/m3 . The thermal energy
equation (4.146) predicts thermal energy (or internal energy) increases due to
• negative gradients in heat flux (more heat enters than leaves),

• pressure force accompanied by a mean negative volumetric deformation (that is, a

uniform compression; note that ∂i vi is the relative expansion rate), or

• viscous force associated with a deformation4 (we will worry about the sign later).
In contrast to mechanical energy, thermal energy changes do not require surface force imbal-
ances; instead they require kinematic deformation. Moreover, body forces have no influence
on thermal energy. The work done by a body force is partitioned entirely to the mechanical
energy of a body.
We can also say
dp
∂o (ρh) + ∂i (ρvi h) = − ∂i qi + τij ∂i vj , (4.148)
dt
∂ dp
(ρh) + ∇T · (ρvh) = − ∇T · q + τ : ∇vT . (4.149)
∂t dt

[Link] Non-conservative energy equation

We can obtain the commonly used non-conservative form of the energy equation, also known
as the energy equation following a fluid particle, by the following operations. First expand
the thermal energy equation (4.146):

ρ∂o e + e∂o ρ + ρvi ∂i e + e∂i (ρvi ) = −∂i qi − p∂i vi + τij ∂i vj . (4.150)

Then regroup and notice terms common from mass conservation, Eq. (4.7):

ρ (∂o e + vi ∂i e) +e (∂o ρ + ∂i (ρvi )) = −∂i qi − p∂i vi + τij ∂i vj , (4.151)

| {z } | {z }
de =0 by mass
dt

4
For a general fluid, this includes a mean volumetric deformation as well as a deviatoric deformation. If
the fluid satisfies Stokes’ assumption, it is only the deviatoric deformation that induces a change in internal
energy in the presence of viscous stress.

© 05 December 2025, Joseph M. Powers. All rights reserved.

4.4. ENERGY 147

so we get
de
ρ = −∂i qi − p∂i vi + τij ∂i vj , (4.152)
dt
de
ρ = −∇T · q − p∇T · v + τ : ∇vT . (4.153)
dt
We can get an equation that is reminiscent of elementary thermodynamics, valid for
small volumes V by multiplying Eq. (4.152) by V and using Eq. (3.187) to replace ∂i vi by
its known value in terms of the relative expansion rate to obtain
de dV
ρV = −V ∂i qi − p + V τij ∂i vj . (4.154)
dt dt
The only term not usually found in elementary thermodynamics texts is the third on the
right hand side, that is a viscous work term.

[Link] Energy equation in terms of enthalpy

Often the energy equation is fully cast in terms of enthalpy. This is generally valid, but
especially useful in constant pressure environments. Now starting with the energy equation
following a particle (4.152), we can use one form of the mass equation, Eq. (4.14), to eliminate
the relative expansion rate ∂i vi in favor of the material derivative of density to get
de p dρ
ρ = −∂i qi + + τij ∂i vj . (4.155)
dt ρ dt
Rearranging, we get  
de p dρ
ρ − = −∂i qi + τij ∂i vj . (4.156)
dt ρ2 dt
Now differentiating Eq. (4.136), h = e + p/ρ, we find
p 1
dh = de −
2
dρ + dp, (4.157)
ρ ρ
dh de p dρ 1 dp
= − 2 + , (4.158)
dt dt ρ dt ρ dt
de p dρ dh 1 dp
− 2 = − , (4.159)
dt ρ dt dt ρ dt
de p dρ dh dp
ρ − = ρ − . (4.160)
dt ρ dt dt dt
So, using Eq. (4.160) to eliminate de/dt in Eq. (4.156) in favor of dh/dt, the energy equation
in terms of enthalpy becomes
dh dp
ρ = − ∂i qi + τij ∂i vj , (4.161)
dt dt
dh dp
ρ = − ∇T · q + τ : ∇vT . (4.162)
dt dt

© 05 December 2025, Joseph M. Powers. All rights reserved.

148 CHAPTER 4. EVOLUTION AXIOMS

[Link] Energy equation in terms of entropy

By using standard relations from thermodynamics, we can write the energy equation in terms
of entropy. It is important to note that this is just an algebraic substitution. The physical
principle that this equation will represent is still energy conservation.
Recall the Gibbs equation from thermodynamics, that serves to define entropy s:
T ds = de + p dv̂. (4.163)
Here T is the absolute temperature, and v̂ is the specific volume, v̂ = V /m = 1/ρ. In terms
of ρ, the Gibbs equation is
p
T ds = de − 2 dρ. (4.164)
ρ
Taking the material derivative of Eq. (4.164) , that is operationally equivalent to dividing
by dt, and solving for de/dt,we get
de ds p dρ
=T + 2 . (4.165)
dt dt ρ dt
This is still essentially a thermodynamic definition of s. Now use Eq. (4.165) in the non-
conservative energy equation (4.152) to get an alternate expression for the first law:
ds p dρ
ρT + = −∂i qi − p∂i vi + τij ∂i vj . (4.166)
dt ρ dt
Recalling Eq. (4.14), −∂i vi = (1/ρ)(dρ/dt), we have
ds
ρT = −∂i qi + τij ∂i vj , (4.167)
dt
ds 1 1
ρ = − ∂i qi + τij ∂i vj . (4.168)
dt T T
Using the fact that from the quotient rule we have ∂i (qi /T ) = (1/T )∂i qi − (qi /T 2 )∂i T , we
can then say
ds q  1 1
i
ρ = −∂i − 2 qi ∂i T + τij ∂i vj , (4.169)
dt T T T
ds q 1 1
ρ = −∇T · − 2 qT · ∇T + τ : ∇vT . (4.170)
dt T T T
From this statement, we can conclude from the first law of thermodynamics that the entropy
of a fluid particle changes due to heat transfer and to deformation in the presence of viscous
stress. We will make a more precise statement about entropy changes after we introduce the
second law of thermodynamics.
The energy equation in terms of entropy can be written in conservative or divergence
form by adding the product of s and the mass equation, s∂o ρ + s∂i (ρvi ) = 0, to Eq. (4.169)
to obtain
q  1 1
i
∂o (ρs) + ∂i (ρvi s) = −∂i − 2 qi ∂i T + τij ∂i vj , (4.171)
T T T
∂ q 1 1
(ρs) + ∇T · (ρvs) = −∇T · − 2 qT · ∇T + τ : ∇vT . (4.172)
∂t T T T

© 05 December 2025, Joseph M. Powers. All rights reserved.

4.5. ENTROPY INEQUALITY 149

4.5 Entropy inequality

Let us use a non-rigorous method to suggest a form of the entropy inequality that is consistent
with classical thermodynamics. Recall the mathematical statement of the entropy inequality
from classical thermodynamics:
dQ
dS ≥ . (4.173)
T
Here S is the extensive entropy, with SI units J/K, and Q is the heat energy into a system
with SI units of J. Notice that entropy can go up or down in a process, depending on the
heat transferred. If the process is adiabatic, dQ = 0, and the entropy can either remain fixed
or rise. Now for our continuous material we have
dS = ρs dV, (4.174)
dQ = −qi ni dA dt. (4.175)
Here we have used s for the specific entropy, that has SI units J/kg/K. Notice we must
be careful with our sign convention. When the heat flux vector is aligned with the outward
normal, heat leaves the system. Because we want positive dQ to represent heat into a system,
we need the negative sign.
The second law becomes then
qi
ρs dV ≥ − ni dA dt. (4.176)
T
Now integrate over the finite geometry: on the left side this is a volume integral and the
right side this is an area integral.
Z Z 
qi
ρs dV ≥ − ni dA dt. (4.177)
Vm (t) Am (t) T
Differentiating with respect to time and then applying our typical machinery to the
second law gives rise to
Z Z
d qi
ρs dV ≥ − ni dA, (4.178)
dt Vm (t) Am (t) T
Z Z Z
qi
∂o (ρs) dV + ρsvi ni dA ≥ − ni dA, (4.179)
Vm (t) Am (t) Am (t) T
Z Z q 
i
(∂o (ρs) + ∂i (ρsvi )) dV ≥ −∂i dV, (4.180)
Vm (t) Vm (t) T
Z Z q  Z
i
(∂o (ρs) + ∂i (ρsvi )) dV = −∂i dV + σ̇ dV. (4.181)
Vm (t) Vm (t) T Vm (t)

Here we have defined the irreversible entropy production rate, σ̇ ≥ 0, as a positive semi-
definite scalar. It is simply a convenience to replace the inequality with an equality. Then
we have the conservative form
q 
i
∂o (ρs) + ∂i (ρsvi ) = −∂i + σ̇, σ̇ ≥ 0. (4.182)
T

© 05 December 2025, Joseph M. Powers. All rights reserved.

150 CHAPTER 4. EVOLUTION AXIOMS

Invoking mass conservation, Eq. (4.7), we easily get the non-conservative form
ds q 
i
ρ = −∂i + σ̇, σ̇ ≥ 0. (4.183)
dt T
This is the second law. Now if we subtract from this the first law written in terms of entropy,
Eq. (4.169), we get the result
1 1
σ̇ = − 2
qi ∂i T + τij ∂i vj ≥ 0. (4.184)
T T | {z }
Φ
As an aside, we have defined the commonly used viscous dissipation function Φ as

Φ ≡ τij ∂i vj . (4.185)

For symmetric stress tensors, we also have Φ = τij ∂(i vj) . Now because σ̇ ≥ 0, we can view
the entirety of the second law as the following constraint, sometimes called the weak form of
the Clausius-Duhem56 inequality:
1 1
− qi ∂i T + τij ∂i vj ≥ 0, (4.186)
T2 T
1 1
− 2 qT · ∇T + τ : ∇vT ≥ 0, (4.187)
T T
1 T 1
− 2 q · ∇T + τ : L ≥ 0. (4.188)
T T
Recalling that τij is symmetric by the angular momenta principle for no external couples, and,
consequently, that its tensor inner product with the velocity gradient only has a contribution
from the symmetric part of the velocity gradient (that is, the deformation rate or strain rate
tensor), the entropy inequality reduces slightly to
1 1
2
−
qi ∂i T + τij ∂(i vj) ≥ 0, (4.189)
T T

T !
1 T 1 ∇v + ∇vT
T
− 2 q · ∇T + τ : ≥ 0, (4.190)
T T 2
1 T 1
− q · ∇T + τ : D ≥ 0. (4.191)
T2 T
We shall see in upcoming sections that we will be able to specify qi and τij in such a fashion
that is both consistent with experiment and satisfies the entropy inequality.
5
Rudolf Clausius, 1822-1888, Prussian-born German mathematical physicist, key figure in making ther-
modynamics a science, author of well-known statement of the second law of thermodynamics, taught at
Zürich Polytechnikum, University of Würzburg, and University of Bonn.
6
Pierre Maurice Marie Duhem, 1861-1916, French physicist, mathematician, and philosopher, taught at
Lille, Rennes, and the University of Bordeaux.

© 05 December 2025, Joseph M. Powers. All rights reserved.

4.5. ENTROPY INEQUALITY 151

The more restrictive (and in some cases, overly restrictive) strong form of the Clausius-
Duhem inequality requires each term to be greater than or equal to zero. For our system
the strong form, realizing that the absolute temperature T > 0, is
−qi ∂i T ≥ 0, τij ∂(i vj) ≥ 0, (4.192)
| {z }
Φ
T

!
T T
∇v + ∇v
−qT · ∇T ≥ 0, τ : ≥ 0. (4.193)
2

It is straightforward to show that terms that generate entropy due to viscous work also
dissipate mechanical energy. This can be cleanly demonstrated by considering the mecha-
nisms that cause mechanical energy the change within a finite fixed control volume V . First
consider the non-conservative form of the mechanical energy equation, Eq. (4.144):
d  vj vj 
ρ = ρvj fj − vj ∂j p + vj ∂i τij . (4.194)
dt 2
Now use the product rule to restate the pressure and viscous work terms so as to achieve
d  vj vj 
ρ = ρvj fj − ∂j (vj p) + p∂j vj + ∂i (τij vj ) − τij ∂i vj . (4.195)
dt 2 | {z }
=Φ≥0
So, here we see what induces local changes in mechanical energy. We see that body forces,
pressure forces and viscous forces in general can induce the mechanical energy to rise or fall.
However that part of the viscous stresses that is associated with the viscous dissipation, Φ,
is guaranteed to induce a local decrease in mechanical energy. It is sometimes said that this
is a transformation in which mechanical energy dissipates into thermal energy.
To study global changes in mechanical energy, we consider the conservative form of the
mechanical energy equation, Eq. (4.142), here written in the same way that takes advantage
of application of the product rule to the pressure and viscous terms:
 vv   vv 
j j j j
∂o ρ + ∂i ρvi = ρvj fj − ∂j (vj p) + p∂j vj + ∂i (τij vj ) − τij ∂i vj . (4.196)
2 2
Now integrate over a fixed control volume with closed boundaries, so that
Z  vv  Z  vv  Z Z Z
j j j j
∂o ρ dV + ∂i ρvi dV = ρvj fj dV − ∂j (vj p) dV + p∂j vj dV
V 2 V 2 VZ V V
Z
+ ∂i (τij vj ) dV − τij ∂i vj dV. (4.197)
V V

Applying Leibniz’s rule, Eq. (2.270), and Gauss’s law, Eq. (2.252), we get
Z Z Z Z Z
d vj vj vj vj
ρ dV + ni ρvi dA = ρvj fj dV − nj vj p dA + p∂j vj dV
dt V 2 A 2 VZ A V
Z
+ ni (τij vj ) dA − τij ∂i vj dV. (4.198)
A V

© 05 December 2025, Joseph M. Powers. All rights reserved.

152 CHAPTER 4. EVOLUTION AXIOMS

Now on the surface of the closed fixed volume, the velocity is zero, so we get
Z Z Z Z
d vj vj
ρ dV = ρvj fj dV + p∂j vj dV − τij ∂i vj dV. (4.199)
dt V 2 V V V | {z }
positive

Now the strong form of the second law requires that τij ∂i vj = τij ∂(i vj) ≥ 0. So, we see for a
finite fixed closed volume of fluid that a body force and pressure force in conjunction with
local volume changes can cause the global mechanical energy to either grow or decay, the
viscous stress always induces a decay of global mechanical energy; in other words it is a
dissipative effect.

4.6 Integral forms

• Our governing equations are formulated based upon laws that apply to a material
element.
• We are not often interested in an actual material element but in some other fixed or
moving region in space.
• Rules for such systems can by formulated with Leibniz’s rule in conjunction with the
differential forms of our axioms.
Let us first recall Leibniz’s rule (2.275) for an arbitrary scalar function f (that has no
relation to our body force term) over a time-dependent arbitrary region Va (t):
Z Z Z
d
f dV = ∂o f dV + ni wi f dA. (4.200)
dt Va (t) Va (t) Aa (t)

Recall that wi is the velocity of the arbitrary surface, not necessarily the particle velocity.

4.6.1 Mass
We rewrite the mass conservation, Eq. (4.7), as
∂o ρ = −∂i (ρvi ). (4.201)
Now let us use this, and let f = ρ in Leibniz’s rule, Eq. (4.200), to get
Z Z Z
d
ρ dV = ∂o ρ dV + ni wi ρ dA, (4.202)
dt Va (t) Va (t) Aa (t)
Z Z Z
d
ρ dV = (−∂i (ρvi )) dV + ni wi ρ dA. (4.203)
dt Va (t) Va (t) Aa (t)

Invoking Gauss, Eq. (2.252), we get

Z Z
d
ρ dV = ni ρ(wi − vi ) dA. (4.204)
dt Va (t) Aa (t)

Now consider three special cases.

© 05 December 2025, Joseph M. Powers. All rights reserved.

4.6. INTEGRAL FORMS 153

[Link] Fixed region

We take wi = 0. So the arbitrary volume that is a function of time, Va (t), becomes a fixed
volume, V . It is bounded by a fixed surface A.
Z Z
d
ρ dV = − ni ρvi dA, (4.205)
dt V A
Z Z
d
ρ dV + ni ρvi dA = 0. (4.206)
dt V A

[Link] Material region

Here we take wi = vi .
Z
d
ρ dV = 0. (4.207)
dt Vm (t)

[Link] Moving solid enclosure with holes

Say the region considered is a solid enclosure with holes through with fluid can enter and
exit. The our arbitrary surface Aa (t) can be specified as

Aa (t) = Ae (t) area of entrances and exits (4.208)

+As (t) solid moving surface with wi = vi (4.209)
+As fixed solid surface with wi = vi = 0. (4.210)

Then we get Z Z
d
ρ dV + ρni (vi − wi ) dA = 0. (4.211)
dt Va (t) Ae (t)

Example 4.2
Consider the volume sketched in Fig. 4.7. Water enters a circular hole of diameter D1 = 1” with
velocity v1 = 3 ft/s. Water enters another circular hole of diameter D2 = 3” with velocity v2 = 2 ft/s.
The cross sectional area of the cylindrical tank is A = 2 ft2 . The tank has height H. Water at density
ρw exists in the tank at height h(t). Air at density ρa fills the remainder of the tank. Find the rate of
rise of the water dh/dt.

Consider two control volumes

• V1 : the fixed region enclosing the entire tank, and

• V2 (t): the material region attached to the air.

First, let us write mass conservation for the material region 2:

© 05 December 2025, Joseph M. Powers. All rights reserved.

154 CHAPTER 4. EVOLUTION AXIOMS

air

water

h(t)

Figure 4.7: Sketch of volume with water and air being filled with water.

Z
d
ρa dV = 0, (4.212)
dt V2
Z H
d
ρa A dz = 0. (4.213)
dt h(t)

Mass conservation for V1 is

Z Z
d
ρ dV + ρvi ni dA = 0. (4.214)
dt V1 Ae

Now break up V1 and write Ae explicitly

Z Z Z Z
d h(t) d H
ρw A dz + ρa A dz = − ρw vi ni dA − ρw vi ni dA, (4.215)
dt 0 dt h(t) A1 A2
| {z }
=0
Z h(t)
d
ρw A dz = ρw v1 A1 + ρw v2 A2 , (4.216)
dt 0
ρw π
= (v1 D12 + v2 D22 ), (4.217)
4
Z h(t)
d ρw π
ρw A dz = (v1 D12 + v2 D22 ), (4.218)
dt 0 4
dh π
= (v1 D12 + v2 D22 ), (4.219)
dt 4A
  2   2 !
π ft 1 ft 3
= 3 ft + 2 ft , (4.220)
4(2 ft2 ) s 12 s 12
7π ft ft
= = 0.057 . (4.221)
384 s s

© 05 December 2025, Joseph M. Powers. All rights reserved.

4.6. INTEGRAL FORMS 155

4.6.2 Linear momenta

Let us perform the same exercise for the linear momenta equation. First, in a strictly
mathematical step, apply Leibniz’s rule, Eq. (4.200), to linear momenta, f = ρvi :
Z Z Z
d
ρvi dV = ∂o (ρvi ) dV + nj wj ρvi dA. (4.222)
dt Va (t) Va (t) Aa (t)

Now invoke the physical linear momenta axiom, Eq. (4.52). Here the axiom gives us an
expression for ∂o (ρvi ). We will also convert volume integrals to surface integrals via Gauss’s
theorem, Eq. (2.252), to get
Z Z Z
d
ρvi dV = − (ρnj (vj − wj )vi + ni p − nj τij ) dA + ρfi dV. (4.223)
dt Va (t) Aa (t) Va (t)

Now momenta flux terms only have values at entrances and exits (at solid surfaces, we get
vi = wi ), so we can say
Z Z Z Z Z
d
ρvi dV + ρnj (vj − wj )vi dA = − ni p dA + nj τij dA + ρfi dV.
dt Va (t) Ae (t) Aa (t) Aa (t) Va (t)
(4.224)
The surface forces are evaluated along all surfaces, not just entrances and exits.

4.6.3 Energy
Applying the same analysis to the energy equation, we obtain
Z   Z  
d 1 1
ρ e + vj vj dV = − ρni (vi − wi ) e + vj vj dA
dt Va (t) 2 Aa (t) 2
Z
− ni qi dA
Va (t)
Z
− (ni vi p − ni τij vj ) dA
Aa (t)
Z
+ ρvi fi dV. (4.225)
Va (t)

4.6.4 General expression

If we have a governing equation from a physical principle that is of the form

∂o fj + ∂i (vi fj ) = ∂i gj + hj , (4.226)

© 05 December 2025, Joseph M. Powers. All rights reserved.

156 CHAPTER 4. EVOLUTION AXIOMS

then we can say for an arbitrary volume that

Z Z Z Z
d
fj dV = − ni fj (vi − wi ) dA + ni gj dA + hj dV . (4.227)
dt Va (t) Aa (t) Aa (t) Va (t)
| {z } | {z } | {z } | {z }
change of fj flux of fj effect of g j effect of hj

4.7 Summary of axioms in differential form

Here we pause to summarize the mathematical form of our axioms. We give the Cartesian
index, Gibbs, and the full non-orthogonal index notation. All details of development of the
non-orthogonal index notation are omitted, and the reader is referred to Aris (1962) for a
full development. We will first present the conservative form and then the non-conservative
form.

4.7.1 Conservative form

[Link] Cartesian index form

∂o ρ + ∂i (ρvi ) = 0, (4.228)
∂o (ρvi ) + ∂j (ρvj vi ) = ρfi − ∂i p + ∂j τji , (4.229)
τij = τji , (4.230)
     
1 1
∂o ρ e + vj vj + ∂i ρvi e + vj vj = −∂i qi − ∂i (pvi ) + ∂i (τij vj )
2 2
+ρvi fi , (4.231)
q 
i
∂o (ρs) + ∂i (ρsvi ) ≥ −∂i . (4.232)
T
If there is no body force, fi = 0, we can rearrange to bring many terms within the
divergence operator, yielding

∂o ρ + ∂i (ρvi ) = 0, (4.233)
∂o (ρvi ) + ∂j (ρvj vi + δji p − τji ) = 0, (4.234)
τij − τji = 0, (4.235)
      
1 1 p
∂o ρ e + vj vj + ∂i ρvi e + vj vj + − τij vj + qi = 0, (4.236)
2 2 ρ
 qi 
∂o (ρs) + ∂i ρsvi + ≥ 0. (4.237)
T

[Link] Gibbs form

∂ρ
+ ∇T · (ρv) = 0, (4.238)
∂t

© 05 December 2025, Joseph M. Powers. All rights reserved.

4.7. SUMMARY OF AXIOMS IN DIFFERENTIAL FORM 157

∂
T
T
(ρv) + ∇T · (ρvvT ) = ρf − ∇p + ∇T · τ , (4.239)
∂t
τ = τT, (4.240)
     
∂ 1 1
ρ e + vT · v + ∇T · ρv e + vT · v = −∇T · q − ∇T · (pv)
∂t 2 2
+∇T · (τ · v) + ρvT · f , (4.241)
∂ q
(ρs) + ∇T · (ρsv) ≥ −∇T · . (4.242)
∂t T
If there is no body force, f = 0, we can rearrange to bring many terms within the divergence
operator, yielding

∂ρ
+ ∇T · (ρv) = 0, (4.243)
∂t
∂
T
(ρv) + ∇T · (ρvvT + pI − τ ) = 0, (4.244)
∂t
τ − τ T = 0, (4.245)
      
∂ 1 T T 1 T p
ρ e+ v ·v + ∇ · ρv e + v · v + −τ ·v+q = 0, (4.246)
∂t 2 2 ρ
∂  q
(ρs) + ∇T · ρsv + ≥ 0. (4.247)
∂t T
If there is a time-independent conservative body force, then we can rewrite f = −∇ϕ; then the
linear momenta and energy axioms can be written as

∂
T
(ρv) + ∇T · (ρvvT + pI − τ ) = −ρ∇ϕ,
∂t
(4.248)
      
∂ 1 1 p
ρ e + vT · v + ϕ + ∇T · ρv e + vT · v + ϕ + −τ ·v+q = 0. (4.249)
∂t 2 2 ρ

[Link] Non-orthogonal index form

Here we present the governing equations for general non-orthogonal coordinate systems. Few
texts give a proper exposition of the conservative form of the equations in non-orthogonal
coordinates. Here we have extended the development of Vinokur7 to include the effects of
momenta and energy diffusion. This extension has been guided by general notions found in
standard works such as Aris (1962) as well as Liseikin (2010).

∂ √ ∂ √

( g ρ) + k g ρv k = 0, (4.250)
 ∂t
  ∂x 
i i i
∂ √ j ∂ξ ∂ √ j k ∂ξ √ j ∂ξ
g ρv + g ρv v = g ρf
∂t ∂xj ∂xk ∂xj ∂xj
7
Vinokur, M., 1974, “Conservation equations of gasdynamics,” Journal of Computational Physics, 14(2):
105-125.

© 05 December 2025, Joseph M. Powers. All rights reserved.

158 CHAPTER 4. EVOLUTION AXIOMS
 i

∂ √ jk ∂ξ
− k g pg
∂x ∂xj
 
∂ √ jk ∂ξ i
+ k gτ ,
∂x ∂xj
(4.251)
     
∂ √ 1 i j ∂ √ 1 ∂ √ k

g ρ e + gij v v + k g ρv k e + gij v i v j = − k gq
∂t 2 ∂x 2 ∂x
∂ √

− k g pv k
∂x
∂ √

+ k g gij v j τ ik
∂x
√
+ g ρgij v j f i , (4.252)
 
∂ √ ∂ √ k

∂ √ qk
( g ρs) + k g ρsv ≥ − k g .
∂t ∂x ∂x T
(4.253)

4.7.2 Non-conservative form

[Link] Cartesian index form

dρ
= −ρ∂i vi , (4.254)
dt
dvi
ρ = ρfi − ∂i p + ∂j τji , (4.255)
dt
τij = τji , (4.256)
de
ρ = −∂i qi − p∂i vi + τij ∂i vj , (4.257)
dt
ds q 
i
ρ ≥ −∂i . (4.258)
dt T

[Link] Gibbs form

dρ
= −ρ∇T · v, (4.259)
dt
dv
T
ρ = ρf − ∇p + ∇T · τ , (4.260)
dt
τ = τT, (4.261)
de
ρ = −∇T · q − p∇T · v + τ : ∇vT , (4.262)
dt
ds q
ρ ≥ −∇T · . (4.263)
dt T

© 05 December 2025, Joseph M. Powers. All rights reserved.

4.7. SUMMARY OF AXIOMS IN DIFFERENTIAL FORM 159

[Link] Non-orthogonal index form

This version is likely correct but has not been checked carefully!

∂ρ ∂ρ ρ ∂ √ i

+ vi i = − √ gv , (4.264)
∂t ∂x g ∂xi
 i  i 
∂v j ∂v i l i ij ∂p 1 ∂ √ ij

ρ +v + Γ jl v = ρf − g + √ g τ + Γijk τ jk , (4.265)
∂t ∂xj ∂xj g ∂xj
 
∂e i ∂e 1 ∂ √ i
p ∂ √ i

ρ +v = − √ g q − √ gv
∂t ∂xi g ∂xi g ∂xi
 i 
kj ∂v i l
+gik τ + Γjl v , (4.266)
∂xj
   
∂s i ∂s 1 ∂ √ qi
ρ +v ≥ −√ g . (4.267)
∂t ∂xi g ∂xi T

The term Γijl v l in the linear momenta equation can be shown to represent the effects of
non-Cartesian terms such as centripetal and Coriolis accelerations, to be fully explored in
Ch. 7.1.1.

4.7.3 Physical interpretations

Each term in the governing axioms represents a physical mechanism. This approach is em-
phasized in the classical text by Bird, Stewart, and Lightfoot (2007) on transport processes.
In general, the equations that are partial differential equations can be represented in the
following form:
local change = advection + diffusion + source. (4.268)
Here we consider advection and diffusion to be types of transport phenomena. If we have a
fixed volume of material, a property of that material, such as its thermal energy, can change
because an outside flow sweeps energy in from outside due to bulk fluid motion. That is
advection.8 The thermal energy can also change because random molecular motions allow
slow leakage to the outside or leakage in from the outside. That is diffusion. Or the material
can undergo intrinsic changes inside, such as viscous work, that converts kinetic energy into
thermal energy.
8
Often the term “convection” is used in a similar fashion as “advection.” While there is not a univer-
sal consensus on the distinctions between these two common terms, “advection” does seem to always be
associated with bulk fluid motion that has vi 6= 0. Such motion is macroscopic and identifiable. It is thus
associated with information and thus order. Advection can apply to mass, momentum, or energy. “Convec-
tion” is more commonly restricted to energy, and sometimes includes both energy advection and diffusion
effects. Diffusion is associated with disordered, random motion of molecules. Diffusion is non-zero, even if
there is no bulk fluid motion, vi = 0. That is to say, a fluid at rest at the macroscale is generally in motion
at the microscale, where the motion is random.

© 05 December 2025, Joseph M. Powers. All rights reserved.

160 CHAPTER 4. EVOLUTION AXIOMS

Let us write the Gibbs form of the non-conservative equations of mass, linear momenta,
and energy in a slightly different way to illustrate these mechanisms:
∂ρ
= local change in mass
∂t
−vT · ∇ρ advection of mass
+0 diffusion of mass
T
−ρ∇ · v, volume expansion source, (4.269)
∂v
ρ = local change in linear momenta
∂t

−ρ vT · ∇ v advection of linear momenta

T
+ ∇T · τ , diffusion of linear momenta (4.270)
+ρf body force source of linear momenta
−∇p pressure gradient source of linear momenta,
∂e
ρ = local change in thermal energy
∂t
−ρvT · ∇e advection of thermal energy
T
−∇ · q diffusion of thermal energy
T
−p∇ · v pressure work thermal energy source
T
+τ : ∇v viscous work thermal energy source. (4.271)
Briefly considering the second law, we note that the irreversible entropy production rate σ̇
is solely associated with diffusion of linear momenta and diffusion of energy. This makes sense
in that diffusion is associated with random molecular motions and thus disorder. Advection
is associated with an ordered motion of matter in that we retain knowledge of the position
of the matter. Pressure-volume work is a reversible work and does not contribute to entropy
changes. A portion of the heat transfer can be considered to be reversible. All of the work
done by the viscous forces is irreversible work.

4.8 Incompleteness of the axioms

The beauty of these axioms is that they are valid for any material that can be modeled as a
continuum under the influence of the forces we have mentioned. Specifically, they are valid
for both solid and fluid mechanics, which is remarkable.
While the axioms are complete, the equations are not! We have twenty-three unknowns
here ρ(1), vi (3), fi (3), p(1), τij (9), e(1), qi (3), T (1), s(1), and only eight equations (one mass,
three linear momenta, three independent angular momenta, one energy). We cannot really
count the second law as an equation, as it is an inequality. Whatever result we get must
be consistent with it. Whatever the case, we are short a number of equations. We will see
in Ch. 5 how we use constitutive equations, equations founded in empiricism, that in some
sense model sub-continuum effects that we have ignored, to complete our system.

© 05 December 2025, Joseph M. Powers. All rights reserved.

Chapter 5

Constitutive equations

see Powers, Chapter 5,

see Panton, Chapter 6,
see Hughes and Gaylord, Chapter 1,
see Whitaker,
see Aris, Chapters 5 and 6.

In this chapter, we return to the problem of completing our set of equations as introduced
in Ch. 4.8. Constitutive equations are additional equations based on experiment that are
not as fundamental as the previously developed axioms that can complete our continuum
description. They can be rather ad hoc relations that in some sense model the sub-continuum
nano-structure. In some cases, for example, the sub-continuum kinetic theory of gases, we
can show that when the sub-continuum is formally averaged, that we obtain commonly used
constitutive equations. In most cases however, constitutive equations simply represent curve
fits to basic experimental results, that can vary widely from material to material. As is
briefly discussed below, constitutive equations are not completely arbitrary. Whatever is
proposed must allow our final equations to be invariant under Galilean transformations and
rotations as well as satisfy the entropy inequality.
For example, we might hope to develop a constitutive equation for the heat flux vector
qi . Being naı̈ve, we might in general expect it to be a function of a large number of variables
as well as their gradients:

qi = qi (ρ, p, T, vj , τjk , fj , e, s, ∂j ρ, ∂j p, ∂j T, ∂j vk , . . .). (5.1)

The principles of continuum mechanics will rule out some possibilities, but still allow a broad
range of forms.

5.1 Frame and material indifference

Our choice of a constitutive law must be invariant under a Galilean transformation (frame
invariance) a rotation (material indifference). Say for example, we propose that the heat

161
162 CHAPTER 5. CONSTITUTIVE EQUATIONS

flux vector is proportional to the velocity vector

qi = avi , trial constitutive relation. (5.2)

If we changed frames such that velocities in the moving frame were ui = vi −V , we would have
qi = a(ui + V ). With this constitutive law, we find a physical quantity is dependent on the
frame velocity, that we observe to be non-physical; hence we rule out this trial constitutive
relation.
A commonly used constitutive law for stress in a one-dimensional experiment is

τ12 = b(∂1 v2 )a (∂1 u2 )b , (5.3)

where u2 is the displacement of particle. While this may fit one-dimensional data well, it is
in no way clear how one could simply extend this to write an expression for τij , and many
propositions will fail to satisfy material indifference.

5.2 Second law restrictions and Onsager relations

The entropy inequality from the second law of thermodynamics provides additional restric-
tions on the form of constitutive equations. Recall the second law (equivalently, the weak
form of the Clausius-Duhem inequality, Eq. (4.189)) tells us that
1 1
− 2
qi ∂i T + τij ∂(i vj) ≥ 0. (5.4)
T T
We would like to find forms of qi and τij that are consistent with the weak form of the
entropy inequality, Eq. (5.4).

5.2.1 Weak form of the Clausius-Duhem inequality

The weak form suggests that we may want to consider both qi and τij to be functions
involving the temperature gradient ∂i T and the deformation tensor ∂(i vj) .

[Link] Non-physical motivating example

To see that this is actually too general of an assumption, it suffices to consider a one-
dimensional limit. In the one-dimensional limit, the weak form of the entropy inequality,
Eq. (4.189), reduces to
1 ∂T 1 ∂u
− 2q + τ ≥ 0. (5.5)
T ∂x T ∂x
We can write this in a vector form as
 q 
1 ∂T 1 ∂u
( − T ∂x u ∂x ) τTu ≥ 0. (5.6)
T

© 05 December 2025, Joseph M. Powers. All rights reserved.

5.2. SECOND LAW RESTRICTIONS AND ONSAGER RELATIONS 163

A factor of u/u was introduced to the viscous stress term. This allows for a necessary
dimensional consistency in that q/T has the same units as τ u/T . Let us then hypothesize a
linear relationship exists between the generalized fluxes q/T and τ u/T and the generalized
driving gradients −(1/T )∂T /∂x and (1/u)∂u/∂x:
 
q 1 ∂T 1 ∂u
= C11 − + C12 , (5.7)
T T ∂x u ∂x
 
τu 1 ∂T 1 ∂u
= C21 − + C22 . (5.8)
T T ∂x u ∂x

In matrix form this becomes

 q
   
T
C11 C12 − T1 ∂T
∂x
τu = 1 ∂u . (5.9)
T
C21 C22 u ∂x

We then substitute this hypothesized relationship into the entropy inequality to obtain
   1 ∂T 
1 ∂T 1 ∂u C11 C12 − T ∂x
( − T ∂x u ∂x ) 1 ∂u ≥ 0. (5.10)
C21 C22 u ∂x

We next segregate the matrix Cij into symmetric and anti-symmetric parts to get
 C12 +C21
  C12 −C21
  1 ∂T 
1 ∂T 1 ∂u C11 0 − T ∂x
( − T ∂x u ∂x ) C21 +C12
2 + C21 −C12 2
1 ∂u ≥ 0. (5.11)
2
C 22 2
0 u ∂x

Distributing the multiplication, we find

 C12 +C21
  1 ∂T 
1 ∂T 1 ∂u C11 − T ∂x
( − T ∂x u ∂x ) C21 +C12 2
1 ∂u
2
C 22 u ∂x
 C12 −C21
  1 ∂T 
1 ∂T 1 ∂u 0 − T ∂x
+ ( − T ∂x u ∂x ) C21 −C12 2
1 ∂u ≥ 0. (5.12)
2
0 u ∂x
| {z }
=0

The second term is identically zero for all values of temperature and velocity gradients. So
what remains is the inequality involving only a symmetric matrix:
 C12 +C21
  1 ∂T 
1 ∂T 1 ∂u C11 − T ∂x
( − T ∂x u ∂x ) C21 +C12 2
1 ∂u ≥ 0. (5.13)
2
C 22 u ∂x

Now in a well known result from linear algebra, a necessary and sufficient condition for
satisfying this inequality is that the new coefficient matrix be positive semi-definite. Further,
the matrix will be positive semi-definite if it has positive semi-definite eigenvalues. The
eigenvalues of the new coefficient matrix can be shown to be
1 p 
λ= (C11 + C22 ) ± (C11 − C22 )2 + (C12 + C21 )2 . (5.14)
2

© 05 December 2025, Joseph M. Powers. All rights reserved.

164 CHAPTER 5. CONSTITUTIVE EQUATIONS

Because the terms inside the radical are positive semi-definite, the eigenvalues must be real.
This is a consequence of the parent matrix being symmetric. Now we require two positive
semi-definite eigenvalues. First, if C11 + C22 < 0, we obviously have at least one negative
eigenvalue, so we demand that C11 + C22 ≥ 0. We then must have
p
C11 + C22 ≥ (C11 − C22 )2 + (C12 + C21 )2 . (5.15)

This gives rise to

(C11 + C22 )2 ≥ (C11 − C22 )2 + (C12 + C21 )2 . (5.16)
Expanding and simplifying, one gets
 2
C12 + C21
C11 C22 ≥ . (5.17)
2

Now the right side is positive semi-definite, so the left side must be also. Thus

C11 C22 ≥ 0. (5.18)

The only way for the sum and product of C11 and C22 to be positive semi-definite is to
demand that C11 ≥ 0 and C22 ≥ 0. Thus we arrive at the final set of conditions to satisfy
the second law:

C11 ≥ 0, (5.19)
C22 ≥ 0, (5.20)
 2
C12 + C21
C11 C22 ≥ . (5.21)
2

Now an important school of thought, founded by Onsager1 in twentieth century thermo-

dynamics takes an extra step and makes the further assertion that the original matrix Cij
itself must be symmetric. That is C12 = C21 . This remarkable assertion is independent of the
second law, and is, for other scenarios, consistent with experimental results. Consequently,
the second law in combination with Onsager’s independent demand, requires that

C11 ≥ 0, (5.22)
C22 ≥ 0, (5.23)
p
C12 ≤ C11 C22 . (5.24)

All this said, we must dismiss our hypothesis in this specific case on other physical
grounds, namely that such a hypothesis results in an infinite shear stress for a fluid at rest!
In the special case in which ∂T /∂x = 0, our hypothesis predicts τ = C22 (T /u2 )(∂u/∂x).
1
Lars Onsager, 1903-1976, Norwegian-born American physical chemist, earned Ph.D. and taught at Yale,
developed a systematic theory for irreversible chemical processes.

© 05 December 2025, Joseph M. Powers. All rights reserved.

5.2. SECOND LAW RESTRICTIONS AND ONSAGER RELATIONS 165

Obviously this is inconsistent with any observation and so we reject this hypothesis. Addi-
tionally, this assumed form is not frame invariant because of the velocity dependency. So,
why did we go to this trouble? First, we now have confidence that we should not expect
to find heat flux to depend on deformation. Second, it illustrates some general techniques
in continuum mechanics. Moreover, the techniques we used have actually been applied to
other more complex phenomena that are physical, and of great practical importance.

[Link] Real physical effects

That such a matrix such as we studied in the previous section was asserted to be symmetric is
a manifestation of what is known as a general Onsager relation, developed by Onsager in 1931
with a statistical mechanics basis for more general systems and for which he was awarded
a Nobel Prize in chemistry in 1968. These actually describe a surprising variety of physical
phenomena, and are described in detail many texts, including Fung and Woods. A well-
known example is the Peltier2 effect in which conduction of both heat and electrical charge
is influenced by gradients of charge and temperature. This forms the basis of the operation of
a thermocouple. Other relations exist are the Soret3 effect in which diffusive mass fluxes are
induced by temperature gradients, the Dufour effect in which a diffusive energy flux is induced
by a species concentration gradient, the Hall4 effect for coupled electrical and magnetic effects
(that explains the operation of an electric motor), the Seeback5 effect in which electromotive
forces are induced by different conducting elements at different temperatures, the Thomson6
effect in which heat is transferred when electric current flows in a conductor in which there
is a temperature gradient, and the principle of detailed balance for multi-species chemical
reactions.

5.2.2 Strong form of the Clausius-Duhem inequality

A less general way to satisfy the second law is to take the sufficient (but not necessary!)
condition that each individual term in the entropy inequality to be greater than or equal to
zero:
1
− qi ∂i T ≥ 0, and (5.25)
T2
1
τij ∂(i vj) ≥ 0. (5.26)
T
2
Jean Charles Athanase Peltier, 1785-1845, French clockmaker, retired at 30 to study science.
3
Charles Soret, 1854-1904, Swiss physicist and chemist.
4
Edwin Herbert Hall, 1855-1938, Maine-born American physicist, educated at Johns Hopkins University
where he discovered the Hall effect while working on his dissertation, taught at Harvard.
5
Thomas Johann Seebeck, 1770-1831, German medical doctor who studied at Berlin and Göttingen.
6
William Thomson (Lord Kelvin), 1824-1907, Belfast-born British mathematician and physicist, grad-
uated and taught at Glasgow University, key figure in all of 19th century engineering science including
mathematics, thermodynamics, and electrodynamics.

© 05 December 2025, Joseph M. Powers. All rights reserved.

166 CHAPTER 5. CONSTITUTIVE EQUATIONS

Once again, this is called the strong form of the entropy inequality (or the strong form of
the Clausius-Duhem inequality), and is potentially overly restrictive.

5.3 Fourier’s law

Let us examine the restriction on qi from the strong form of the entropy inequality to infer
the common constitutive relation known as Fourier’s law.7 The portion of the strong form
of the entropy inequality with which we are concerned here is Eq. (5.25):

1
− qi ∂i T ≥ 0. (5.27)
T2
Now one way to guarantee this inequality is satisfied is to specify the constitutive relation
for the heat flux vector as

qi = −k∂i T, with k ≥ 0. (5.28)

This is the well known Fourier’s law for an isotropic material, where k is the thermal con-
ductivity. It has the proper behavior under Galilean transformations and rotations; more
importantly, it is consistent with macroscale experiments for isotropic materials and can be
justified from an underlying microscale theory. Substitution of Fourier’s law for an isotropic
material into the strong form of entropy inequality, Eq. (5.25), yields

1
k(∂i T )(∂i T ) ≥ 0, (5.29)
T2
that for k ≥ 0 is a true statement. The second law allows other forms as well. The expression
qi = −k((∂j T )(∂j T ))∂i T is consistent with the second law. It does not match experiments
well for most materials however.
Following Duhamel,8 we can also generalize Fourier’s law for an anisotropic material. Let
us only consider anisotropic materials for which the conductivity in any given direction is
a constant. For such materials, the thermal conductivity is a tensor kij , and Fourier’s law
generalizes to
qi = −kij ∂j T. (5.30)
This effectively states that for a fixed temperature gradient, the heat flux depends on the
orientation. This is characteristic of anisotropic substances such as layered materials. Sub-
stitution of the generalized Fourier’s law into the strong form of the entropy inequality,
7
Jean Baptiste Joseph Fourier, 1768-1830, French mathematician and Egyptologist who studied the trans-
fer of heat and the representation of mathematical functions by infinite series summations of other functions.
Son of a tailor.
8
Jean Marie Constant Duhamel, 1797-1872, highly regarded mathematics teacher at École Polytechnique
in Paris who applied mathematics to problems in heat transfer, mechanics, and acoustics.

© 05 December 2025, Joseph M. Powers. All rights reserved.

5.3. FOURIER’S LAW 167

Eq. (5.25), gives now

1
kij (∂j T )(∂i T ) ≥ 0, (5.31)
T2
1
(∂i T )kij (∂j T ) ≥ 0. (5.32)
T2
Now 1/T 2 > 0, so we must have (∂i T )kij (∂j T ) ≥ 0 for all possible values of ∇T . Now any
possible anti-symmetric portion of kij cannot contribute to the inequality. We can see this
by expanding kij in the entropy inequality to get
 
1 1
∂i T (kij + kji ) + (kij − kji ) ∂j T ≥ 0, (5.33)
2 2


∂i T k(ij) + k[ij] ∂j T ≥ 0, (5.34)
(∂i T )k(ij) (∂j T ) + (∂i T )k[ij](∂j T ) ≥ 0, (5.35)
| {z }
=0
(∂i T )k(ij) (∂j T ) ≥ 0. (5.36)

The anti-symmetric part of kij makes no contribution to the entropy generation because it
involves the tensor inner product of a symmetric tensor with an anti-symmetric tensor, that
is identically zero.
Next, we again use the well-known result from linear algebra that the entropy inequality
is satisfied if k(ij) is a positive semi-definite tensor. This will be the case if all the eigenvalues
of k(ij) are non-negative. Let us recast the second law for this system, Eq. (5.36), in Gibbs
form:

(∇T )T · Ks · ∇T ≥ 0, (5.37)

where Ks = k(ij) , the symmetric part of the conductivity tensor. Because Ks is symmetric,
it has a diagonalization of the form equivalent to that of Eq. (3.141):

Ks = Q · Λ · QT , (5.38)

where Q is a rotation matrix, and Λ is a matrix of eigenvalues of Ks that are guaranteed to

be real. Thus the second law, Eq. (5.37) can be recast as

(∇T )T · Q · Λ · QT · ∇T ≥ 0, (5.39)
(QT · ∇T )T · Λ · (QT · ∇T ) ≥ 0. (5.40)

We recognize this form and the fact that it can be true iff each of the eigenvalues that reside
in the diagonal matrix Λ are positive semi-definite.
Further now, when we consider the contribution of the heat flux vector to the energy
equation, we see any possible anti-symmetric portion of the conductivity tensor will be

© 05 December 2025, Joseph M. Powers. All rights reserved.

168 CHAPTER 5. CONSTITUTIVE EQUATIONS

inconsequential as well. This is seen by the following analysis, that considers only relevant
terms in the energy equation
de
ρ = −∂i qi + . . . , (5.41)
dt
= ∂i (kij ∂j T ) + . . . , (5.42)
= kij ∂i ∂j T + . . . , (5.43)


= k(ij) + k[ij] ∂i ∂j T + . . . , (5.44)
= k(ij) ∂i ∂j T + k[ij]∂i ∂j T + . . . , (5.45)
| {z }
=0
= k(ij) ∂i ∂j T + . . . . (5.46)
So, it seems any possible anti-symmetric portion of kij will have no consequence as far
as the first or second laws are concerned. However, an anti-symmetric portion of kij would
induce a heat flux orthogonal to the direction of the temperature gradient. In a remarkable
confirmation of Onsager’s principle, experimental measurements on anisotropic crystalline
materials demonstrate that there is no component of heat flux orthogonal to the temperature
gradient, and thus, the conductivity matrix kij in fact has zero anti-symmetric part, and thus
is symmetric, kij = kji . For our particular case with a tensorial conductivity, the competing
effects are the heat fluxes in three directions, caused by temperature gradients in three
directions:     
q1 k11 k12 k13 ∂1 T
 q2  = −  k21 k22 k23   ∂2 T  . (5.47)
q3 k31 k32 k33 ∂3 T
The symmetry condition, Onsager’s principle, requires that k12 = k21 , k13 = k31 , and k23 =
k32 . So, the experimentally verified Onsager’s principle further holds that the heat flux for
an anisotropic material is given by
    
q1 k11 k12 k13 ∂1 T
 q2  = −  k12 k22 k23   ∂2 T  . (5.48)
q3 k13 k23 k33 ∂3 T
Now it is well known that the conductivity matrix kij will be positive semi-definite if all
its eigenvalues are non-negative. The eigenvalues will be guaranteed real upon adopting
Onsager symmetry. The characteristic polynomial for the eigenvalues is given by
(1) (2) (3)
λ3 − Ik λ2 + Ik λ − Ik = 0, (5.49)
where the invariants of the conductivity tensor kij , are given by the standard
(1)
Ik = kii = tr K, (5.50)
(2) 1

Ik = (kii kjj − kij kji ) = (det K) tr K−1 , (5.51)
2
(3)
Ik = ǫijk k1j k2j k3j = det K. (5.52)

© 05 December 2025, Joseph M. Powers. All rights reserved.

5.3. FOURIER’S LAW 169

In a standard result from linear algebra, one can show that if all three invariants are positive
semi-definite, then the eigenvalues are all positive semi-definite, and as a result, the matrix
itself is positive semi-definite. Hence, in order for kij to be positive semi-definite we demand
that
(1)
Ik ≥ 0, (5.53)
(2)
Ik ≥ 0, (5.54)
(3)
Ik ≥ 0, (5.55)

that is equivalent to demanding that

k11 + k22 + k33 ≥ 0, (5.56)

2 2 2
k11 k22 + k11 k33 + k22 k33 − k12 − k13 − k23 ≥ 0, (5.57)
k13 (k12 k23 − k22 k13 ) + k23 (k12 k13 − k11 k23 ) + k33 (k11 k22 − k12 k12 ) ≥ 0. (5.58)

If det K 6= 0, the conditions reduce to

tr K ≥ 0, (5.59)
tr K−1 ≥ 0, (5.60)
det K > 0. (5.61)

Now by considering ∂i T = (1, 0, 0)T , and demanding (∂i T )kij (∂j T ) ≥ 0, we conclude that
k11 ≥ 0. Similarly, by considering ∂i T = (0, 1, 0)T and ∂i T = (0, 0, 1)T , we conclude that
k22 ≥ 0 and k33 ≥ 0, respectively. Thus tr K ≥ 0 is automatically satisfied. In equation
form, we then have

k11 ≥ 0, (5.62)
k22 ≥ 0, (5.63)
k33 ≥ 0, (5.64)
2 2 2
k11 k22 + k11 k33 + k22 k33 − k12 − k13 − k23 ≥ 0, (5.65)
k13 (k12 k23 − k22 k13 ) + k23 (k12 k13 − k11 k23 ) + k33 (k11 k22 − k12 k12 ) ≥ 0. (5.66)

While by no means a proof, numerical experimentation gives strong indication that the
remaining conditions can be satisfied if, loosely stated, k11 , k22 , k33 ≫ |k12 |, |k23|, |k13 |. That
is, for positive semi-definiteness,

• each diagonal element must be positive semi-definite,

• off-diagonal terms can be positive or negative, and

• diagonal terms must have amplitudes that are, loosely speaking, larger than the am-
plitudes of off-diagonal terms.

© 05 December 2025, Joseph M. Powers. All rights reserved.

170 CHAPTER 5. CONSTITUTIVE EQUATIONS

Example 5.1
Let us consider heat conduction in the limit of two dimensions and a constant anisotropic conduc-
tivity tensor, without imposing Onsager’s conditions.

Let us take then     

q1 k11 k12 ∂1 T
=− . (5.67)
q2 k21 k22 ∂2 T
The second law demands that
  
k11 k12 ∂1 T
( ∂1 T ∂2 T ) ≥ 0. (5.68)
k21 k22 ∂2 T
This is expanded as
 k12 +k21   k12 −k21   
k11 2 0 2 ∂1 T
( ∂1 T ∂2 T ) k21 +k12 + k21 −k12 ≥ 0. (5.69)
2 k22 2 0 ∂2 T
As before, the anti-symmetric portion makes no contribution to the left hand side, giving rise to
 k12 +k21   
k11 2 ∂1 T
( ∂1 T ∂2 T ) k21 +k12 ≥ 0. (5.70)
2 k22 ∂2 T
And, demanding that the eigenvalues of the symmetric part of the conductivity tensor be positive gives
rise to the conditions, identical to that of an earlier analysis, that
k11 ≥ 0, (5.71)
k22 ≥ 0, (5.72)
 2
k12 + k21
k11 k22 ≥ . (5.73)
2
The energy equation becomes
de
ρ = −∂i qi + . . . , (5.74)
dt   
k11 k12 ∂1 T
= ( ∂1 ∂2 ) + ..., (5.75)
k21 k22 ∂2 T
 
k11 ∂1 T + k12 ∂2 T
= ( ∂1 ∂2 ) + ..., (5.76)
k21 ∂1 T + k22 ∂2 T
= k11 ∂1 ∂1 T + (k12 + k21 )∂1 ∂2 T + k22 ∂2 ∂2 T + . . . , (5.77)
∂2T ∂2T ∂2T
= k11 2 + (k12 + k21 ) + k22 2 + . . . . (5.78)
∂x1 ∂x1 ∂x2 ∂x2
One sees that the energy evolution depends only on the symmetric part of the conductivity tensor.
Imposition of Onsager’s relations gives simply k12 = k21 , giving rise to second law restrictions of
k11 ≥ 0, (5.79)
k22 ≥ 0, (5.80)
2
k11 k22 ≥ k12 , (5.81)
and an energy equation of
de ∂2T ∂2T ∂2T
ρ = k11 2 + 2k12 + k22 2 + . . . . (5.82)
dt ∂x1 ∂x1 ∂x2 ∂x2

© 05 December 2025, Joseph M. Powers. All rights reserved.

5.3. FOURIER’S LAW 171

Example 5.2
Consider the ramifications of a heat flux vector in violation of Onsager’s principle: flux in which
the anisotropic conductivity is purely anti-symmetric. For simplicity consider an incompressible solid
with constant specific heat cp . For the heat flux, we take
    
q1 0 −β ∂1 T
=− . (5.83)
q2 β 0 ∂2 T

This holds that heat flux in the 1 direction is induced only by temperature gradients in the 2 direction
and heat flux in the 2 direction is induced only by temperature gradients in the 1 direction.

The second law demands that

  
−β
0 ∂1 T
( ∂1 T ∂2 T ) ≥ 0, (5.84)
β0 ∂2 T
 
−β∂2 T
( ∂1 T ∂2 T ) ≥ 0, (5.85)
β∂1 T
−β(∂1 T )(∂2 T ) + β(∂1 T )(∂2 T ) ≥ 0, (5.86)
0 ≥ 0. (5.87)

So, the second law holds.

For the incompressible solid with constant heat capacity, the velocity field is zero, and the energy
equation reduces to the simple
∂T
ρcp = −∂i qi . (5.88)
∂t
Imposing our unusual expression for heat flux, we get
  
∂T 0 −β ∂1 T
ρcp = ( ∂1 ∂2 ) , (5.89)
∂t β 0 ∂2 T
 
−β∂2 T
= ( ∂1 ∂2 ) , (5.90)
β∂1 T
= −β∂1 ∂2 T + β∂1 ∂2 T, (5.91)
= 0. (5.92)

So, this unusual heat flux vector is one that induces no change in temperature. In terms of the first law
of thermodynamics, a net energy flux into a control volume in the 1 direction is exactly counterbalanced
by an net energy flux out of the same control volume in the 2 direction. Thus the first law holds as
well.
Let us consider a temperature distribution for this unusual material. And let us consider it to apply
to the domain x ∈ [0, 1], y ∈ [0, 1], t ∈ [0, ∞]. Take

T (x1 , x2 , t) = x2 . (5.93)

Obviously this satisfies the first law as ∂T /∂t = 0. Let us check the heat flux.

q1 = β∂2 T = β, (5.94)
q2 = −β∂1 T = 0. (5.95)

© 05 December 2025, Joseph M. Powers. All rights reserved.

172 CHAPTER 5. CONSTITUTIVE EQUATIONS

Now the lower boundary at x2 = 0 has T = 0. The upper boundary has x2 = 1 so T = 1. And this
constant temperature gradient in the 2 direction is inducing a constant heat flux in the 1 direction,
q1 = −β. The energy flux that enters at x1 = 0 departs at x1 = 1, maintaining energy conservation.
One can consider an equivalent problem in cylindrical coordinates. Taking

x1 = r cos θ, x2 = r sin θ, (5.96)

and applying the chain rule,     

∂r ∂θ ∂
∂1 ∂x1 ∂x1 ∂r
= ∂r ∂θ ∂ , (5.97)
∂2 ∂x2 ∂x2 ∂θ
one finds
    
∂1 cos θ − sinr θ ∂
∂r
= cos θ ∂ . (5.98)
∂2 sin θ r ∂θ

So, transforming q1 = β∂2 T , and q2 = −β∂1 T gives

   cos θ   ∂T 
q1 sin θ r ∂r
=β sin θ ∂T . (5.99)
q2 − cos θ r ∂θ

Standard trigonometry gives

    
qr cos θ sin θ q1
= . (5.100)
qθ − sin θ cos θ q2
| {z }
rotation matrix

Applying the rotation matrix to both sides gives then

     cos θ   ∂T 
cos θ sin θ q1 cos θ sin θ sin θ r ∂r
= β sin θ ∂T , (5.101)
− sin θ cos θ q2 − sin θ cos θ − cos θ r ∂θ
   1
  ∂T 
qr 0 r ∂r
= β ∂T , (5.102)
qθ −1 0 ∂θ

or simply
β ∂T
qr = , (5.103)
r ∂θ
∂T
qθ = −β . (5.104)
∂r
Now the steady state temperature distribution in the annular region 1/2 < r < 1, T = r, describes
a domain with an inner boundary held at T = 1/2 and an outer boundary held at T = 1. Such a
temperature distribution would induce a heat flux in the θ direction only, so that qr = 0 and qθ = −β.
That is, the heat goes round and round the domain, but never enters or exits at any boundary.
Now such a flux is counterintuitive precisely because it has never been observed or measured. It is
for this reason that we can adopt Onsager’s hypothesis and demand that, independent of the first and
second laws of thermodynamics,
β = 0, (5.105)
and the conductivity tenser is purely symmetric.

© 05 December 2025, Joseph M. Powers. All rights reserved.

5.4. STRESS-STRAIN RATE RELATION FOR A NEWTONIAN FLUID 173

cross sectional area A

F

Figure 5.1: Sketch of simple Couette flow experiment with measurements of stress versus
strain rate.

5.4 Stress-strain rate relation for a Newtonian fluid

We now seek to satisfy the second part of the strong form of the entropy inequality, Eq. (5.26).
Recalling that T > 0, this reduces to

τij ∂(i vj) ≥ 0. (5.106)

| {z }
Φ
This form suggests that we seek a constitutive equation for the viscous stress tensor τij
that is a function of the deformation tensor ∂(i vj) . Fortunately, such a form exists, that
moreover agrees with macroscale experiments and microscale theories. Here we will focus on
the simplest of such theories, for what is known as a Newtonian fluid, a fluid that is isotropic
and whose viscous stress varies linearly with strain rate. In general, this is a discipline unto
itself known as rheology.

5.4.1 Underlying experiments

We can pull a flat plate over a fluid and measure the force necessary to maintain a specified
velocity. This situation and some expected results are sketched in Fig. 5.1. We observe that

© 05 December 2025, Joseph M. Powers. All rights reserved.

174 CHAPTER 5. CONSTITUTIVE EQUATIONS

• At the upper and lower plate surfaces, the fluid has the same velocity of each plate.
This is called the no-slip condition.

• The faster the velocity v of the upper plate is, the higher the force necessary to pull
the plate is. The increase can be linear or non-linear.

• When experiments are carried out with different plate area and different gap width, a
single universal curve results when F/A is plotted against v/h.

• The velocity profile is linear with increasing x2 .

In a way similar on a molecular scale to energy diffusion, this experiment is describing

a diffusion of momentum from the pulled plate into the fluid below it. The constitutive
equation we develop for viscous stress, when combined with the governing axioms, will
model momentum diffusion.
We can associate F/A with a shear stress: τ21 , recalling stress on the 2 face in the 1
direction. We can associate v/h with a velocity gradient, here ∂2 v1 . We note that considering
the velocity gradient is essentially equivalent to considering the deformation gradient, as far
as the second law is concerned, and so we will be loose here in our use of the term. We
loosely define the coefficient of viscosity µ for this configuration as
τ21 viscous stress
µ= = . (5.107)
∂2 v1 strain rate
We shall see in the next section that µ is better described as the first coefficient of viscosity
and that a second coefficient is also required to fully characterize an isotropic fluid. The
viscosity is the analog of Young’s9 modulus in solid mechanics, that is the ratio of stress to
strain. In general µ is a thermodynamic property of a material. It is often a strong function
of temperature, but can vary with pressure as well. A Newtonian fluid has a viscosity
that does not depend on strain rate (but could depend on temperature and pressure). A
non-Newtonian fluid has a viscosity that is strain rate-dependent (and possible temperature
and pressure). Some typical behavior is sketched in Fig. 5.2. We shall focus here on fluids
whose viscosity is not a function of strain rate. Much of our development will be valid
for temperature- and pressure-dependent viscosity, while most actual examples will consider
only constant viscosity.

5.4.2 Analysis for isotropic Newtonian fluid

Here we shall outline the method described by Whitaker (pp. 139-145) to describe the viscous
stress as a function of strain rate for an isotropic fluid with constant viscosity. An isotropic
fluid has no directional dependencies when subjected to a force. A fluid composed of aligned
9
Thomas Young, 1773-1829, English physician and physicist whose experiments in interferometry revived
the wave theory of light, Egyptologist who helped decipher the Rosetta stone, worked on surface tension in
fluids, gave the word “energy” scientific significance, and developed Young’s modulus in elasticity.

© 05 December 2025, Joseph M. Powers. All rights reserved.

5.4. STRESS-STRAIN RATE RELATION FOR A NEWTONIAN FLUID 175

dilatant
Bingham
plastic

Newtonian

1
pseudo-plastic

Figure 5.2: Variation of viscous stress with strain rate for typical fluids.

long chain polymers is an example of a fluid that is most likely not isotropic. Following
Whitaker, we

• postulate that stress is a function of deformation rate (strain rate) only:10

τij = fij (∂(k vl) ). (5.108)

Written out in more detail, we have postulated a relationship of the form

τ11 = f11 (∂(1 v1) , ∂(2 v2) , ∂(3 v3) , ∂(1 v2) , ∂(2 v3) , ∂(3 v1) ∂(2 v1) , ∂(3 v2) , ∂(1 v3) ), (5.109)
τ12 = f12 (∂(1 v1) , ∂(2 v2) , ∂(3 v3) , ∂(1 v2) , ∂(2 v3) , ∂(3 v1) ∂(2 v1) , ∂(3 v2) , ∂(1 v3) ), (5.110)
..
.
τ33 = f33 (∂(1 v1) , ∂(2 v2) , ∂(3 v3) , ∂(1 v2) , ∂(2 v3) , ∂(3 v1) ∂(2 v1) , ∂(3 v2) , ∂(1 v3) ). (5.111)

• require that τij = 0 if ∂(i vj) = 0; hence, no strain rate, no stress.

• require that stress is linearly related to strain rate:

τij = Ĉijkl ∂(k vl) . (5.112)

This is the imposition of the assumption of a Newtonian fluid. Here Ĉijkl is a fourth
10
Thus, we are not allowing viscous stress to be a function of the rigid body rotation rate. While it seems
intuitive that rigid body rotation should not induce viscous stress, Batchelor (2000) mentions that there is
no rigorous proof for this; hence, we describe our statement as a postulate.

© 05 December 2025, Joseph M. Powers. All rights reserved.

176 CHAPTER 5. CONSTITUTIVE EQUATIONS

order tensor. It can be considered as a matrix of matrices:

     
Ĉ1111 Ĉ1112 Ĉ1113 Ĉ1211 Ĉ1212 Ĉ1213 Ĉ1311 Ĉ1312 Ĉ1313
  Ĉ1121 Ĉ1122 Ĉ1123   Ĉ1221 Ĉ1222 Ĉ1223   Ĉ1321 Ĉ1322 Ĉ1323  
 
 Ĉ1131 Ĉ1132 Ĉ1133 Ĉ1333  
   Ĉ1231 Ĉ1232 Ĉ1233   Ĉ1331 Ĉ1332 
 Ĉ2111 Ĉ2112 Ĉ2113 Ĉ2211 Ĉ2212 Ĉ2213 Ĉ2311 Ĉ2312 Ĉ2313 
 
Ĉijkl =  
 Ĉ2121 Ĉ21222 Ĉ2123
  Ĉ2221 Ĉ2222 Ĉ2223   Ĉ2321 Ĉ2322 Ĉ2323  
.
 Ĉ2131 Ĉ2132 Ĉ2133 Ĉ2333  

 Ĉ3111 Ĉ3112 Ĉ3113   Ĉ2231 Ĉ2232 Ĉ2233   Ĉ2331 Ĉ2332 
 Ĉ3211 Ĉ3212 Ĉ3213 Ĉ3311 Ĉ3312 Ĉ3313  
  Ĉ3121 Ĉ3122 Ĉ3123   Ĉ3221 Ĉ3222 Ĉ3223   Ĉ3321 Ĉ3322 Ĉ3323  
Ĉ3131 Ĉ3132 Ĉ3133 Ĉ3231 Ĉ3232 Ĉ3233 Ĉ3331 Ĉ3332 Ĉ3333
(5.113)

In what is known as a flattening process, we can recast the 3 × 3 τij as a 9 × 1 vector,

the 3 × 3 ∂(i vj) as a 9 × 1 vector, and the 3 × 3 × 3 × 3 Ĉijkl as a 9 × 9 matrix. Thus
we have in matrix form
    
τ11 Ĉ1111 Ĉ1122 Ĉ1133 Ĉ1112 Ĉ1123 Ĉ1131 Ĉ1121 Ĉ1132 Ĉ1113 ∂(1 v1)
 τ22   Ĉ2211 Ĉ2222 Ĉ2233 Ĉ2212 Ĉ2223 Ĉ2231 Ĉ2221 Ĉ2232 Ĉ2213   
     ∂(2 v2) 
 τ33   Ĉ3311 Ĉ3322 Ĉ3333 Ĉ3312 Ĉ3323 Ĉ3331 Ĉ3321 Ĉ3332  
Ĉ3313   ∂(3 v3) 
   
 τ12   Ĉ1211 Ĉ1222 Ĉ1233 Ĉ1212 Ĉ1223 Ĉ1231 Ĉ1221 Ĉ1232 Ĉ1213   
     ∂(1 v2) 
 τ23  =  Ĉ2311 Ĉ2322 Ĉ2333 Ĉ2312 Ĉ2323 Ĉ2331 Ĉ2321 Ĉ2332  
Ĉ2313   ∂(2 v3) 
   .
 τ31   Ĉ3111 Ĉ3122 Ĉ3133 Ĉ3112 Ĉ3123 Ĉ3131 Ĉ3121 Ĉ3132 Ĉ3113   ∂(3 v1) 
    
 τ21   Ĉ2111 Ĉ2122 Ĉ2133 Ĉ2112 Ĉ2123 Ĉ2131 Ĉ2121 Ĉ2132 Ĉ2113   
     ∂(2 v1) 
 τ32   Ĉ3211 Ĉ3222 Ĉ3233 Ĉ3212 Ĉ3223 Ĉ3231 Ĉ3221 Ĉ3232 Ĉ3213   ∂(3 v2) 
τ13 Ĉ1311 Ĉ1322 Ĉ1333 Ĉ1312 Ĉ1323 Ĉ1331 Ĉ1321 Ĉ1332 Ĉ1313 ∂(1 v3)
(5.114)
There are 34 = 81 unknown coefficients Ĉijkl . We found one of them in our simple
experiment in which we found

τ21 = τ12 = µ∂2 v1 = µ(2∂(1 v2) ). (5.115)

Hence in this special case Ĉ1212 = 2µ.

Now we could do eighty-one separate experiments, or we could take advantage of the
assumption that the fluid has no directional dependency. We will take the following approach.
Observer A conducts an experiment to measure the stress tensor in reference frame A. The
observer begins with the “viscosity matrix” Ĉijkl . The experiment is conducted by varying
strain rate and measuring stress. With complete knowledge A feels confident this knowledge
could be used to predict the stress in rotated frame A′.
Consider observer A′ who is oriented to frame A′ . Oblivious to observer A, A′ conducts
the same experiment to measure what for her or him is τij′ The value that A′ measures must
be the same that A predicts in order for the system to be isotropic. This places restrictions
on the viscosity matrix Ĉijkl . We intend to show that if the fluid is isotropic, only two of the
eighty-one coefficients are distinct and non-zero.

© 05 December 2025, Joseph M. Powers. All rights reserved.

5.4. STRESS-STRAIN RATE RELATION FOR A NEWTONIAN FLUID 177

We first use symmetry properties of the stress and strain rate tensor to reduce to thirty-six
unknown coefficients. We note that in actuality there are only six independent components of
stress and six independent components of deformation because both are symmetric tensors.
Consequently, we can write our linear stress-strain rate relation as
    
τ11 Ĉ1111 Ĉ1122 Ĉ1133 Ĉ1112 + Ĉ1121 Ĉ1123 + Ĉ1132 Ĉ1131 + Ĉ1113 ∂(1 v1)
 τ22   Ĉ2211 Ĉ2222 Ĉ2233 Ĉ2212 + Ĉ2221 Ĉ2223 + Ĉ2232 Ĉ2231 + Ĉ2213   
     ∂(2 v2) 
 τ33   Ĉ3311 Ĉ3322 Ĉ3333 Ĉ3312 + Ĉ3321 Ĉ3323 + Ĉ3332  
Ĉ3331 + Ĉ3313   ∂(3 v3) 
   .
 τ12  =  Ĉ1211 Ĉ1222 Ĉ1233 Ĉ1212 + Ĉ1221 Ĉ1223 + Ĉ1232 Ĉ1231 + Ĉ1213   
     ∂(1 v2) 
 τ23   Ĉ2311 Ĉ2322 Ĉ2333 Ĉ2312 + Ĉ2321 Ĉ2323 + Ĉ2332 Ĉ2331 + Ĉ2313   ∂(2 v3) 
τ31 Ĉ3111 Ĉ3122 Ĉ3133 Ĉ3112 + Ĉ3121 Ĉ3123 + Ĉ3132 Ĉ3131 + Ĉ3113 ∂(3 v1)
(5.116)
Now adopting Whitaker’s notation for simplification, we define this matrix of Ĉ’s as a new
matrix of C’s. Here, now C itself is not a tensor, while Ĉ is a tensor. We take equivalently
then     
τ11 C11 C12 C13 C14 C15 C16 ∂(1 v1)
 τ22   C21 C22 C23 C24 C25 C26   ∂(2 v2) 
    
 τ33   C31 C32 C33 C34 C35 C36   ∂(3 v3) 
    
 τ12  =  C41 C42 C43 C44 C45 C46   ∂(1 v2)  . (5.117)
    
 τ23   C51 C52 C53 C54 C55 C56   ∂(2 v3) 
τ31 C61 C62 C63 C64 C65 C66 ∂(3 v1)
This less standard formulation of stress and strain rate is a type of Voigt11 notation. Further
background is given by Ting (1996, p. 35-64).
Next, recalling that for tensorial quantities

τij′ = ℓki ℓlj τkl , (5.118)

∂(i′ vj)
′
= ℓki ℓlj ∂(k vl) , (5.119)

let us subject our fluid to a battery of rotations and see what can be concluded by enforcing
material indifference.
• 180◦ rotation about x3 axis
For this rotation, sketched in Fig. 5.3. we have direction cosines
 
ℓ11 = −1 ℓ12 = 0 ℓ13 = 0
ℓki =  ℓ21 = 0 ℓ22 = −1 ℓ23 = 0  . (5.120)
ℓ31 = 0 ℓ32 = 0 ℓ33 = 1

So, applying Eq. (2.10), we see

 
−1 0 0
( x′1 x′2 x′3 ) = ( x1 x2 x3 )  0 −1 0  , (5.121)
0 0 1
11
Woldemar Voigt, 1850-1919, German mathematician.

© 05 December 2025, Joseph M. Powers. All rights reserved.

178 CHAPTER 5. CONSTITUTIVE EQUATIONS

Figure 5.3: Rotation of 180◦ about x3 axis.

that yields
x′1 = −x1 , x′2 = −x2 , x′3 = x3 , (5.122)

that is consistent with Fig. 5.3. Because det ℓki = 1, the transformation is a rotation.
Applying the transformation rules to each term in the shear stress tensor, we get

′
τ11 = ℓk1 ℓl1 τkl = (−1)2 τ11 = τ11 , (5.123)
′
τ22 = ℓk2 ℓl2 τkl = (−1)2 τ22 = τ22 , (5.124)
′
τ33 = ℓk3 ℓl3 τkl = (1)2 τ33 = τ33 , (5.125)
′
τ12 = ℓk1 ℓl2 τkl = (−1)2 τ12 = τ12 , (5.126)
′
τ23 = ℓk2 ℓl3 τkl = (−1)(1)τ23 = −τ23 , (5.127)
′
τ31 = ℓk3 ℓl1 τkl = (1)(−1)τ31 = −τ31 . (5.128)

We could find the same in Gibbs notation. With the rotation matrix of
 
−1 0 0
ℓki = Q =  0 −1 0  , (5.129)
0 0 1

we have from Eq. (2.61) that

τ ′ = QT · τ · Q, (5.130)
   
−1 0 0 τ11 τ12 τ31 −1 0 0
=  0 −1 0   τ12 τ22 τ23   0 −1 0  , (5.131)
0 0 1 τ31 τ23 τ33 0 0 1
 
τ11 τ12 −τ31
=  τ12 τ22 −τ23  . (5.132)
−τ31 −τ23 τ33

© 05 December 2025, Joseph M. Powers. All rights reserved.

5.4. STRESS-STRAIN RATE RELATION FOR A NEWTONIAN FLUID 179

Likewise, for strain rate we find that

′ ′
∂(1 v1) = ∂(1 v1) , (5.133)
′ ′
∂(2 v2) = ∂(2 v2) , (5.134)
′ ′
∂(3 v3) = ∂(3 v3) , (5.135)
′ ′
∂(1 v2) = ∂(1 v2) , (5.136)
′ ′
∂(2 v3) = −∂(2 v3) , (5.137)
′ ′
∂(3 v1) = −∂(3 v1) . (5.138)

Now our observer A′ who is in the rotated system would say, for instance that
′ ′ ′ ′ ′ ′ ′ ′ ′ ′ ′ ′ ′
τ11 = C11 ∂(1 v1) + C12 ∂(2 v2) + C13 ∂(3 v3) + C14 ∂(1 v2) + C15 ∂(2 v3) + C16 ∂(3 v1) , (5.139)
′
while our observer A who used tensor algebra to predict τ11 would say
′ ′ ′ ′ ′ ′ ′ ′ ′ ′ ′ ′ ′
τ11 = C11 ∂(1 v1) + C12 ∂(2 v2) + C13 ∂(3 v3) + C14 ∂(1 v2) − C15 ∂(2 v3) − C16 ∂(3 v1) . (5.140)

Because we want both predictions to be the same, we must require that

C15 = C16 = 0. (5.141)

In matrix form, our observer A would predict for the rotated frame that
 ′    ′ ′ 
τ11 C11 C12 C13 C14 C15 C16 ∂(1 v1)
 τ22′   C21 C22 C23 
C24 C25 C26   ∂(2 ′ ′
v2) 
 ′   
 τ33   C31 C32 C33 C34 C35 C36   ′ ′ 
 ′ =   ∂(3 v3) . (5.142)
 τ   C41 C42 C43 C44 C45 C46   ∂(1 v2) 
  ′ ′
 12′   
 −τ23   C51 C52 C53 C54 C55 C56   −∂(2 ′ ′ 
v3)
′ ′ ′
−τ31 C61 C62 C63 C64 C65 C66 −∂(3 v1)

To retain material indifference between the predictions of our two observers, we thus
require that C15 = C16 = C25 = C26 = C35 = C36 = C45 = C46 = C51 = C52 = C53 =
C54 = C61 = C62 = C63 = C64 = 0. This eliminates 16 coefficients and gives our
viscosity matrix the form
 
C11 C12 C13 C14 0 0
 C21 C22 C23 C24 0 0 
 
 C31 C32 C33 C34 0 0 
 . (5.143)
 C41 C42 C43 C44 0 0 
 
 0 0 0 0 C55 C56 
0 0 0 0 C65 C66

with only 20 independent coefficients.

© 05 December 2025, Joseph M. Powers. All rights reserved.

180 CHAPTER 5. CONSTITUTIVE EQUATIONS

Figure 5.4: Rotation of 180◦ about x1 axis.

• 180◦ rotation about x1 axis

This rotation is sketched in Fig. 5.4. Applying Eq. (2.10), we see
 
1 0 0
′ ′ ′ 
( x1 x2 x3 ) = ( x1 x2 x3 ) 0 −1 0 , (5.144)
0 0 −1

that yields
x′1 = x1 , x′2 = −x2 , x′3 = −x3 , (5.145)
that is consistent with Fig. 5.4. Leaving out the rest of the details of the previous
section, this rotation has a set of direction cosines of
 
1 0 0
ℓij =  0 −1 0  . (5.146)
0 0 −1

Application of this rotation leads to the conclusion that the viscosity matrix must be
of the form  
C11 C12 C13 0 0 0
 C21 C22 C23 0 0 0 
 
 C31 C32 C33 0 0 0 
 . (5.147)
 0 0 0 C44 0 0 
 
 0 0 0 0 C55 0 
0 0 0 0 0 C66
with only 12 independent coefficients.

• 180◦ rotation about x2 axis

One is tempted to perform this rotation as well, but nothing new is learned from it!

© 05 December 2025, Joseph M. Powers. All rights reserved.

5.4. STRESS-STRAIN RATE RELATION FOR A NEWTONIAN FLUID 181

Figure 5.5: Rotation of 90◦ about x1 axis.

• 90◦ rotation about x1 axis

Having exhausted 180◦ rotations, let us turn to 90◦ rotations. We first rotate about
the x1 axis. This rotation is sketched in Fig. 5.5. This rotation has a set of direction
cosines of  
1 0 0
ℓij =  0 0 −1  . (5.148)
0 1 0
Because det ℓki = 1, the transformation is a rotation. Certainly also the length of each
column vector is unity, and each column vector dotted into other column vectors has
a value of zero, so the column vectors are orthonormal. Application of this rotation
leads to the conclusion that the viscosity matrix must be of the form
 
C11 C12 C12 0 0 0
 C21 C22 C23 0 0 0 
 
 C21 C23 C22 0 0 0 
 . (5.149)
 0 0 0 C44 0 0 
 
 0 0 0 0 C55 0 
0 0 0 0 0 C66

with only 8 independent coefficients.

• 90◦ rotation about x3 axis

This rotation is sketched in Fig. 5.6. This rotation has a set of direction cosines of
 
0 1 0
ℓij =  −1 0 0  . (5.150)
0 0 1

© 05 December 2025, Joseph M. Powers. All rights reserved.

182 CHAPTER 5. CONSTITUTIVE EQUATIONS

Figure 5.6: Rotation of 90◦ about x3 axis.

Because det ℓki = 1, the transformation is a rotation. Application of this rotation leads
to the conclusion that the viscosity matrix must be of the form
 
C11 C12 C12 0 0 0
 C12 C11 C12 0 0 0 
 
 C12 C12 C11 0 0 0 
 . (5.151)
 0 0 0 C 0 0 
 44 
 0 0 0 0 C44 0 
0 0 0 0 0 C44
with only 3 independent coefficients.
• 90◦ rotation about x2 axis
We learn nothing from this rotation.
• 45◦ rotation about x3 axis
This rotation is sketched in Fig. 5.7. This rotation has a set of direction cosines of
√ √ 
√2/2 −√ 2/2 0
ℓij =  2/2 2/2 0  . (5.152)
0 0 1
Because det ℓki = 1, the transformation is a rotation. After a lot of algebra, application
of this rotation leads to the conclusion that the viscosity matrix must be of the form
 
C44 + C12 C12 C12 0 0 0
 C12 C44 + C12 C12 0 0 0 
 
 C12 C12 C44 + C12 0 0 0 
 . (5.153)
 0 0 0 C 0 0 
 44 
 0 0 0 0 C44 0 
0 0 0 0 0 C44

© 05 December 2025, Joseph M. Powers. All rights reserved.

5.4. STRESS-STRAIN RATE RELATION FOR A NEWTONIAN FLUID 183

Figure 5.7: Rotation of 45◦ about x3 axis.

with only 2 independent coefficients.

Try as we might, we cannot reduce this any further with more rotations. It can be proved
more rigorously, as shown in most books on tensor analysis, that this is the furthest reduction
that can be made. So, for an isotropic Newtonian fluid, we can expect two independent
coefficients to parameterize the relation between strain rate and viscous stress. The relation
between stress and strain rate can be expressed in detail as


τ11 = C44 ∂(1 v1) + C12 ∂(1 v1) + ∂(2 v2) + ∂(3 v3) , (5.154)


τ22 = C44 ∂(2 v2) + C12 ∂(1 v1) + ∂(2 v2) + ∂(3 v3) , (5.155)


τ33 = C44 ∂(3 v3) + C12 ∂(1 v1) + ∂(2 v2) + ∂(3 v3) , (5.156)
τ12 = C44 ∂(1 v2) , (5.157)
τ23 = C44 ∂(2 v3) , (5.158)
τ31 = C44 ∂(3 v1) . (5.159)

Using traditional notation, we take

• C44 ≡ 2µ, where µ is the first coefficient of viscosity, and

• C12 ≡ λ, where λ is the second coefficient of viscosity.

There are a variety of other nomenclatures for µ and λ. Following Paolucci (2016), we can
also call µ the shear viscosity and λ the dilatational viscosity. We also can define the bulk
viscosity, ζ, as
2
ζ ≡ λ + µ, (5.160)
3

© 05 December 2025, Joseph M. Powers. All rights reserved.

184 CHAPTER 5. CONSTITUTIVE EQUATIONS

that is a term in common usage. A similar analysis in solid mechanics leads one to conclude
for an isotropic material in which the stress tensor is linearly related to the strain (rather
than the strain rate) gives rise to two independent coefficients, the elastic modulus and the
shear modulus. In solids, these both can be measured, and they are independent. In terms
of our original fourth order tensor, we can write the linear relationship τij = Ĉijkl∂(i vj) as
    
τ11 2µ + λ λ λ 0 0 0 0 0 0 ∂(1 v1)
 τ22   λ 2µ + λ λ 0 0 0 0 0 0   
     ∂(2 v2) 
 τ33   λ λ 2µ + λ 0 0 0 0 0  
0   ∂(3 v3) 
   
 τ12   0 0 0 2µ 0 0 0 0 0   
     ∂(1 v2) 
 τ23  =  0 0 0 0 2µ 0 0 0  
0   ∂(2 v3) 
   . (5.161)
 τ31   0 0 0 0 0 2µ 0 0 0   ∂(3 v1) 
    
 τ21   0 0 0 0 0 0 2µ 0 0   
     ∂(2 v1) 
 τ32   0 0 0 0 0 0 0 2µ 0   ∂(3 v2) 
τ13 0 0 0 0 0 0 0 0 2µ ∂(1 v3)

We note that because of the symmetry of ∂(i vj) that this representation is not unique in that
the following, as well as other linear combinations, is an identically equivalent statement:
    
τ11 2µ + λ λ λ 0 0 0 0 0 0 ∂(1 v1)
 τ22   λ 2µ + λ λ 0 0 0 0 0 0  
     ∂(2 v2) 
 τ33   λ λ 2µ + λ 0 0 0 0 0  
0   ∂(3 v3) 
   
 τ12   0 0 0 µ 0 0 µ 0 0  
     ∂(1 v2) 
 τ23  =  0 0 0 0 µ 0 0 µ 0  
     ∂(2 v3)  . (5.162)
 τ31   0 0 0 0 0 µ 0 0  
µ   ∂(3 v1) 
   
 τ21   0 0 0 µ 0 0 µ 0 0  
     ∂(2 v1) 
 τ32   0 0 0 0 µ 0 0 µ 0   ∂(3 v2) 
τ13 0 0 0 0 0 µ 0 0 µ ∂(1 v3)

In shorthand Cartesian index and Gibbs notation, the viscous stress tensor is given by

τij = 2µ∂(i vj) + λ∂k vk δij , (5.163)

 
∇vT + (∇vT )T
τ = 2µ + λ(∇T · v)I. (5.164)
2

By performing minor algebraic manipulations, the viscous stress tensor can be cast in
a way that elucidates more of the physics of how strain rate influences stress. It is easily
verified by direct expansion that the viscous stress tensor can be written as
   
 ∂k vk   1 
τij = (2µ + 3λ) δij  + 2µ ∂(i vj) − ∂k vk δij , (5.165)
3}
| {z | {z3 }
mean strain rate deviatoric strain rate
| {z } | {z }
mean viscous stress deviatoric viscous stress
 T   
∇ ·v ∇vT + (∇vT )T 1 T


τ = (2µ + 3λ) I + 2µ − ∇ ·v I . (5.166)
3 2 3

© 05 December 2025, Joseph M. Powers. All rights reserved.

5.4. STRESS-STRAIN RATE RELATION FOR A NEWTONIAN FLUID 185

Here it is seen that a mean strain rate, really a volumetric change, induces a mean viscous
stress, as long as λ 6= −(2/3)µ. If either λ = −(2/3)µ or ∂k vk = 0, all viscous stress is
deviatoric. Further, for µ 6= 0, a deviatoric strain rate induces a deviatoric viscous stress.
Eliminating λ in favor of the bulk viscosity ζ, we can say
 
1
τij = ζ∂k vk δij + 2µ ∂(i vj) − ∂k vk δij , (5.167)
3
 
T

∇vT + (∇vT )T 1 T


τ = ζ ∇ · v I + 2µ − ∇ ·v I , (5.168)
2 3
 
1
= ζϑI + 2µ D − ϑI . (5.169)
3

We can form the mean viscous stress by contracting the viscous stress tensor:
 
1 2
τii = µ + λ ∂k vk = ζ∂k vk = ζϑ. (5.170)
3 3

The mean viscous stress is a scalar, and is thus independent of orientation; it is directly
proportional to the first invariant of the viscous stress tensor. Obviously the mean viscous
stress is zero if λ = −(2/3)µ, that occurs if the bulk viscosity is zero. Now the total stress
tensor is given by

Tij = −pδij + 2µ∂(i vj) + λ∂k vk δij , (5.171)

 
∇vT + (∇vT )T
T = −pI + 2µ + λ(∇T · v)I. (5.172)
2

We notice the stress tensor has three components, 1) a uniform diagonal tensor with the
hydrostatic pressure, 2) a tensor that is directly proportional to the strain rate tensor, and
3) a uniform diagonal tensor that is proportional to the first invariant of the strain rate
(1)
tensor: Iǫ̇ = tr (∂(i vk) ) = ∂k vk = ϑ. Consequently, the stress tensor can be written as
 
(1)
Tij = −p + λIǫ̇ δij + 2µ∂(i vj) , (5.173)
| {z } | {z }
isotropic linear in strain rate
   
(1) ∇vT + (∇vT )T
T = −p + λIǫ̇ I + 2µ . (5.174)
2
(1)
Recalling that δij = I as well as Iǫ̇ are invariant under a rotation of coordinate axes, we
deduce that the stress is related linearly to the strain rate. Moreover when the axes are
rotated to be aligned with the principal axes of strain rate, the stress is purely normal stress
and takes on its principal value.
Let us next consider two typical elements to aid in interpreting the relation between
viscous stress and strain rate for a general Newtonian fluid.

© 05 December 2025, Joseph M. Powers. All rights reserved.

186 CHAPTER 5. CONSTITUTIVE EQUATIONS

[Link] Diagonal component

Consider a typical diagonal component of the viscous stress tensor, say τ11 :
   
 
 ∂1 v1 + ∂2 v2 + ∂3 v3  1
τ11 =  (2µ + 3λ)  + 2µ  
∂1 v1 − (∂1 v1 + ∂2 v2 + ∂3 v3 ) .(5.175)
 3  3
| {z } | {z }
deviatoric strain rate
| mean strain rate
{z } | {z }
mean viscous stress deviatoric viscous stress

If we choose our axes to be the principal axes of the strain-rate tensor, then these terms
will appear on the diagonal of the stress tensor and there will be no off-diagonal elements.
Thus, the fundamental physics of the stress-strain relationship are completely embodied in
a natural way in this expression.

[Link] Off-diagonal component

If we are not aligned with the principal axes, then off-diagonal terms will be non-zero. A
typical off-diagonal component of the viscous stress tensor, say τ12 , has the following form:
τ12 = 2µ∂(1 v2) + λ∂k vk δ12 , (5.176)
|{z}
=0
= 2µ∂(1 v2) , (5.177)
= µ(∂1 v2 + ∂2 v1 ). (5.178)
Note that this is associated with shear deformation for elements aligned with the 1 and 2
axes, and that it is independent of the value of λ, that is only associated with the mean
strain rate.

5.4.3 Stokes’ assumption

It is a straightforward matter to measure µ. It is not at all straightforward to measure λ. As
discussed earlier in Ch. 4.2.2, Stokes in the mid-nineteenth century suggested to require that
the mechanical pressure (that is the average normal stress) be equal to the thermodynamic
pressure; see Eq. (4.43) and the surrounding discussion. We have seen that the consequence
of this is Eq. (4.46): τii = 0. If we enforce this on our expression for τij , we get
τii = 0 = 2µ∂(i vi) + λ∂k vk δii , (5.179)
= 2µ∂i vi + 3λ∂k vk , (5.180)
= 2µ∂i vi + 3λ∂i vi , (5.181)
= (2µ + 3λ)∂i vi . (5.182)
Because in general ∂i vi 6= 0, Stokes’ assumption implies that
2
λ = − µ, iff Stokes’ assumption satisfied. (5.183)
3

© 05 December 2025, Joseph M. Powers. All rights reserved.

5.4. STRESS-STRAIN RATE RELATION FOR A NEWTONIAN FLUID 187

Stated another way, a fluid that satisfies Stokes’ assumption has a bulk viscosity of zero:

ζ = 0, iff Stokes’ assumption satisfied. (5.184)

So, a Newtonian fluid satisfying Stokes’ assumption has the following constitutive equation
for viscous stress
 
1
τij = 2µ ∂(i vj) − ∂k vk δij , (5.185)
3
| {z }
deviatoric strain rate
| {z }
deviatoric viscous stress
 
(∇vT + (∇vT )T ) 1 T
τ = 2µ − (∇ · v)I . (5.186)
2 3

Incompressible flows have ∂i vi = 0; thus, λ plays no role in determining the viscous stress
in such flows. For the fluid that obeys Stokes’ assumption, the viscous stress is entirely
deviatoric and is induced only by a deviatoric strain rate.
The full viscous stress tensor for the isotropic Newtonian fluid that satisfies Stokes’
assumption can be written as
       
∂v1 1 ∂v1 ∂v2 ∂v3 1 ∂v1 ∂v2 1 ∂v1 ∂v3
− + + + +
 ∂x1 3  ∂x1 ∂x2  ∂x3 2 ∂x2
 ∂x1
 2 ∂x3
 ∂x1
 
 1 ∂v2 ∂v1 ∂v2 1 ∂v1 ∂v2 ∂v3 1 ∂v2 ∂v3 
τ = 2µ  2 ∂x1
+ ∂x2 ∂x2
− 3 ∂x1 + ∂x2 + ∂x3 2 ∂x3
+ ∂x2 .
      
1 ∂v3 ∂v1 1 ∂v3 ∂v2 ∂v3 ∂v1 ∂v2 ∂v3
2 ∂x1
+ ∂x 3 2 ∂x2
+ ∂x 3 ∂x3
− 13 ∂x 1
+ ∂x 2
+ ∂x 3

(5.187)

5.4.4 Second law restrictions

Recall that in order that the constitutive equation for viscous stress be consistent with second
law of thermodynamics, that it is sufficient (but perhaps overly restrictive) to require that
Eq. (5.26) hold:
1
τij ∂(i vj) ≥ 0. (5.188)
T | {z }
=Φ

In terms of the viscous dissipation function Φ, this is

1
Φ ≥ 0. (5.189)
T
Invoking our constitutive equation for viscous stress, and realizing that the absolute tem-
perature T > 0, we have then that the viscous dissipation function Φ must satisfy

Φ = (2µ∂(i vj) + λ∂k vk δij )(∂(i vj) ) ≥ 0. (5.190)

© 05 December 2025, Joseph M. Powers. All rights reserved.

188 CHAPTER 5. CONSTITUTIVE EQUATIONS

This reduces to the sum of two squares:

Φ = 2µ∂(i vj) ∂(i vj) + λ∂k vk ∂i vi ≥ 0. (5.191)

We then seek restrictions on µ and λ such that this is true. Obviously requiring µ ≥ 0 and
λ ≥ 0 guarantees satisfaction of the second law. However, Stokes’ assumption of λ = −2µ/3
does not meet this criterion, and so we are motivated to check more carefully to see if we
actually need to be that restrictive.

[Link] One-dimensional systems

Let us first check the criterion for a strictly one-dimensional system. For such a system, our
second law restriction reduces to

2µ∂(1 v1) ∂(1 v1) + λ∂1 v1 ∂1 v1 ≥ 0, (5.192)

(2µ + λ)∂1 v1 ∂1 v1 ≥ 0, (5.193)
2µ + λ ≥ 0, (5.194)
λ ≥ −2µ. (5.195)

Obviously if µ > 0 and λ = −2µ/3, the entropy inequality is satisfied. We also could satisfy
the inequality for negative µ with sufficiently large positive λ.

[Link] Two-dimensional systems

Extending this to a two-dimensional system is more complicated. For such systems, expan-
sion of our second law condition gives

2µ∂(1 v1) ∂(1 v1) + 2µ∂(1 v2) ∂(1 v2) + 2µ∂(2 v1) ∂(2 v1) + 2µ∂(2 v2) ∂(2 v2)



+λ ∂(1 v1) + ∂(2 v2) ∂(1 v1) + ∂(2 v2) ≥ 0. (5.196)

Taking advantage of symmetry of the deformation tensor, we can say

2µ∂(1 v1) ∂(1 v1) + 4µ∂(1 v2) ∂(1 v2) + 2µ∂(2 v2) ∂(2 v2) + λ ∂(1 v1) + ∂(2 v2) ∂(1 v1) + ∂(2 v2) ≥ 0.
(5.197)
Expanding the product and regrouping gives

(2µ + λ)∂(1 v1) ∂(1 v1) + 4µ∂(1 v2) ∂(1 v2) + (2µ + λ)∂(2 v2) ∂(2 v2) + 2λ∂(1 v1) ∂(2 v2) ≥ 0. (5.198)

In matrix form, we can write this inequality in the form known from linear algebra as a
quadratic form (see Powers and Sen (2015), Ch. 7):
  
(2µ + λ) λ 0 ∂(1 v1)
Φ = ( ∂(1 v1) ∂(2 v2) ∂(1 v2) )  λ (2µ + λ) 0   ∂(2 v2)  ≥ 0. (5.199)
0 0 4µ ∂(1 v2)

© 05 December 2025, Joseph M. Powers. All rights reserved.

5.4. STRESS-STRAIN RATE RELATION FOR A NEWTONIAN FLUID 189

As we have discussed before, the condition that this hold for all values of the deformation
is that the symmetric part of the coefficient matrix have eigenvalues that are greater than
or equal to zero. In fact, here the coefficient matrix is purely symmetric. Let us find the
eigenvalues κ of the coefficient matrix. The eigenvalues are found by evaluating the following
equation
(2µ + λ) − κ λ 0
λ (2µ + λ) − κ 0 = 0. (5.200)
0 0 4µ − κ
We get the characteristic polynomial


(4µ − κ) (2µ + λ − κ)2 − λ2 = 0. (5.201)
This has roots
κ = 4µ, (5.202)
κ = 2µ, (5.203)
κ = 2(µ + λ). (5.204)
For the two-dimensional system, we see now formally that we must satisfy both
µ ≥ 0, (5.205)
λ ≥ −µ. (5.206)
This is more restrictive than for the one-dimensional system, but we see that a fluid obeying
Stokes’ assumption λ = −2µ/3 still satisfies this inequality.

[Link] Three-dimensional systems

For a full three-dimensional variation, the entropy inequality (2µ∂(i vj) +λ∂k vk δij )(∂(i vj) ) ≥ 0,
when expanded, is equivalent to the following quadratic form
  
λ + 2µ λ λ 0 0 0 ∂(1 v1)
λ λ + 2µ λ 0 0 0 ∂(2 v2)
   
∂(3 v1) ) 
λ λ λ + 2µ 0 0 0   ∂(3 v3)  ≥ 0.
Φ = ( ∂(1 v1) ∂(2 v2) ∂(3 v3) ∂(1 v2) ∂(2 v3)
 0 0 0 4µ 0 0   ∂(1 v2) 
0 0 0 0 4µ 0 ∂(2 v3)
0 0 0 0 0 4µ ∂(3 v1)
(5.207)
Again this must hold for arbitrary values of the deformation, so we must require that the
eigenvalues κ of the interior matrix be greater than or equal to zero to satisfy the entropy
inequality. It is easy to show that the six eigenvalues for the interior matrix are
κ = 2µ, (5.208)
κ = 2µ, (5.209)
κ = 4µ, (5.210)
κ = 4µ, (5.211)
κ = 4µ, (5.212)
κ = 3λ + 2µ. (5.213)

© 05 December 2025, Joseph M. Powers. All rights reserved.

190 CHAPTER 5. CONSTITUTIVE EQUATIONS

Two of the eigenvalues are degenerate, but this is not a particular problem. We need now
that κ ≥ 0, so the entropy inequality requires that

µ ≥ 0, (5.214)
2
λ ≥ − µ. (5.215)
3
Obviously a fluid that satisfies Stokes’ assumption does not violate the entropy inequality,
but it does give rise to a minimum level of satisfaction. This does not mean the fluid is
isentropic! It simply means one of the six eigenvalues is zero.
Now using standard techniques from linear algebra for quadratic forms (see Powers and
Sen (2015), Ch. 7), the entropy inequality can, after much effort, be manipulated into the
form
2

Φ = µ (∂(1 v1) − ∂(2 v2) )2 + (∂(2 v2) − ∂(3 v3) )2 + (∂(3 v3) − ∂(1 v1) )2
3  
2
+ λ + µ (∂(1 v1) + ∂(2 v2) + ∂(3 v3) )2
3
+4µ((∂(1 v2) )2 + (∂(2 v3) )2 + (∂(3 v1) )2 ) ≥ 0. (5.216)

Obviously, this is a sum of perfect squares, and holds for all values of the strain rate tensor.
It can be verified by direct expansion that this term is identical to the strong form of the
entropy inequality for viscous stress. It can further be verified by direct expansion that the
entropy inequality can also be written more compactly as
    
1 1 2
Φ = 2µ ∂(i vj) − ∂k vk δij ∂(i vj) − ∂k vk δij + λ + µ (∂i vi )(∂j vj ) ≥ 0. (5.217)
3 3 3 | {z }
| {z } (mean strain rate)2
(deviatoric strain rate)2

We could also say

   
1 1
Φ = 2µ D − ϑI : D − ϑI + ζϑ2 ≥ 0. (5.218)
3 3

So, we see that for a Newtonian fluid that the increase in entropy due to viscous dissipation is
attributable to two effects: deviatoric strain rate and mean strain rate. The terms involving
both are perfect squares, so as long as µ ≥ 0 and λ ≥ −2µ/3 (or equivalently µ ≥ 0, ζ ≥ 0),
the second law is not violated by viscous effects.
We can also write the strong form of the entropy inequality for a Newtonian fluid
(2µ∂(i vj) + λ∂k vk δij )(∂(i vj) ) ≥ 0, in terms of the principal invariants of strain rate. Leaving
out details, that can be verified by direct expansion of all terms, we find the following form
      
2 (1) 2 (2) 2 (1)
2
Φ = 2µ Iǫ̇ − 2Iǫ̇ + λ + µ Iǫ̇ ≥ 0. (5.219)
3 3

© 05 December 2025, Joseph M. Powers. All rights reserved.

5.5. EQUATIONS OF STATE 191

Because this is in terms of the invariants, we are assured that it is independent of the
orientation of the coordinate system.
It is, however, not obvious that this form is positive semi-definite. We can use the
definitions of the invariants of strain rate to rewrite the inequality as
   
1 2
Φ = 2µ ∂(i vj) ∂(j vi) − (∂i vi ) (∂j vj ) + λ + µ (∂i vi ) (∂j vj ) ≥ 0. (5.220)
3 3

In terms of the eigenvalues of the strain rate tensor, κ1 , κ2 , and κ3 , this becomes
   
1 2
Φ = 2µ κ1 + κ2 + κ3 − (κ1 + κ2 + κ3 ) + λ + µ (κ1 + κ2 + κ3 )2 ≥ 0.
2 2 2 2
(5.221)
3 3

This then reduces to a positive semi-definite form:

 
2
2
Φ = µ (κ1 − κ2 ) + (κ1 − κ3 ) + (κ2 − κ3 ) + λ + µ (κ1 + κ2 + κ3 )2 ≥ 0.
2 2 2
(5.222)
3 3

Because the eigenvalues are invariant under rotation, this form is invariant.
We summarize by noting relations between mean and deviatoric stress and strain rates
for Newtonian fluids. The influence of each on each has been seen or is easily shown to be
as follows:

• A mean strain rate will induce a time rate of change in the mean thermodynamic stress
via traditional thermodynamic relations12 and will induce an additional mean viscous
stress for fluids that do not obey Stokes’ assumption.

• A deviatoric strain rate will not directly induce a mean stress.

• A deviatoric strain rate will directly induce a deviatoric stress.

• A mean strain rate will induce entropy production only for a fluid that does not obey
Stokes’ assumption.

• A deviatoric strain rate will always induce entropy production in a viscous fluid.

5.5 Equations of state

Thermodynamic equations of state provide algebraic relations between variables such as
pressure, temperature, energy, and entropy. They do not involve velocity. They are formally
valid for materials at rest. As long as the times scales of equilibration of the thermodynamic
variables are much faster than the finest time scales of fluid dynamics, it is a valid assumption
to use an ordinary equations of state. Such assumptions can be violated in high speed flows
12
e.g. for an isothermal ideal gas dp/dt = RT (dρ/dt) = −ρRT ∂i vi .

© 05 December 2025, Joseph M. Powers. All rights reserved.

192 CHAPTER 5. CONSTITUTIVE EQUATIONS

in which vibrational and rotational modes of oscillation become excited. They may also be
invalid in highly rarefied flows such as might occur in the upper atmosphere.
Typically, we will require two types of equations, a thermal equation of state that gives
the pressure as a function of two independent thermodynamic variables, e.g.

p = p(ρ, T ), (5.223)

and a caloric equation of state that gives the internal energy as a function of two independent
thermodynamic variables, e.g.
e = e(ρ, T ). (5.224)
There are additional conditions regarding internal consistency of the equations of state; that
is, just any stray functional forms will not do.
We outline here a method for generating equations of state with internal consistency based
on satisfying the entropy inequality. First let us define a new thermodynamic variable, â,
the Helmholtz13 free energy:
â = e − T s. (5.225)
We can take the material time derivative of Eq. (5.225) to get

dâ de ds dT
= −T −s . (5.226)
dt dt dt dt
It is shown in thermodynamics texts that there are a set of natural, “canonical,” variables
for describing â that are T and ρ. That is, we take â = â(T, ρ). Taking the time derivative
of this form of â and using the chain rule tells us another form for dâ/dt:

dâ ∂â dT ∂â dρ

= + . (5.227)
dt ∂T ρ dt ∂ρ T dt

Now we also have the energy equation and entropy inequality:

de
ρ = −∂i qi − p∂i vi + τij ∂i vj , (5.228)
dt
ds q 
i
ρ ≥ −∂i . (5.229)
dt T
Using Eq. (5.226) to eliminate de/dt in favor of dâ/dt in the energy equation, Eq. (5.228),
gives a modified energy equation:
 
dâ ds dT
ρ +T +s = −∂i qi − p∂i vi + τij ∂i vj . (5.230)
dt dt dt
13
Hermann von Helmholtz, 1821-1894, Potsdam-born German physicist and philosopher, descendant of
William Penn, the founder of Pennsylvania, empiricist and refuter of the notion that scientific conclusions
could be drawn from philosophical ideas, graduated from medical school, wrote convincingly on the science
and physiology of music, developed theories of vortex motion as well as thermodynamics and electrodynamics.

© 05 December 2025, Joseph M. Powers. All rights reserved.

5.5. EQUATIONS OF STATE 193

Next, we use Eq. (5.227) to eliminate dâ/dt in Eq. (5.230) to get

!
∂â dT ∂â dρ ds dT
ρ + +T +s = −∂i qi − p∂i vi + τij ∂i vj . (5.231)
∂T ρ dt ∂ρ T dt dt dt

Now in this modified energy equation, we solve for ρ ds/dt to get

ds 1 p 1 ρ ∂â dT ρ ∂â dρ ρs dT
ρ = − ∂i qi − ∂i vi + τij ∂i vj − − − . (5.232)
dt T T T T ∂T ρ dt T ∂ρ T dt T dt

Substituting this version of the energy conservation equation into the second law, Eq. (5.229),
gives

1 p 1 ρ ∂â dT ρ ∂â dρ ρs dT q 
i
− ∂i qi − ∂i vi + τij ∂i vj − − − ≥ −∂i . (5.233)
T T T T ∂T ρ dt T ∂ρ T dt T dt T

Rearranging and using the mass conservation relation to eliminate ∂i vi , we get

 
qi p 1 dρ 1 ρ ∂â dT ρ ∂â dρ ρs dT
− 2 ∂i T − − + τij ∂i vj − − − ≥ 0, (5.234)
T T ρ dt T T ∂T ρ dt T ∂ρ T dt T dt
qi p dρ ∂â dT ∂â dρ dT
− ∂i T + + τij ∂i vj − ρ −ρ − ρs ≥ 0, (5.235)
T ρ dt ∂T ρ dt ∂ρ T dt dt
  !
qi 1 dρ ∂â dT ∂â
− ∂i T + τij ∂i vj + p − ρ2 −ρ s+ ≥ 0. (5.236)
T ρ dt ∂ρ T dt ∂T ρ
| {z } | {z }
→0 →0

Now in our discussion of the strong form of the energy inequality, we have already found
forms for qi and τij for which the terms involving these phenomena are positive semi-definite.
Now dρ/dt and dT /dt can be positive or negative. So to guarantee the remaining two terms
are consistent with the second law, and are associated with reversible processes, we require
that that
∂â ∂â
p = ρ2 , s=− . (5.237)
∂ρ T ∂T ρ

For example, if we take the non-obvious, but experimentally defensible choice for a of
   
T ρ
â = cv (T − To ) − cv T ln + RT ln , (5.238)
To ρo

then we get for pressure  

∂â 2 2 RT
p=ρ =ρ = ρRT. (5.239)
∂ρ T ρ

© 05 December 2025, Joseph M. Powers. All rights reserved.

194 CHAPTER 5. CONSTITUTIVE EQUATIONS

This equation for pressure is a thermal equation of state for an ideal gas, and R is known as
the gas constant. It is the ratio of the universal gas constant and the molecular mass of the
particular gas.
Solving for entropy s, we get
   
∂â T ρ
s=− = cv ln − R ln . (5.240)
∂T ρ To ρo

Then, we get for e

e = â + T s = cv (T − To ). (5.241)
We call this equation for energy a caloric equation of state for calorically perfect gas. It is
calorically perfect because the specific heat at constant volume cv is assumed a true constant
here. In general for ideal gases, it can be shown to be at most a function of temperature.

© 05 December 2025, Joseph M. Powers. All rights reserved.

Chapter 6

Governing equations: summary and

special cases

see Powers, Chapter 6.

see Panton, Chapters 5 and 8,
see Hughes and Gaylord, Chapter 1,
see Saffman, Chapter 1.

In this chapter, we consider a variety of secondary topics related to the governing equations.
We briefly discuss boundary and interface conditions, necessary for a complete system, sum-
marize the partial differential equations in various forms, present some special cases of the
governing equations, present the equations in a dimensionless form, and consider a few cases
where the linear momenta equation can be integrated once.

6.1 Boundary and interface conditions

At fluid solid interfaces, it is observed in the continuum regime that the fluid sticks to the
solid boundary, so that we can safely take the fluid and solid velocities to be identical at the
interface. This is called the no-slip condition. As one approaches the molecular level, this
breaks down.
At the interface of two distinct, immiscible fluids, one requires that stress and energy flux
both be continuous across the interface. Density need not be continuous in the absence of
mass diffusion. Were mass diffusion present, the fluids would not be immiscible, and density
would be a continuous variable. Additionally the effect of surface tension may need to be
accounted for. We shall not consider surface tension in this course, but many texts give a
complete treatment.

195
196 CHAPTER 6. GOVERNING EQUATIONS: SUMMARY AND SPECIAL CASES

6.2 Complete set of compressible Navier-Stokes equa-

tions
Here we pause once more to write a complete set of equations, the compressible Navier1
-Stokes equations, written here for a fluid that satisfies Stokes’ assumption, but for which
the viscosity µ (as well as thermal conductivity k) may be variable. They are an expanded
version of those presented in Ch. 4.7.

6.2.1 Conservative form

[Link] Cartesian index form

∂o ρ + ∂i (ρvi ) = 0, (6.1)
  
1
∂o (ρvi ) + ∂j (ρvj vi ) = ρfi − ∂i p + ∂j 2µ ∂(j vi) − ∂k vk δji , (6.2)
3
     
1 1
∂o ρ e + vj vj + ∂i ρvi e + vj vj = ρvi fi − ∂i (pvi ) + ∂i (k∂i T )
2 2
   
1
+∂i 2µ ∂(i vj) − ∂k vk δij vj , (6.3)
3
p = p(ρ, T ), (6.4)
e = e(ρ, T ), (6.5)
µ = µ(ρ, T ), (6.6)
k = k(ρ, T ). (6.7)

[Link] Gibbs form

∂ρ
+ ∇T · (ρv) = 0, (6.8)
∂t
   T
∂
T ∇vT + (∇vT )T 1
(ρv) + ∇T · (ρvvT ) = ρf − ∇p + ∇T · 2µ − (∇T · v)I ,
∂t 2 3
| {z }
=µ(∇2 v+∇ϑ/3) if µ constant

(6.9)
     
∂ 1 1
ρ e + vT · v + ∇T · ρv e + vT · v = ρvT · f − ∇T · (pv) + ∇T · (k∇T )
∂t 2 2
   
∇vT + (∇vT )T 1
+ ∇T · 2µ − (∇T · v)I · v , (6.10)
2 3
| {z }
=Φ+µvT ·(∇2 v+∇ϑ/3) if µ constant

p = p(ρ, T ), (6.11)
1
Claude Louis Marie Henri Navier, 1785-1836, Dijon-born French civil engineer and mathematician, stud-
ied under Fourier, taught applied mechanics at École des Ponts et Chaussées, replaced Cauchy as professor
at École Polytechnique, specialist in road and bridge building, did not fully understand shear stress in a fluid
and used faulty logic in arriving at his equations.

© 05 December 2025, Joseph M. Powers. All rights reserved.

6.2. COMPLETE SET OF COMPRESSIBLE NAVIER-STOKES EQUATIONS 197

e = e(ρ, T ), (6.12)
µ = µ(ρ, T ), (6.13)
k = k(ρ, T ). (6.14)

6.2.2 Non-conservative form

[Link] Cartesian index form

dρ
= −ρ∂i vi , (6.15)
dt   
dvi 1
ρ = ρfi − ∂i p + ∂j 2µ ∂(j vi) − ∂k vk δji , (6.16)
dt 3
 
de 1
ρ = −p∂i vi + ∂i (k∂i T ) + 2µ ∂(i vj) − ∂k vk δij ∂i vj , (6.17)
dt 3
p = p(ρ, T ), (6.18)
e = e(ρ, T ), (6.19)
µ = µ(ρ, T ), (6.20)
k = k(ρ, T ). (6.21)

[Link] Gibbs form

dρ
= −ρ∇T · v, (6.22)
dt
   T
dv T ∇vT + (∇vT )T 1
ρ = ρf − ∇p + ∇ · 2µ − (∇T · v)I , (6.23)
dt 2 3
| {z }
µ(∇2 v+∇ϑ/3) if µ constant
 
de T T ∇vT + (∇vT )T 1
ρ = −p∇ · v + ∇ · (k∇T ) + 2µ − (∇T · v)I : ∇vT , (6.24)
dt 2 3
| {z }
Φ
p = p(ρ, T ), (6.25)
e = e(ρ, T ), (6.26)
µ = µ(ρ, T ), (6.27)
k = k(ρ, T ). (6.28)

We take µ, and k to be thermodynamic properties of temperature and density. In practice,

both dependencies are often weak, especially the dependency of µ and k on density. We also
assume we know the form of the external body force per unit mass fi . We also no longer
formally require the angular momenta principle, as it has been absorbed into our constitutive
equation for viscous stress. We also need not write the second law, as we can guarantee its
satisfaction as long as µ ≥ 0, k ≥ 0.

© 05 December 2025, Joseph M. Powers. All rights reserved.

198 CHAPTER 6. GOVERNING EQUATIONS: SUMMARY AND SPECIAL CASES

In summary, we have nine unknowns, ρ, vi (3), p, e, T , µ, and k, and nine equations,

mass, linear momenta (3), energy, thermal state, caloric state, and thermodynamic relations
for viscosity and thermal conductivity. When coupled with initial, interface, and boundary
conditions, all dependent variables can, in principle, be expressed as functions of position xi
and time t, and this knowledge utilized to design devices of practical importance.

6.3 Fluid statics

Fluids that are static have v = 0. In this limit, the equations simplify considerably.

6.3.1 Cartesian index form

∂o ρ = 0, (6.29)
0 = ρfi − ∂i p, (6.30)
ρ∂o h = ∂o p + ∂i (k∂i T ), (6.31)
p = p(ρ, T ), (6.32)
h = h(ρ, T ), (6.33)
k = k(ρ, T ). (6.34)

6.3.2 Gibbs form

∂ρ
= 0, (6.35)
∂t
0 = ρf − ∇p, (6.36)
∂h ∂p
ρ = + ∇T · (k∇T ), (6.37)
∂t ∂t
p = p(ρ, T ), (6.38)
h = h(ρ, T ), (6.39)
k = k(ρ, T ). (6.40)

Example 6.1
Consider the limit in which the density ρ is a constant, and the body force is f = −gk, thermal
conductivity k is constant, and there is no time-dependency. For z = 0, we have p = po , T = To . For
z = h, we have T = T1 . Find the pressure and temperature fields.

The mass equation is trivially satisfied. The momentum equation becomes

∂p ∂p ∂p
0 = 0i + 0j − ρgk − i− j− k. (6.41)
∂x ∂y ∂z

© 05 December 2025, Joseph M. Powers. All rights reserved.

6.4. INCOMPRESSIBLE NAVIER-STOKES EQUATIONS WITH CONSTANT
PROPERTIES 199
This gives us

∂p ∂p ∂p
= 0, = 0, = −ρg. (6.42)
∂x ∂y ∂z

The solution for pressure that satisfies is

p = −ρgz + po . (6.43)

The energy equation becomes

∇2 T = 0. (6.44)

If we seek solutions of the form T = T (z), this reduces to

d2 T
= 0. (6.45)
dz 2
The variation in T must be linear in z. The obvious solution that satisfies the boundary conditions is
z
T = To + (T1 − To ) . (6.46)
h

6.4 Incompressible Navier-Stokes equations with con-

stant properties
If we make the assumption, that can be justified in the limit when fluid particle velocities
are small relative to the velocity of sound waves in the fluid, that density changes following a
particle are negligible (that is, dρ/dt → 0), the Navier-Stokes equations simplify considerably.
This does not imply the density is constant everywhere in the flow. Our assumption allows
for stratified flows, for which the density of individual particles still can remain constant.
We shall also assume viscosity µ, and thermal conductivity k are constants, though this is
not necessary.
Let us examine the mass, linear momenta, and energy equations in this limit.

6.4.1 Mass
Expanding the mass equation
∂o ρ + ∂i (ρvi ) = 0, (6.47)
we get
∂o ρ + vi ∂i ρ +ρ∂i vi = 0. (6.48)
| {z }
dρ
dt
→0

© 05 December 2025, Joseph M. Powers. All rights reserved.

200 CHAPTER 6. GOVERNING EQUATIONS: SUMMARY AND SPECIAL CASES

We are assuming the first two terms in this expression, that form dρ/dt, go to zero; hence
the mass equation becomes ρ∂i vi = 0. Because ρ > 0, we can say

∂i vi = 0, (6.49)
T
∇ · v = 0. (6.50)

So, for an incompressible fluid, the relative expansion rate for a fluid particle is zero, by
Eq. (3.187).

6.4.2 Linear momenta

Let us first consider the viscous term:
  
1
∂j 2µ ∂(j vi) − ∂k vk δij  , (6.51)
3 |{z}
=0



∂j 2µ ∂(j vi) , (6.52)
∂j (µ (∂i vj + ∂j vi )) . (6.53)

Because µ is constant here, we get

µ (∂j ∂i vj + ∂j ∂j vi ) , (6.54)
 

µ ∂i ∂j vj +∂j ∂j vi  , (6.55)

|{z}
=0
µ∂j ∂j vi . (6.56)

Everything else in the linear momenta equation is unchanged; hence, we get

ρ∂o vi + ρvj ∂j vi = ρfi − ∂i p + µ∂j ∂j vi , (6.57)

dv
ρ = ρf − ∇p + µ∇2 v. (6.58)
dt
In the incompressible constant viscosity limit, the mass and linear momenta equations form
a complete set of four equations in four unknowns: p,vi . We will see that in this limit the
energy equation is coupled to mass, and linear momenta, but it is only a one-way coupling.

6.4.3 Energy
The compressible energy equation in full generality is given by Eq. (4.161):

dh dp
ρ = − ∂i qi + τij ∂i vj . (6.59)
dt dt

© 05 December 2025, Joseph M. Powers. All rights reserved.

6.4. INCOMPRESSIBLE NAVIER-STOKES EQUATIONS WITH CONSTANT
PROPERTIES 201
It is shown in Powers (2024) that if h = h(T, p) that
1
dh = cp dT + (1 − βT ) dp, (6.60)
ρ
where the specific heat at constant pressure cp is defined as
∂h
cp = , (6.61)
∂T p

and the thermal expansion coefficient β is defined as

1 ∂ρ
β=− . (6.62)
ρ ∂T p

For common liquids or solids, β ∼ 0. For an ideal gas, it is easy to show β = 1/T , and this
yields dh = cp dT for the ideal gas. We operate on Eq. (6.60) to get
dh dT dp
ρ = ρcp + (1 − βT ) . (6.63)
dt dt dt
We use Eq. (6.63) to rewrite the compressible energy equation, Eq. (6.59) as
dT dp
ρcp = βT − ∂i qi + τij ∂i vj . (6.64)
dt dt
Imposing our constitutive equations and assumption of incompressibility, we get
 
dT dp 1
ρcp = βT − ∂i (−k∂i T ) + 2µ ∂(i vj) − ∂k vk δij  ∂i vj , (6.65)
dt dt 3 |{z}
=0
dp
= βT + k∂i ∂i T + 2µ∂(i vj) ∂i vj , (6.66)
dt  
dp  
= βT + k∂i ∂i T + 2µ ∂(i vj) ∂(i vj) + ∂[i vj]  , (6.67)
dt | {z } | {z } | {z }
sym. sym. anti−sym.
dp
= βT + k∂i ∂i T + 2µ∂(i vj) ∂(i vj) , (6.68)
dt | {z }
Φ
   
dT dp 2 ∇vT + (∇vT )T ∇vT + (∇vT )T
ρcp = βT + k∇ T + 2µ : . (6.69)
dt dt 2 2
| {z }
Φ

For incompressible flows, we have ϑ = 0, and the viscous dissipation function Φ from
Eq. (5.218) reduces to

Φ = 2µ∂(i vj) ∂(i vj) = 2µD : D = 2µ tr (D · D). (6.70)

© 05 December 2025, Joseph M. Powers. All rights reserved.

202 CHAPTER 6. GOVERNING EQUATIONS: SUMMARY AND SPECIAL CASES

It is a scalar function and obviously positive for µ > 0 because it is a tensor inner product of
a tensor with itself. Because ϑ = 0, all of the strain rate tensor D is completely constituted
of deviatoric strain rate. The mean strain rate is zero. So all of the viscous dissipation is
due to isochoric fluid deformation.

6.4.4 Poisson equation for pressure

In the numerical solution of incompressible Navier-Stokes equations in multiple dimensions,
one is often required to solve a Poisson2 equation for the pressure field. Let us see how such
an equation arises. We will not solve this equation as there are complicated issues associated
with the boundary conditions.
For simplicity, make two assumptions. First, consider the body force to be zero, fi = 0.
Second let us consider constant density flow, thus requiring that ρ be the same constant on
all streamlines and not allowing stratified flow. Then take the divergence of the constant
density, constant property linear momenta equation under these limits, Eq. (6.57):

∂i (ρ∂o vi + ρvj ∂j vi ) = ∂i (−∂i p + µ∂j ∂j vi ) , (6.71)

ρ∂o ∂i vi +ρ∂i (vj ∂j vi ) = −∂i ∂i p + µ∂j ∂j ∂i vi , (6.72)
|{z} |{z}
=0 =0

Because mass is conserved, we have ∂i vi = 0, by Eq. (6.49), so

ρ∂i (vj ∂j vi ) = −∂i ∂i p, (6.73)

ρ(vj ∂j ∂i vi +(∂i vj )(∂j vi )) = −∂i ∂i p, (6.74)
|{z}
=0
−∂i ∂i p = ρ(∂i vj )(∂j vi ), −∇2 p = ρ∇vT : ∇vT . (6.75)

After splitting ∇vT into its symmetric and anti-symmetric parts, it can be verified by direct
expansion that this is equivalent to the Poisson equation
 
2 ωT · ω
−∇ p = ρ D : D − . (6.76)
2

Because of the positive definite nature of the Laplacian operator −∇2 , we can deduce that
deformation induces positive pressure on the interior of a domain, and that vorticity has the
opposite effect. Expanding, we find
 2   2  2  2  !
∂ p ∂2p ∂2p ∂u ∂v ∂w ∂u ∂v ∂u ∂w ∂v ∂w
− + + =ρ + + +2 + + .
∂x2 ∂y 2 ∂z 2 ∂x ∂y ∂z ∂y ∂x ∂z ∂x ∂z ∂y
(6.77)
2
Siméon Denis Poisson, 1781-1840, French mathematician taught by Laplace, Lagrange, and Legendre,
studied partial differential equations, potential theory, elasticity, and electrodynamics.

© 05 December 2025, Joseph M. Powers. All rights reserved.

6.4. INCOMPRESSIBLE NAVIER-STOKES EQUATIONS WITH CONSTANT
PROPERTIES 203
For two-dimensional Cartesian flows, we have w = 0 and no variation with z. Using in-
compressibility, ∂u/∂x + ∂v/∂y = 0, the Poisson equation for pressure reduces in the two-
dimensional limit to the much simpler
 2   
∂ p ∂2p ∂u ∂v ∂u ∂v
− + = 2ρ − . (6.78)
∂x2 ∂y 2 ∂y ∂x ∂x ∂y
It is interesting to note that the Laplacian of the pressure field is balanced by terms that
arise from fluid inertia but not from viscous stresses. Of course the more primitive pressure
and velocity fields do depend on viscosity. If we assume a nonzero conservative body force
that be written as the gradient of a potential, f = −∇ϕ, then in the constant density limit,
Eq. (6.76) becomes
 
2 ωT · ω
−∇ (p + ρϕ) = ρ D : D − . (6.79)
2
For a nonconservative body force, we simply have
 
2 ωT · ω
−∇ p = ρ D : D − − ρ∇T · f. (6.80)
2

6.4.5 Vector potential

Because ∇T · v = 0, it is possible to define a vector potential, ψ, that must be constructed
such that the following holds:

v = ∇ × ψ. (6.81)

Here ψ is the three-dimensional extension of the well-known scalar stream function ψ(x, y)
that will be introduced for two-dimensional incompressible flows in Sec. 8.1. Now by the
properties of vector calculus, we are ensured that the incompressibility condition will hold
as

∇T · ∇ × ψ = 0, (6.82)
| {z }
v

for all ψ. Next recall the definition of vorticity ω, Eq. (3.112), and operate on it:

ω = ∇ × v, (6.83)
= ∇ × ∇ × ψ, (6.84)
= ∇(∇T · ψ) − ∇2 ψ. (6.85)

We note the vector potential ψ is not uniquely determined, for example, if the curl of ψ
maps to the physical velocity vector v, so does ψ + ∇χ, where χ is any scalar function. Let

© 05 December 2025, Joseph M. Powers. All rights reserved.

204 CHAPTER 6. GOVERNING EQUATIONS: SUMMARY AND SPECIAL CASES

us choose to constrain ψ such that ∇T · ψ = 0. This is a common exercise in magneto-electro

dynamics, and is known as choosing the gauge, see Feynman, et al. (1963, Volume 2, Chapter
18). With this choice, we get a Poisson equation for ψ:

−∇2 ψ = ω. (6.86)

In the limit of zero inertia, the incompressible Navier-Stokes equations can be shown to yield
∇2 ω = 0. Thus taking the Laplacian of our Poisson equation for ψ, we get the biharmonic
equation
∇4 ψ = 0.
This is valid under the constraint ∇T · ψ = 0 and in the limit of zero inertia.
Taking ψ = (ψx , ψy , ψz )T , we also see that
∂ψz ∂ψy ∂ψz ∂ψx ∂ψx ∂ψy
u= − , v=− + , w=− + . (6.87)
∂y ∂z ∂x ∂z ∂y ∂x

In the two-dimensional limit with v = (u(x, y), v(x, y), 0)T and ω = (0, 0, ωz (x, y))T , we
recover ψ = (0, 0, ψz )T with
∂ψz ∂ψz
u= , v=− . (6.88)
∂y ∂x
This is equivalent to the stream function we will define in Eq. (8.9) when we consider two-
dimensional potential flows in Ch. 8.

6.4.6 Summary of incompressible constant property equations

The incompressible constant property equations for an isotropic liquid are summarized below
in Gibbs notation:

∇T · v = 0, (6.89)
dv
ρ = ρf − ∇p + µ∇2 v, (6.90)
dt
dT dp
ρcp = βT + k∇2 T + Φ. (6.91)
dt dt

6.4.7 Limits for one-dimensional diffusion

For a static fluid (vi = 0), we have d/dt = ∂/∂t and Φ = 0; hence the energy equation can
be written in a familiar form
∂T
= α∇2 T. (6.92)
∂t
Here we take α
k
α= , (6.93)
ρcp

© 05 December 2025, Joseph M. Powers. All rights reserved.

6.5. EULER EQUATIONS 205

to be defined as the thermal diffusivity. In SI, thermal diffusivity has units of m2 /s. For
one-dimensional cases where all variation is in the x2 direction, we get

∂T ∂2T
=α 2. (6.94)
∂t ∂x2

Compare this to the momentum equation for a specific form of the velocity field, namely,
vi (xi ) = v1 (x2 , t). When we also have no pressure gradient and no body force, the linear
momenta principle reduces to
∂v1 ∂ 2 v1
=ν 2. (6.95)
∂t ∂x2
Here we take ν
µ
ν= , (6.96)
ρ
to be defined as the momentum diffusivity, more commonly known as the kinematic viscosity.
In SI, momentum diffusivity has units of m2 /s; these are the same as for thermal diffusivity.
This equation has an identical form to that for one-dimensional energy diffusion. In fact the
physical mechanism governing both, random molecular collisions, is the same.

6.5 Euler equations

The Euler equations are best described as the special case of the compressible Navier-Stokes
equations in the limit in which diffusion is negligibly small; thus, we consider τij → 0, qi → 0.
For an isotropic fluid that obeys Fourier’s law, we could also insist that µ → 0, k → 0. As
Euler equations are typically used for compressible flows for which body forces are often
negligible, we also take fi → 0, though this can be relaxed. A version of these equations
was first presented by Euler,3 although he only considered the mass and linear momenta
principles. With the later nineteenth century development of thermodynamics, the Euler
equations have been taken to include the first law as well. For a complete set, they must be
supplemented by appropriate thermodynamic equations of state; we leave these in a general
form here.
Because the Euler equations neglect entropy-generating diffusive mechanisms, for contin-
uous regions of flow, the second law tells us that the entropy is constant. However, we shall
see later in Ch. 9.4 that shock discontinuities induce entropy changes. Also, because viscous
stress is negligible, the angular momenta principle is irrelevant for the Euler equations. We
summarize some of the various forms of the Euler equations next. They represent seven
equations for the seven unknowns ρ, vi , p, e, T .
3
Euler, L., 1757, “Principes généraux du mouvement des fluides,” Mémoires de l’Académie des Sciences
de Berlin, 11: 274-315.

© 05 December 2025, Joseph M. Powers. All rights reserved.

206 CHAPTER 6. GOVERNING EQUATIONS: SUMMARY AND SPECIAL CASES

6.5.1 Conservative form

[Link] Cartesian index form

∂o ρ + ∂i (ρvi ) = 0, (6.97)
∂o (ρvi ) + ∂j (ρvj vi + pδji ) = 0, (6.98)
     
1 1 p
∂o ρ e + vj vj + ∂i ρvi e + vj vj + = 0, (6.99)
2 2 ρ
p = p(ρ, T ), (6.100)
e = e(ρ, T ). (6.101)

[Link] Gibbs form

∂ρ
+ ∇T · (ρv) = 0, (6.102)
∂t
∂
T
(ρv) + ∇T · (ρvvT ) + pI = 0, (6.103)
   ∂t   
∂ 1 T T 1 T p
ρ e+ v ·v + ∇ · ρv e + v · v + = 0, (6.104)
∂t 2 2 ρ
p = p(ρ, T ), (6.105)
e = e(ρ, T ). (6.106)

6.5.2 Non-conservative form

[Link] Cartesian index form

dρ
= −ρ∂i vi , (6.107)
dt
dvi
ρ = −∂i p, (6.108)
dt
de
ρ = −p∂i vi , (6.109)
dt
p = p(ρ, T ), (6.110)
e = e(ρ, T ). (6.111)

[Link] Gibbs form

dρ
= −ρ∇T · v, (6.112)
dt
dv
ρ = −∇p, (6.113)
dt

© 05 December 2025, Joseph M. Powers. All rights reserved.

6.5. EULER EQUATIONS 207

de
ρ = −p∇T · v, (6.114)
dt
p = p(ρ, T ), (6.115)
e = e(ρ, T ). (6.116)

6.5.3 Alternate forms of the energy equation

The neglect of entropy-generating mechanisms allows the energy equation in the Euler equa-
tions to be cast in some simple forms that clearly illuminate the fluid’s behavior. One can
specialize Eq. (4.155) to cast the energy equation as

de p dρ
ρ
= , (6.117)
dt ρ dt
de p dρ
− 2 = 0. (6.118)
dt ρ dt

In terms of differentials this is simply

p
de − dρ = 0, (6.119)
ρ2

that, when compared to the Gibbs equation, Eq. (4.164), tells us that this flow is isentropic,
ds = 0, on a particle pathline. Moreover, using the definition of specific volume, v̂ = 1/ρ,
the non-conservative form of the energy equation in the Euler equations is simply

de = −p dv̂. (6.120)

That is to say the change is energy is attributable solely to the reversible work done by a
pressure force acting to change the volume. In terms of the material derivative, one would
say

de dv̂
= −p . (6.121)
dt dt
Equivalently, one could state the energy equation in terms of entropy by considering Eq. (4.169)
in the limit of qi = 0, τij = 0:

ds
= 0, (6.122)
dt
∂o s + vi ∂i s = 0, (6.123)
∂s
+ vT · ∇s = 0. (6.124)
∂t
Integrating on a particle pathline, we get s = C, where the constant C may vary from
pathline to pathline. We adopt the following nomenclature:

© 05 December 2025, Joseph M. Powers. All rights reserved.

208 CHAPTER 6. GOVERNING EQUATIONS: SUMMARY AND SPECIAL CASES

• isentropic flow: the entropy s remains constant on a pathline, but may vary from
pathline to pathiline.
• homeoentropic flow: the entropy s is the same constant throughout all of the flow field.
For the special case in which the fluid is a CPIG, we have p = ρRT , e = cv T + ê, and
the first law, Eq. (6.119) reduces to
p
cv dT − 2 dρ = 0, (6.125)
ρ
 
p p
cv d − 2 dρ = 0, (6.126)
ρR ρ
 
cv 1 p p
dp − 2 dρ − 2 dρ = 0, (6.127)
R ρ ρ ρ
 
cv p p
dp − dρ − dρ = 0, (6.128)
cp − cv ρ ρ
 
1 p p
dp − dρ − dρ = 0, (6.129)
γ−1 ρ ρ
dp dρ
= γ , (6.130)
p ρ
p ρ
ln = γ ln , (6.131)
po ρo
 γ
p ρ
= , (6.132)
po ρo
p
= C. (6.133)
ργ
This is the well-known relation for the isentropic behavior of a CPIG, where C is a constant.
We really confined ourselves to a particle pathline as we were considering the material time
derivative. So the “constant” C actually can take on different values on different pathlines.
Another way to cast this version of the energy equation (for an inviscid CPIG) is
 
d p
= 0, (6.134)
dt ργ
   
p p
∂o γ
+ vi ∂i = 0, (6.135)
ρ ργ
   
∂ p T p
γ
+v ·∇ = 0. (6.136)
∂t ρ ργ

6.6 Dimensionless compressible Navier-Stokes equations

Here we discuss how to scale the Navier-Stokes equations into a set of dimensionless equa-
tions. Panton (2024) gives a general background for scaling. White and Majdalani (2022)
give a detailed discussion of the dimensionless form of the Navier-Stokes equations.

© 05 December 2025, Joseph M. Powers. All rights reserved.

6.6. DIMENSIONLESS COMPRESSIBLE NAVIER-STOKES EQUATIONS 209

Consider the Navier-Stokes equations for a CPIG that has Newtonian behavior, satisfies
Stokes’ assumption, and has constant viscosity, thermal conductivity, and specific heat at
constant volume:

∂o ρ + ∂i (ρvi ) = 0, (6.137)
  
1
∂o (ρvi ) + ∂j (ρvj vi ) = ρfi − ∂i p + µ∂j 2 ∂(j vi) − ∂k vk δji ,(6.138)
3
     
1 1
∂o ρ e + vj vj + ∂i ρvi e + vj vj = ρvi fi − ∂i (pvi ) + k∂i ∂i T
2 2
   
1
+µ∂i 2 ∂(i vj) − ∂k vk δij vj , (6.139)
3
p = ρRT, (6.140)
e = cv T + ê. (6.141)

Here R is the gas constant for the particular gas we are considering, that is the ratio of the
universal gas constant ℜ and the gas’s molecular mass M: R = ℜ/M. Also ê is a constant.
Now solutions to these equations, that may be of the form, for example, of p(x1 , x2 , x3 , t),
are necessarily parameterized by the constants from constitutive laws such as cv , R, µ, k, fi ,
in addition to parameters from initial and boundary conditions. That is our solutions will
really be of the form
p(x1 , x2 , x3 , t; cv , R, µ, k, fi , . . .). (6.142)
It is desirable for many reasons to reduce the number of parametric dependencies of these
solutions. Some of these reasons include

• identification of groups of terms that truly govern the features of the flow,

• efficiency of presentation of results, and

• efficiency of design of experiments.

The Navier-Stokes equations (and nearly all sets of physically motivated equations) can be
reduced in complexity by considering scaled versions of the same equations. For a given
problem, the proper scales are non-unique, though some choices will be more helpful than
others. One generally uses the following rules of thumb in choosing scales:

• reduce variables so that their scaled value is near unity,

• demonstrate that certain physical mechanisms may be negligible relative to other phys-
ical mechanisms, and

• simplify initial and boundary conditions.

In forming dimensionless equations, one must usually look for

• characteristic length scale L, and

© 05 December 2025, Joseph M. Powers. All rights reserved.

210 CHAPTER 6. GOVERNING EQUATIONS: SUMMARY AND SPECIAL CASES

Figure 6.1: Sketch of known flow from infinity approaching body with characteristic length
L.

• characteristic time scale tc .

Often an ambient velocity or sound speed exists that can be used to form either a length
or time scale, for example
L
• given vo , L −→ tc = vo
,

• given vo , tc −→ L = vo tc .

If for example our physical problem involves the flow over a body of length L (and
whose other dimensions are of the same order as L), and free-stream conditions are known
to be p = po , vi = (vo , 0, 0)T , ρ = ρo , as sketched in Fig. 6.1, Knowledge of free-stream
pressure and density fixes all other free-stream thermodynamic variables, e.g. e, T , via the
thermodynamic relations. For this problem, let the ∗ subscript represent a dimensionless
variable. Define the following scaled dependent variables:
ρ p vi ρo R ρo
ρ∗ = , p∗ = , v∗i = , T∗ = T, e∗ = e. (6.143)
ρo po vo po po
Define the following scaled independent variables:
xi vo
x∗i = , t∗ = t. (6.144)
L L
With these definitions, the operators must also be scaled, that is,
∂ dt∗ ∂ vo ∂ vo
∂o = = = = ∂∗o ,
∂t dt ∂t∗ L ∂t∗ L
L
∂∗o = ∂o .
vo
∂ dx∗i ∂ 1 ∂ 1
∂i = = = = ∂∗i ,
∂xi dxi ∂x∗i L ∂x∗i L
∂∗i = L∂i . (6.145)

© 05 December 2025, Joseph M. Powers. All rights reserved.

6.6. DIMENSIONLESS COMPRESSIBLE NAVIER-STOKES EQUATIONS 211

6.6.1 Mass
Let us make these substitutions into the mass equation:
∂o ρ + ∂i (ρvi ) = 0, (6.146)
vo 1
∂∗o (ρo ρ∗ ) + ∂∗i (ρo ρ∗ vo v∗i ) = 0, (6.147)
L L
ρo vo
(∂∗o ρ∗ + ∂∗i (ρ∗ v∗i )) = 0, (6.148)
L
∂∗o ρ∗ + ∂∗i (ρ∗ v∗i ) = 0. (6.149)
The mass equation is unchanged in form when we transform to a dimensionless version.

6.6.2 Linear momenta

We have a similar analysis for the linear momenta equation.
∂o (ρvi ) + ∂j (ρvj vi ) = ρfi − ∂i p
  
1
+µ∂j 2 ∂(j vi) − ∂k vk δji , (6.150)
3
vo
∂∗o (ρo vo ρ∗ v∗i )
L
1 1
+ ∂∗j (ρo ρ∗ vo v∗j vo v∗i ) = ρo ρ∗ fi − ∂∗i (po p∗ )
L L  
µ 2 1
+ ∂∗j ∂(∗j vo v∗i) − ∂∗k vo v∗k δji , (6.151)
L L 3
ρo vo2
∂∗o (ρ∗ v∗i )
L
ρo v 2 po
+ o ∂∗j (ρ∗ v∗j v∗i ) = ρo ρ∗ fi − ∂∗i (p∗ )
L L 
 
µvo 1
+ 2 ∂∗j 2 ∂(∗j v∗i) − ∂∗k v∗k δji , (6.152)
L 3
fi L po
∂∗o (ρ∗ v∗i ) + ∂∗j (ρ∗ v∗j v∗i ) = ρ ∗ − ∂∗i (p∗ )
vo2 ρo vo2
 
2µ 1
+ ∂∗j ∂(∗j v∗i) − ∂∗k v∗k δji . (6.153)
ρo vo L 3
With this scaling, we have generated three distinct dimensionless groups of terms that drive
the linear momenta equation:
fi L po µ
, , and . (6.154)
vo2 ρo vo2 ρo vo L
These groups are closely related to the following groups of terms, that have the associated
interpretations indicated:

© 05 December 2025, Joseph M. Powers. All rights reserved.

212 CHAPTER 6. GOVERNING EQUATIONS: SUMMARY AND SPECIAL CASES

• Froude number F r:4 With the body force per unit mass fi = gĝi , where g > 0 is the
gravitational acceleration magnitude and ĝi is a unit vector pointing in the direction
of gravitational acceleration,
vo2 flow kinetic energy
F r2 ≡ = . (6.155)
gL gravitational potential energy
• Mach number Mo :5 With the Mach number Mo defined as the ratio of the ambient
velocity to the ambient sound speed, and recalling that for a CPIG that the square
of the ambient sound speed, c2o is c2o = γpo /ρo , where γ is the ratio of specific heats
γ = cp /cv = (1 + R/cv ), we have
vo2 vo2 ρo vo2 vo2 flow kinetic energy
Mo2 ≡ 2
= po = = = . (6.156)
co γ ρo γpo γRTo thermal energy
Here we have taken To = po /ρo /R.
• Reynolds number Re: We have
ρo vo L ρo v 2 dynamic pressure
Re ≡ = voo = . (6.157)
µ µL viscous stress
With these definitions, we get
1 1 1
∂∗o (ρ∗ v∗i ) + ∂∗j (ρ∗ v∗j v∗i ) = 2
ĝi ρ∗ − ∂∗i (p∗ )
Fr γ Mo2
 
2 1
+ ∂∗j ∂(∗j v∗i) − ∂∗k v∗k δji . (6.158)
Re 3
The relative magnitudes of F r, Mo , and Re play a crucial role in determining that physical
mechanisms are most influential in changing the fluid’s linear momenta.
Another important number is the Knusden number, Kn. For a calorically perfect ideal
gas it is
r
Mo γπ
Kn = , (6.159)
Re 2
where γ is the constant ratio of specific heats. For low Kn < 0.01, the flow is well modeled
as a continuum. For high Kn > 10, molecular non-continuum effects are important. With
effort, one can also show that
λ
Kn = , (6.160)
L
where λ is the mean free path between molecular collisions, introduced in Eq. (1.9), and L
is a characteristic geometric length scale.
4
William Froude, 1810-1879, English engineer and naval architect, Oxford educated.
5
Ernst Mach, 1838-1926, Viennese physicist and philosopher who worked in optics, mechanics, and wave
dynamics, received doctorate at University of Vienna and taught mathematics at University of Graz and
physics at Charles University of Prague, developed fundamental ideas of inertia that influenced Einstein.

© 05 December 2025, Joseph M. Powers. All rights reserved.

6.6. DIMENSIONLESS COMPRESSIBLE NAVIER-STOKES EQUATIONS 213

6.6.3 Energy

The analysis is of the exact same form, but more tedious, for the energy equation.

     
1 1
∂o ρ e + vj vj + ∂i ρvi e + vj vj = k∂i ∂i T − ∂i (pvi )
2 2
   
1
+µ∂i 2 ∂(i vj) − ∂k vk δij vj
3
+ρvi fi , (6.161)
  
vo po 1
∂∗o ρo ρ∗ e∗ + vo2 v∗j v∗j
L ρ 2
  o 
1 po 1 2 k po
+ ∂i∗ ρo ρ∗ vo vi∗ e∗ + vo v∗j v∗j = 2
∂∗i ∂∗i T∗
L ρo 2 L ρo R
1
− ∂∗i (po p∗ vo v∗i )
L    
µ 2 1
+ ∂∗i ∂(∗i vo v∗j) − ∂∗k vo v∗k δij vo v∗j
L L 3
+ρo ρ∗ vo v∗i fi , (6.162)
!!
ρo vo po 1 γvo2
∂∗o ρ∗ e∗ + v∗j v∗j
L ρo 2 γ ρpoo
!!
ρo vo po 1 γvo2 k po 1
+ ∂∗i ρ∗ v∗i e∗ + po v∗j v∗j = ∂∗i ∂∗i T∗
L ρo 2 γ ρo L2 ρo R
po vo
− ∂∗i (p∗ v∗i )
L   
2µvo2 1
+ 2 ∂∗i ∂(∗i v∗j) − ∂∗k v∗k δij v∗j
L 3
+ρo vo fi ρ∗ v∗i , (6.163)
!!
1 γvo2
∂∗o ρ∗ e∗ + v∗j v∗j
2 γ ρpoo
!!
1 γvo2 k
+∂∗i ρ∗ v∗i e∗ + po v∗j v∗j = ∂∗i ∂∗i T∗
2 γ ρo LRρo vo
−∂∗i (p∗ v∗i )
  
2µvo2 L 1 1
+ 2 ∂∗i ∂(∗i v∗j) − ∂∗k v∗k δij v∗j
L ρo vo pρoo 3
fi L
+ po ρ∗ v∗i . (6.164)
ρo

© 05 December 2025, Joseph M. Powers. All rights reserved.

214 CHAPTER 6. GOVERNING EQUATIONS: SUMMARY AND SPECIAL CASES

Now examining the dimensionless groups, we see that

k k cp 1 k cp µ 1 γ 1
= = = . (6.165)
LRρo vo cp R Lρo vo µcp cp − cv ρo vo L P r γ − 1 Re

Here we have a new dimensionless group, the Prandtl6 number, P r, where

µ
µcp ρo momentum diffusivity ν
Pr ≡ = k
= = . (6.166)
k ρo cp
energy diffusivity α

This has employed definitions of diffusivities given earlier in Eqs. (6.93, 6.96). We also see
that

fi L γgLĝi vo2 gL Mo2

po = = γ ĝ i = γ ĝi , (6.167)
ρo
γ ρpoo γ ρpoo vo2 F r2
γvo2
= γMo2 , (6.168)
γ ρpoo
2µvo2 L 1 2µ γvo2 1
2 po = po = 2 γMo2 . (6.169)
L ρo vo ρo ρo vo L γ ρo Re

So, the dimensionless energy equation becomes

  
1 2
∂∗o ρ∗ e∗ + γMo v∗j v∗j
2
  
1 2 γ 1 1
+∂∗i ρ∗ v∗i e∗ + γMo v∗j v∗j = ∂∗i ∂∗i T∗
2 γ − 1 P r Re
−∂∗i (p∗ v∗i )
  
Mo2 1
+2γ ∂∗i ∂(∗i v∗j) − ∂∗k v∗k δij v∗j
Re 3
2
γMo
+ ĝi ρ∗ v∗i . (6.170)
F r2

6.6.4 Thermal state equation

 
po
po p∗ = ρo ρ∗ R T∗ , (6.171)
ρo R
p∗ = ρ∗ T∗ . (6.172)
6
Ludwig Prandtl, 1875-1953, German mechanician and father of aerodynamics, primarily worked at Uni-
versity of Göttingen, discoverer of the boundary layer, pioneer of dirigibles, and advocate of monoplanes.

© 05 December 2025, Joseph M. Powers. All rights reserved.

6.7. FIRST INTEGRALS OF LINEAR MOMENTA 215

6.6.5 Caloric state equation

 
po po
e∗ = cv T∗ + ê, (6.173)
ρo ρo R
cv ρo ê
e∗ = T∗ + , (6.174)
R po
1 ρo ê
e∗ = T∗ + . (6.175)
γ−1 po
|{z}
unimportant

For completeness, we retain the term ρo ê/po . It actually plays no role in this non-reactive
flow because energy only enters via its derivatives. When flows with chemical reactions are
modeled, this term may be important.

6.6.6 Upstream conditions

Scaling the upstream conditions, we get
p∗ = 1, ρ∗ = 1, v∗i = (1, 0, 0)T . (6.176)
With this we then get secondary relationships
1 ρo ê
T∗ = 1, e∗ = + . (6.177)
γ−1 po

6.6.7 Reduction in parameters

We lastly note that our original system had the following ten independent parameters:
ρo , po , cv , R, L, vo , µ, k, fi , ê. (6.178)
Our scaled system however has only six independent parameters:
ρo ê
Re, P r, Mo , F r, γ, . (6.179)
po
We have lost no information, nor made any approximations, and we have a system with
fewer dependencies.

6.7 First integrals of linear momenta

Under special circumstances, we can integrate the linear momenta principle to obtain a
simplified equation. We will consider two cases here, what is known as Bernoulli’s7 equation
7
Daniel Bernoulli, 1700-1782, Dutch-born Swiss mathematician of the prolific and mathematical Bernoulli
family, son of Johann Bernoulli, studied at Heidelberg, Strasbourg, and Basel, receiving M.D. degree, served
in St. Petersburg and lectured at the University of Basel, put forth his fluid mechanical principle in the 1738
Hydrodynamica, in competition with his father’s 1738 Hydraulica.

© 05 December 2025, Joseph M. Powers. All rights reserved.

216 CHAPTER 6. GOVERNING EQUATIONS: SUMMARY AND SPECIAL CASES

and Crocco’s8 equation. We will soon consider the Helmholtz vorticity transport equation
and Kelvin’s circulation theorem, in Secs. 7.4 and 7.5, respectively, that are also first integrals
of linear momenta in special cases.

6.7.1 Bernoulli’s equation

What we commonly call Bernoulli’s equation is really a first integral of the linear momenta
principle. Under different assumptions, we can get different flavors of Bernoulli’s equation.
A first integral of the linear momenta principle exists under the following conditions:
• viscous stresses are negligible relative to other terms, τij ∼ 0,
• the fluid is barotropic, p = p(ρ) or ρ = ρ(p).9
• body forces are conservative, so we can write fi = −∂i ϕ, where ϕ is a known potential
function, and
• either
– the flow is irrotational, ωk = ǫkij ∂i vj = 0, or
– the flow is steady, ∂o = 0.
First consider a version of the general linear momenta equation in non-conservative form,
Eq. (4.60) scaled by ρ:
1 1
∂o vi + vj ∂j vi = − ∂i p + fi + ∂j τji . (6.180)
ρ ρ
Now use our vector identity, Eq. (2.263), to rewrite the advective term, and impose our
assumptions to arrive at
 
1 1
∂o vi + ∂i vj vj − ǫijk vj ωk = − ∂i p − ∂i ϕ. (6.181)
2 ρ
Now let us define, just for this particular analysis, a new function P. We will take P to
be a function of pressure p, and thus implicitly, a function of xi and t. For the barotropic
fluid, we define P as
Z p(xi ,t)
dp̂
P(p(xi , t)) ≡ ρo . (6.182)
po ρ(p̂)
In the special case of incompressible flow with 1/ρ = 1/ρo , we have

P = p − po , incompressible. (6.183)
8
Luigi Crocco, 1909-1986, Sicilian-born, Italian applied mathematician and theoretical aerodynamicist
and rocket engineer, taught at University of Rome, Princeton, and Paris.
9
Three of the most common barotropic conditions are 1) a constant density fluid, ρ = C, 2) an isothermal
ideal gas, e.g. p = ρ(RT ), where R and T are constant, or 3) an adiabatic CPIG, e.g. p/po = (ρ/ρo )γ , where
γ is the ratio of specific heats.

© 05 December 2025, Joseph M. Powers. All rights reserved.

6.7. FIRST INTEGRALS OF LINEAR MOMENTA 217

In the special case of isothermal flow of an ideal gas with 1/ρ = RT /p, we have
p
P = ρo RT ln , isothermal ideal gas. (6.184)
po

In the special case of isentropic flow of CPIG with 1/ρ = (1/ρo )(p/po )−1/γ , see Eq. (6.132),
we have !
  γ−1
γ p γ
P= po −1 , isentropic CPIG. (6.185)
γ−1 po
Recalling Leibniz’s rule for one-dimension, Eq. (2.276),
Z x=b(t) Z x=b(t)
d db da
f (x, t) dx = ∂o f dx + f (b(t), t) − f (a(t), t), (6.186)
dt x=a(t) x=a(t) dt dt

we let ∂/∂xi play the role of d/dt to get

Z p(xi ,t) Z p(xi ,t)  
1 ∂P ∂ dp̂ 1 ∂p 1 ∂po ∂ 1
= = − + dp̂. (6.187)
ρo ∂xi ∂xi po ρ(p̂) ρ(p(xi , t)) ∂xi ρ(po ) ∂xi po ∂xi ρ(p̂)
|{z} | {z }
=0 =0

As po is constant, and the integrand has no explicit dependency on xi , we get

1 ∂P 1 ∂p
= . (6.188)
ρo ∂xi ρ ∂xi

So, our linear momenta principle reduces to

 
1 1
∂o vi + ∂i vj vj − ǫijk vj ωk = − ∂i P − ∂i ϕ, (6.189)
2 ρo
 T 
∂v v ·v 1
+∇ − v × ω = − ∇P − ∇ϕ. (6.190)
∂t 2 ρo

Consider now some special cases:

[Link] Irrotational case

If the fluid is irrotational, we have ωk = ǫklm ∂l vm = 0. Consequently, we can write the
velocity vector as the gradient of a potential function φ, known as the velocity potential:

∂m φ = vm . (6.191)

If the velocity takes this form, then the vorticity is

ωk = ǫklm ∂l ∂m φ. (6.192)

© 05 December 2025, Joseph M. Powers. All rights reserved.

218 CHAPTER 6. GOVERNING EQUATIONS: SUMMARY AND SPECIAL CASES

Because ǫklm is anti-symmetric and ∂l ∂m is symmetric, their tensor inner product must be
zero; hence, such a flow is irrotational: ωk = ǫklm ∂l ∂m φ = 0. So, the linear momenta
principle, Eq. (6.189), reduces to
 
1 1
∂o ∂i φ + ∂i (∂j φ)(∂j φ) = − ∂i P − ∂i ϕ, (6.193)
2 ρo
 
1 P
∂i ∂o φ + (∂j φ)(∂j φ) + +ϕ = 0, (6.194)
2 ρo
1 P
∂o φ + (∂j φ)(∂j φ) + + ϕ = f (t), (6.195)
2 ρo
∂φ 1 T P
+ ∇ φ · ∇φ + + ϕ = f (t). (6.196)
∂t 2 ρo

Here f (t) is an arbitrary function of time, that can be chosen to match conditions in a given
problem.

[Link] Steady case

[Link].1 Streamline integration Here we take ∂o = 0, but ωk 6= 0. Rearranging the
steady version of the linear momenta equation, Eq. (6.189), we get
 
1 1
∂i vj vj + ∂i P + ∂i ϕ = ǫijk vj ωk , (6.197)
2 ρo
 
1 P
∂i vj vj + +ϕ = ǫijk vj ωk . (6.198)
2 ρo

Taking the inner product of both sides with vi , we get

 
1 P
vi ∂i vj vj + +ϕ = vi ǫijk vj ωk , (6.199)
2 ρo
= ǫijk vi vj ωk , (6.200)
| {z }
=0
= 0. (6.201)

The term on the right hand side is zero because it is the tensor inner product of a symmetric
and anti-symmetric tensor.
For a local coordinate system that has component s aligned with the velocity vector vi ,
and the other two directions n, and b, mutually orthogonal, we have vi = (vs , 0, 0)T . Such
a system is sketched in Fig. 6.2, we will get many simplifications. Our linear momenta
principle then reduces to  
∂s []
(vs , 0, 0)  ∂n []  = 0. (6.202)
∂b []

© 05 December 2025, Joseph M. Powers. All rights reserved.

6.7. FIRST INTEGRALS OF LINEAR MOMENTA 219

v
b streamline: everywhere parallel
s to the velocity vector

n locally orthogonal
coordinate system
s, n, b

Figure 6.2: Local orthogonal intrinsic coordinate system oriented with local velocity field.

Forming this dot product yields

 
∂ 1 P
vs vj vj + +ϕ = 0. (6.203)
∂s 2 ρo
For vs 6= 0, we get that
1 P
vj vj + + ϕ = C(n, b). (6.204)
2 ρo
On a particular streamline, the function C(n, b) will be a constant.

[Link].2 Lamb surfaces We can extend the idea of integration along a streamline to
describe what are known as Lamb surfaces10 by again considering the steady, inviscid linear
momenta principle with conservative body forces, Eq. (6.198):
 
1 P
∂i vj vj + + ϕ = ǫijk vj ωk . (6.205)
2 ρo
Now taking the quantity B to be
1 P
B ≡ vj vj + + ϕ, (6.206)
2 ρo
the linear momenta principle, Eq. (6.198), becomes

∂i B = ǫijk vj ωk , ∇B = v × ω. (6.207)

This applies in the steady, inviscid, barotropic limit. In general, we can define the Lamb
vector, l as

l = −(v × ω). (6.208)

Thus, in the steady, inviscid, barotropic limit,

l = −∇B. (6.209)
10
Sir Horace Lamb, 1849-1934, English fluid mechanician, first studied at Owens College, Manchester
followed by mathematics at Cambridge, taught at Adelaide, Australia, then returned to the University of
Manchester; prolific writer of textbooks, including Hydrodynamics (1993).

© 05 December 2025, Joseph M. Powers. All rights reserved.

220 CHAPTER 6. GOVERNING EQUATIONS: SUMMARY AND SPECIAL CASES

Now the vector ǫijk vj ωk is orthogonal to both velocity vj and vorticity ωk because of the
nature of the cross product. Also the vector ∂i B is orthogonal to a surface on which B
is constant. Consequently, the surface on which B is constant must be tangent to both
the velocity and vorticity vectors. Surfaces of constant B thus are composed of families of
streamlines on which the Bernoulli constant has the same value. In addition they contain
families of vortex lines. These are the Lamb surfaces of the flow.

[Link] Irrotational, steady, incompressible case

In this case, we recover the form most commonly used (and misused) of Bernoulli’s equation,
namely,
1 P
vj vj + + ϕ = C. (6.210)
2 ρo

The constant is truly constant throughout the flow field. With P = p − po here and ϕ =
gz z + ϕo (with gz > 0, and rising z corresponding to rising distance from the earth’s surface,
we get f = −∇ϕ = −gz k) for a constant gravitational field, and v the magnitude of the
velocity vector, we get
1 2 p
v + + gz z = C. (6.211)
2 ρ

Here we have absorbed constants po /ρo and ϕo into C.

6.7.2 Crocco’s theorem

It is common, especially in texts on compressible flow, to present what is known as Crocco’s
theorem. The many different versions presented in many standard texts are non-uniform
and often of unclear validity. Its utility is confined mainly to providing an alternative way
of expressing the linear momenta principle that provides some insight into the factors that
influence fluid motion. In special cases, it can be integrated to form a more useful relation-
ship, similar to Bernoulli’s equation, between fundamental fluid variables. The heredity of
this theorem is not always clear, though, as we shall see it is nothing more than a combina-
tion of the linear momenta principle coupled with some definitions from thermodynamics.
Its derivation is often confined to inviscid flows. Here we will first present a result valid
for general viscous flows for the evolution of stagnation enthalpy, that is closely related to
Crocco’s theorem. As introduced earlier on p. 101, a stagnation property, such as stagnation
enthalpy, is the value the property acquires when the fluid is brought to rest at a stagnation
point. Next we will show how one of the restrictions can be relaxed so as to obtain what we
call the extended Crocco’s theorem. We then show how this reduces to a form that is similar
to a form presented in many texts.

© 05 December 2025, Joseph M. Powers. All rights reserved.

6.7. FIRST INTEGRALS OF LINEAR MOMENTA 221

[Link] Stagnation enthalpy variation

First, again consider the general linear momenta equation, Eq. (6.180):
1 1
∂o vi + vj ∂j vi = − ∂i p + fi + ∂j τji . (6.212)
ρ ρ
Now, as before in the development of Bernoulli’s equation, use our vector identity, Eq. (2.263),
to rewrite the advective term, but retain the viscous terms to get
 
1 1 1
∂o vi + ∂i vj vj − ǫijk vj ωk = − ∂i p + fi + ∂j τji . (6.213)
2 ρ ρ

Taking the dot product with vi , and rearranging, we get

   
1 1 1 1
∂o vi vi + vi ∂i vj vj = ǫijk vi vj ωk − vi ∂i p + vi fi + vi ∂j τji . (6.214)
2 2 | {z } ρ ρ
=0

Again, because ǫijk is anti-symmetric and vi vj is symmetric, their tensor inner product is
zero, so we get
   
1 1 1 1
∂o vi vi + vi ∂i vj vj = − vi ∂i p + vi fi + vi ∂j τji . (6.215)
2 2 ρ ρ

Now recall the Gibbs relation from thermodynamics, Eq. (4.164):

p
T ds = de − dρ. (6.216)
ρ2
Also recall the definition of enthalpy h, Eq. (4.136):
p
h=e+ . (6.217)
ρ
Differentiating the equation for enthalpy, we recover Eq. (4.157):
1 p
dh = de + dp − 2 dρ. (6.218)
ρ ρ
Eliminating de in favor of dh in the Gibbs equation gives
1
T ds = dh − dp. (6.219)
ρ
If we choose to apply this relation to the motion following a fluid particle, we can say then
that
ds dh 1 dp
T = − . (6.220)
dt dt ρ dt

© 05 December 2025, Joseph M. Powers. All rights reserved.

222 CHAPTER 6. GOVERNING EQUATIONS: SUMMARY AND SPECIAL CASES

Expanding, we get
1
T (∂o s + vi ∂i s) = ∂o h + vi ∂i h − (∂o p + vi ∂i p). (6.221)
ρ

Rearranging, we get
1 1
T (∂o s + vi ∂i s) − (∂o h + vi ∂i h) + ∂o p = − vi ∂i p. (6.222)
ρ ρ

We then use this identity to eliminate the pressure gradient term from the linear momenta
equation in favor of enthalpy, entropy, and unsteady pressure terms:
   
1 1 1 1
∂o vi vi + vi ∂i vj vj = T (∂o s + vi ∂i s) − (∂o h + vi ∂i h) + ∂o p + vi fi + vi ∂j τji . (6.223)
2 2 ρ ρ

Rearranging slightly, noting that vi vi = vj vj , and assuming the body force is conservative so
that fi = −∂i ϕ, we get
   
1 1 1 1
∂o h + vj vj + ϕ + vi ∂i h + vj vj + ϕ = T (∂o s + vi ∂i s) + ∂o p + vi ∂j τji . (6.224)
2 2 ρ ρ

Here we have made the common assumption that the body force potential ϕ is independent
of time, that allows us to absorb it within the time derivative. If we define, as is common,
the stagnation enthalpy ho as
1
ho = h + vj vj + ϕ, (6.225)
2
we can then state
1 1
∂o ho + vi ∂i ho = T (∂o s + vi ∂i s) + ∂o p + vi ∂j τji , (6.226)
ρ ρ
dho ds 1 ∂p 1 T
T
= T + + v · ∇T · τ (6.227)
dt dt ρ ∂t ρ

The stagnation enthalpy is sometimes known as the total enthalpy. We can use the first
law of thermodynamics written in terms of entropy, Eq. (4.168), ρ(ds/dt) = −(1/T )∂i qi +
(1/T )τij ∂i vj , to eliminate the entropy derivative in favor of those terms that generate entropy
to arrive at
dho
ρ = ∂i (τij vj − qi ) + ∂o p, (6.228)
dt
dho ∂p
ρ = ∇T · (τ · v − q) + . (6.229)
dt ∂t
Thus, we see that the total enthalpy of a fluid particle is influenced by energy and momentum
diffusion as well as an unsteady pressure field.

© 05 December 2025, Joseph M. Powers. All rights reserved.

6.7. FIRST INTEGRALS OF LINEAR MOMENTA 223

[Link] Extended Crocco’s theorem

With a slight modification of the preceding analysis, we can arrive at the extended Crocco’s
theorem. Begin once more with an earlier version of the linear momenta principle, Eq. (6.213):
 
1 1 1
∂o vi + ∂i vj vj − ǫijk vj ωk = − ∂i p + fi + ∂j τji . (6.230)
2 ρ ρ
Now assume we have a functional representation of enthalpy in the form
h = h(s, p). (6.231)
Then we get
∂h ∂h
dh = ds + dp. (6.232)
∂s p ∂p s
We also thus deduce from the Gibbs relation dh = T ds + (1/ρ) dp that
∂h ∂h 1
= T, = . (6.233)
∂s p ∂p s ρ
Now, because we have h = h(s, p), we can take its derivative with respect to each and all of
the coordinate directions to obtain
∂h ∂h ∂s ∂h ∂p
= + . (6.234)
∂xi ∂s p ∂xi ∂p s ∂xi
or
∂h ∂h
∂i h = ∂i s + ∂i p. (6.235)
∂s p ∂p s
Substituting known values for the thermodynamic derivatives, we get
1
∂i h = T ∂i s + ∂i p. (6.236)
ρ
We can use this to eliminate directly the pressure gradient term from the linear momenta
equation to obtain then
 
1 1
∂o vi + ∂i vj vj − ǫijk vj ωk = T ∂i s − ∂i h + fi + ∂j τji . (6.237)
2 ρ
Rearranging slightly, and again assuming the body force is conservative so that fi = −∂i ϕ,
we get the extended Crocco’s theorem:
 
1 1
∂o vi + ∂i h + vj vj + ϕ = T ∂i s + ǫijk vj ωk + ∂j τji . (6.238)
2 ρ
Again, employing the total enthalpy, ho = h + 21 vj vj + ϕ, we write the extended Crocco’s
theorem as
1
∂o vi + ∂i ho = T ∂i s + ǫijk vj ωk + ∂j τji , (6.239)
ρ
∂v 1
+ ∇ho = T ∇s + v × ω + ∇T · τ . (6.240)
∂t ρ

© 05 December 2025, Joseph M. Powers. All rights reserved.

224 CHAPTER 6. GOVERNING EQUATIONS: SUMMARY AND SPECIAL CASES

[Link] Traditional Crocco’s theorem

For a steady, inviscid flow, the extended Crocco’s theorem reduces to what is usually called
Crocco’s theorem:

∂i ho = T ∂i s + ǫijk vj ωk , (6.241)
∇ho = T ∇s + v × ω. (6.242)

If the flow is further required to be homeoentropic, we get

∂i ho = ǫijk vj ωk , (6.243)
∇ho = v × ω. (6.244)

Similar to Lamb surfaces, we find that surfaces on which ho is constant are parallel to both
the velocity and vorticity vector fields. Taking the dot product with vi , we get

vi ∂i ho = vi ǫijk vj ωk , (6.245)
= ǫijk vi vj ωk , (6.246)
= 0. (6.247)

Integrating this along a streamline, as for Bernoulli’s equation, we find

ho = C(n, b), (6.248)

1
h + vj vj + ϕ = C(n, b), (6.249)
2
so we see that the stagnation enthalpy is constant along a streamline and varies from stream-
line to streamline. If the flow is steady, homeoentropic, and irrotational, the total enthalpy
will be constant throughout the flow-field:

1
h + vj vj + ϕ = C. (6.250)
2
In terms of internal energy, we can rewrite this as

1 p
e + vj vj + + ϕ = C. (6.251)
2 ρ

This is in a remarkably similar form to the Bernoulli equation for a steady, incompressible,
irrotational fluid, Eq. (6.211). However, the assumptions for each are very different. Bernoulli
made no appeal to the first law of thermodynamics, while Crocco did. This version of the
Bernoulli equation is restricted to incompressible flows, while this version of the Crocco
equation is fully compressible.

© 05 December 2025, Joseph M. Powers. All rights reserved.

6.8. FLUID ANALOG TO MAXWELL-HEAVISIDE EQUATIONS 225

6.8 Fluid analog to Maxwell-Heaviside equations

Let us examine a special limit in which the Navier-Stokes equations can be cast in a form
analogous to the Maxwell11 -Heaviside12 equations of electromagnetics. Here we are guided
by the studies of Marmanis13 One can consult Jackson (1999) for detailed background of the
Maxwell-Heaviside equations.

6.8.1 Development
Consider a constant density, isotropic, Newtonian fluid with constant viscosity and conserva-
tive body force. Mass and momentum evolution, Eqs. (6.50, 6.58), reduce to a divergence-free
velocity field and a viscous extension of Eq. (6.181):

∇T · v = 0, (6.252)
 
∂v 1 T p
= v×ω−∇ v · v + + ϕ + ν∇2 v. (6.253)
∂t 2 ρ
Defining

v 2 = vT · v, (6.254)

and then specializing Eq. (6.206), we have

1 p
B = v 2 + + ϕ. (6.255)
2 ρ
This lets us rewrite the momentum equation, Eq. (6.253), as
∂v
= v × ω − ∇B + ν∇2 v. (6.256)
∂t
Now if we take the curl of Eq. (6.256) and recall the curl of the gradient of a scalar is zero,
we get
∂ω
= ∇ × (v × ω) + ν∇2 ω. (6.257)
∂t
Now, recall Eq. (6.208) for the Lamb vector:

l = −(v × ω). (6.258)

If we are given the velocity v, we can compute ω = ∇ × v and then l = −(v × ω). However
if we are given l and ω, we cannot compute a unique v. That is because any component
11
James Clerk Maxwell, 1831-1879, Scottish physicist and mathematician.
12
Oliver Heaviside, 1850-1925, English physicist and mathematician.
13
Marmanis, H., 1999, “Analogy between the Navier–Stokes equations and Maxwell’s equations: Applica-
tion to turbulence,” Physics of Fluids, 10(6): 1428-1437.

© 05 December 2025, Joseph M. Powers. All rights reserved.

226 CHAPTER 6. GOVERNING EQUATIONS: SUMMARY AND SPECIAL CASES

of v that is parallel to ω will not not influence l. An unusual flow that happens to have
the velocity vector parallel to the vorticity vector, v = αω will have l = 0 and is known as
a Beltrami14 flow. Beltrami flows must be fully three-dimensional, as a flow confined to a
plane will have a vorticity vector normal to the plane.
The definition of l allows us to rewrite Eq. (6.256) as a version of Eq. (6.209) extended
to the unsteady, viscous limit:

∂v
l=− − ∇B + ν∇2 v. (6.259)
∂t
We take the curl of this equation and get

∂
∇×l = − (∇ × v) − |∇ ×{z∇B} +ν∇2 (∇ × v), (6.260)
∂t
0
∂ω
= − + ν∇2 ω. (6.261)
∂t
We can take the divergence of Eq. (6.259) to get

∂
∇T · l = − (∇
|
T
{z· v}) − ∇2 B + ν∇2 (∇
|
T
{z· v}). (6.262)
∂t
0 0

Because of incompressibility, this reduces to

∇T · l = −∇2 B. (6.263)

Let us define the scalar field q as the fluid charge density:

q ≡ −∇2 B. (6.264)

This has an equivalent interpretation in electromagnetic theory. Thus, Eq. (6.263) reduces
to

∇T · l = q. (6.265)

Next recall that the divergence of the curl of any vector must be zero, so we have

∇T · (∇ × v) = ∇T · ω = 0. (6.266)

Now take the time derivative of Eq. (6.259) to get

∂l ∂2v ∂ ∂v
= − 2 − ∇B + ν∇2 . (6.267)
∂t ∂t ∂t ∂t
14
Eugenio Beltrami, 1835-1900, Italian mathematician.

© 05 December 2025, Joseph M. Powers. All rights reserved.

6.8. FLUID ANALOG TO MAXWELL-HEAVISIDE EQUATIONS 227

Now from the definition of l in Eq. (6.258), we can take a derivative with respect to time
to get
∂l ∂ω ∂v
= ×v+ω× . (6.268)
∂t ∂t ∂t
Next invoke Eqs. (6.259) and (6.261) and to get
∂l


= −∇ × l + ν∇2 ω × v + ω × −l − ∇B + ν∇2 v , (6.269)
∂t

= − (∇ × l) × v + ω × (−l − ∇B) + ν ∇2 ω × v + ω × ∇2 v . (6.270)

Now it is possible to prove the vector identity

− (∇ × l) × v − ω × l = −(vT · ∇)l − (lT · ∇)v. (6.271)

Thus, Eq. (6.270) can be rewritten as

∂l

= −(vT · ∇)l − (lT · ∇)v − ω × ∇B + ν ∇2 ω × v + ω × ∇2 v . (6.272)
∂t
With effort, one can show the viscous term can be recast as


T

∇2 ω × v + ω × ∇2 v = vT · ∇2 ∇v − (∇v)T + ∇2 v · ∇v − (∇v)T . (6.273)

We will not use this form, but it does illustrate the role of the anti-symmetric part of the
velocity gradient tensor.
We also have the vector identity

∇ × (v × l) = v (∇T · l) −l (∇T · v) +(lT · ∇)v − (vT · ∇)l, (6.274)

| {z } | {z }
q 0

= vq + (l · ∇)v − (vT · ∇)l,

T
(6.275)
−(vT · ∇)l = ∇ × (v × l) − vq − (lT · ∇)v. (6.276)

Using this, Eq. (6.272) is rewritten as

∂l

= ∇ × (v × l) − vq − 2(lT · ∇)v − ω × ∇B + ν ∇2 ω × v + ω × ∇2 v . (6.277)
∂t
Recalling v 2 = vT · v, it is possible to prove the vector identity

∇ × (v × l) = v 2 ∇ × ω + ∇v 2 × ω − ∇ × ((vT · ω)v). (6.278)

With this Eq. (6.277) becomes

∂l
= v 2 ∇ × ω − vq − ∇ × ((vT · ω)v) − ω × ∇(B + v 2 ) − 2(lT · ∇)v
∂t

+ν ∇2 ω × v + ω × ∇2 v . (6.279)

© 05 December 2025, Joseph M. Powers. All rights reserved.

228 CHAPTER 6. GOVERNING EQUATIONS: SUMMARY AND SPECIAL CASES

Let us now define the fluid current vector J as

J = vq + ∇ × ((vT · ω)v) + ω × ∇(B + v 2 ) + 2(lT · ∇)v − ν ∇2 ω × v + ω × ∇2 v .
(6.280)

Here J has no relation to the Jacobian matrix that has the same symbol. Then Eq. (6.279)
becomes
∂l
= v 2 ∇ × ω − J. (6.281)
∂t

6.8.2 The fluid Maxwell-Heaviside equations

We now write as a group Eqs. (6.265, 6.266, 6.261, 6.281) the fluid Maxwell-Heaviside equa-
tions:

∇T · l = q, (6.282)
∇T · ω = 0, (6.283)
∂ω
∇×l = − , (6.284)
∂t
1 1 ∂l
∇×ω = 2
J+ 2 . (6.285)
v v ∂t
These have the precise form of the Maxwell-Heaviside equations of electromagnetism. The
Lamb vector l plays the role of the electric field vector; the vorticity plays the role of the
magnetic field vector; the variable q plays the role of charge density scaled by permittivity;
the vector J plays the role of the current density vector.
We see the following:
• A time-varying vorticity vector field induces a Lamb vector field.

• A time-varying Lamb vector field or a fluid current vector induces a vorticity vector
field.

• The vorticity vector is orthogonal to the Lamb vector.

• The velocity vector is orthogonal to the Lamb vector.

Similar to electromagnetics, one can define a Poynting15 vector p as

p = l × ω. (6.286)

Certainly p, l, and ω form an orthogonal triple. In electromagnetics, one can show that
the Poynting vector points in the direction of energy propagation; the interpretation in fluid
mechanics is not straightforward.
15
John Henry Poynting, 1852-1914, English physicist.

© 05 December 2025, Joseph M. Powers. All rights reserved.

6.8. FLUID ANALOG TO MAXWELL-HEAVISIDE EQUATIONS 229

We note that by taking the time derivative of Eq. (6.282), we get

∂l ∂q
∇T · = . (6.287)
∂t ∂t
Then use Eq. (6.285) to eliminate ∂l/∂t:

∂q
∇T · v 2 ∇ × ω − J = . (6.288)
∂t
This reduces to
∂q

+ ∇T · J = ∇ T · v 2 ∇ × ω , (6.289)
∂t
= −(∇T v 2 ) · ∇2 v. (6.290)

In contrast to electromagnetics, we can only interpret J as the flux of q in the limit where
v 2 is a constant. In electromagnetics, we find the constant square of the speed of light, c2 ,
instead of the variable v 2 .

Example 6.2
Analyze the simple rotational flow of an incompressible isotropic Newtonian fluid with velocity field
 ω0 y 
− 2
v =  ω20 x  , (6.291)
0

in the context of the fluid Maxwell-Heaviside equations. Ignore body force effects; take the pressure
p(0, 0, 0) = p0 .

Calculate some basic quantities. The vorticity vector field is given by

 
i j k 0
∂ ∂ ∂ 
ω = ∇ × v = ∂x ∂y ∂z = 0 . (6.292)
− ω20 y ω20 x 0 ω0

The Lamb vector is

 
ω2x
i j k − 20
 
l = −(v × ω) = ω × v = 0 0 ω0 =  − ω02 y  . (6.293)
2
− ω20 y ω0 x
2 0 0
We can verify incompressibility:
 
T ∂ −ω0 y ∂  ω0 x  ∂
∇ ·v = + + (0) = 0. (6.294)
∂x 2 ∂y 2 ∂z
We find the fluid charge density q to enable the first fluid Maxwell-Heaviside equation, Eq. (6.282), to
be satisfied:
   
T ∂ −ω02 x ∂ −ω02 y ∂
q = ∇ ·l = + + (0) = −ω02 . (6.295)
∂x 2 ∂y 2 ∂z

© 05 December 2025, Joseph M. Powers. All rights reserved.

230 CHAPTER 6. GOVERNING EQUATIONS: SUMMARY AND SPECIAL CASES

By inspection the second fluid Maxwell-Heaviside equation, Eq. (6.283), is satisfied:

∂ ∂ ∂
∇T · ω = (0) + (0) + (ω0 ) = 0. (6.296)
∂x ∂y ∂z
The third fluid Maxwell-Heaviside equation, Eq. (6.284), is satisfied as we have
   
i j k 0 0
∂ω ∂ ∂ ∂ ∂    
∇×l+ = ∂x ∂y ∂z + 0 = 0 . (6.297)
∂t ω2 x ω2 y ∂t
− 20 − 20 0 ω0 0

Because our l is curl-free, it is the gradient of a potential Φ: l = −∇Φ. Here we have

ω02 2
Φ= (x + y 2 ). (6.298)
4
We also have Φ satisfying the Poisson equation −∇2 Φ = q. For the fourth fluid Maxwell-Heaviside
equation, Eq. (6.285), we see that ∇ × ω = 0, ∂l/∂t = 0. This then requires that the fluid current
vector J = 0. Detailed calculation for our v and ω that Eq. (6.280) yields J = 0 if we select the
pressure field to be
ρω02 2

p = p0 + x + y2 . (6.299)
8
This choice yields
p0 ω2

B= + 0 x2 + y 2 . (6.300)
ρ 4
6 Φ, but that −∇2 B = −∇2 Φ = q.
Note that B =
Because ω is curl-free, it represents the gradient of a potential Ψ via ω = −∇Ψ. Here we have
Ψ = −ω0 z. (6.301)
We have Ψ satisfying the Laplace equation ∇2 Ψ = 0. Note the viscosity plays no role in this analysis
as the fluid rotates as a solid body with zero deformation:
∇v + (∇v)T
D= = 0. (6.302)
2
We show in Fig. 6.3 the velocity vector field v and the Lamb vector field l for this flow. The velocity
field is such that a fluid particle is rotating about the origin. The Lamb vectors point towards the
origin. Both v and l have no variation with z and may be considered as confined to the (x, y) plane. In
this plane, it is obvious that l is orthogonal to v. The vorticity vector field is uniform and points out of
the page. Thus, ω is orthogonal to both l and v. Thus, the three vectors form a mutually orthogonal
set:
lT · v = lT · ω = vT · ω = 0. (6.303)
We also note that our velocity and pressure fields are consistent with the Poisson equation for
pressure, Eq. (6.76), in that
 
T T 2
ω · ω  = − ρω · ω = − ρω0 .
−∇2 p = ρ D : D} −
| {z (6.304)
2 2 2
0

© 05 December 2025, Joseph M. Powers. All rights reserved.

6.8. FLUID ANALOG TO MAXWELL-HEAVISIDE EQUATIONS 231

y 0

-1

-2
-2 -1 0 1 2
x

Figure 6.3: Velocity vector field v and Lamb vector field l for a fluid rotating as a non-
deforming solid body about the origin.

6.8.3 Statics with q and J

Let us now consider the static limit in which there is no dependency on time; we retain the
effect of q and J source terms. In this limit, Eqs. (6.282-6.285), reduce to

∇T · l = q, (6.305)
∇T · ω = 0, (6.306)
∇ × l = 0, (6.307)
1
∇ × ω = 2 J. (6.308)
v
In this limit, Eq. (6.307) tells us the vector field l is curl-free. This allows us to take the
potential Φ to be defined such that

l = −∇Φ. (6.309)

As with all potentials, Φ is not unique. For example if we add a constant to get Φ′ = Φ + C,
we get the same physical quantity l as ∇Φ = ∇Φ′ . With this definition of Φ, Eq. (6.305),
reduces to a Poisson equation:

−∇T · ∇Φ = −∇2 Φ = q, (6.310)

when we interpret q = q(x, y, z).

Now because of Eq. (6.308), ω is not curl-free, and so it is not the gradient of a scalar
potential. But because it is divergence-free, it satisfies the definition of a solenoidal vector

© 05 December 2025, Joseph M. Powers. All rights reserved.

232 CHAPTER 6. GOVERNING EQUATIONS: SUMMARY AND SPECIAL CASES

field. Vector calculus tells us for solenoidal fields that we can expect ω to be the curl of a
vector function we will call the vector potential. And in fact the definition of vorticity gives
us the vector potential that is v:
ω = ∇ × v. (6.311)
In contrast to electromagnetics, the vector potential is completely defined in terms of a
physical quantity, the fluid velocity v. Electromagnetics does not require the vector potential
to be uniquely defined. Vector calculus tells us that ∇T · ω = ∇T · (∇ × v) = 0 is identically
true. We next substitute into Eq. (6.308) to get
1
∇×∇×v = J. (6.312)
v2
Using a standard vector calculus identity we rewrite this as
1
∇(∇T · v) − ∇2 v = J. (6.313)
v2
Because ∇T · v = 0, Eq. (6.312) becomes
1
−∇2 v = J. (6.314)
v2
So Eqs. (6.310) and (6.314) are both Poisson equations for scalar and vector potentials Φ
and v, respectively. These are fully equivalent to the steady equations in terms of l and ω.

6.8.4 Statics with no q and J

Let us now consider the static limit and neglect source terms, q = 0, J = 0. In this limit,
Eqs. (6.282-6.285), reduce to
∇T · l = 0, (6.315)
∇T · ω = 0, (6.316)
∇×l = 0, (6.317)
∇×ω = 0. (6.318)
So in this limit, the vector fields l and ω are curl-free, and both can be expressed as the
gradients of a potential. We can again employ Eq. (6.309), l = −∇Φ, and with this definition,
Eq. (6.315) becomes Laplace’s equation:
∇T · ∇Φ = ∇2 Φ = 0, (6.319)
and we can use a variety of methods to solve. In three-dimensions, we can adapt the point
source solution, setting aside dimensional scaling constants, to get for an isolated point
source at the origin
1
Φ(r) = . (6.320)
4πr

© 05 December 2025, Joseph M. Powers. All rights reserved.

6.8. FLUID ANALOG TO MAXWELL-HEAVISIDE EQUATIONS 233

The Lamb vector for the point source is

dΦ 1
l = −∇Φ = − er + 0eθ + 0eφ = er . (6.321)
dr 4πr 2
The Lamb vector magnitude decays as 1/r 2.

Example 6.3
Consider a three-dimensional domain with a point source at (x, y, z) = (1, 1, 0) and a second point
source at (x, y, z) = (−1, −1, 0). Find the potential field and the Lamb vector field.

A point source centered at the origin of a three-dimensional field has

1 1
Φ= = p . (6.322)
4πr 4π x + y 2 + z 2
2

We know that this satisfies −∇T · ∇Φ = δ(x)δ(y)δ(z), with the Dirac delta functions each of unit
strength. A pair of sources centered at (1, 1, 0) and at (−1, −1, 0) has the field that arises from super-
position of two fields:
1 1
Φ= p + p . (6.323)
2 2
4π (x − 1) + (y − 1) + z 2 4π (x + 1) + (y + 1)2 + z 2
2
| {z } | {z }
source source

It satisfies
 
T ∂2Φ ∂2Φ ∂2Φ
−∇ · ∇Φ = − + + = δ(x − 1)δ(y − 1)δ(z) + δ(x + 1)δ(y + 1)δ(z). (6.324)
∂x2 ∂y 2 ∂z 2

Direct substitution of the field into Laplace’s equation reveals that as long as (x, y) 6= (1, 1), (x, y) 6=
(−1, −1), the Laplacian is zero. So it is a solution. The Lamb vector is
 x−1 x+1 
4π((x−1)2 +(y−1)2 +z 2 )3/2
+ 4π((x+1)2 +(y+1) 2 +z 2 )3/2
 y−1 y+1 
l = −∇Φ =  4π((x−1)2 +(y−1) 2 +z 2 )3/2
+ 4π((x+1)2 +(y+1) 2 +z 2 )3/2  . (6.325)
z z
4π((x−1)2 +(y−1)2 +z 2 )3/2
+ 4π((x+1)2 +(y+1)2 +z 2 )3/2

Confining attention to the plane of z = 0, a plot of the potential and Lamb vector field is shown in
Fig. 6.4a. We show in Fig. 6.4b the three-dimensional field vector.

Similarly for the curl-free static vorticiy field ω, we can take the scalar potential Ψ to be
defined such that

ω = −∇Ψ. (6.326)

Substituting into Eq. (6.316), we again get Laplace’s equation:

∇T · ∇Ψ = ∇2 Ψ = 0. (6.327)

© 05 December 2025, Joseph M. Powers. All rights reserved.

234 CHAPTER 6. GOVERNING EQUATIONS: SUMMARY AND SPECIAL CASES

2 0.08 0.1
0.08

1 0.2
0.15
0.1
0

0.15
-1 0.2 0.08

0.08 0.1
-2 0.1
-2 -1 0 1 2
a) b)

Figure 6.4: Superposition of two point sources in a three-dimensional geometry: a) potential

contours, Φ(x, y, z = 0), and Lamb field vector, l(x, y, z = 0), b) Lamb vector field l(x, y, z).

Source-sink pairs are admitted.

Example 6.4
Consider a three-dimensional domain with a point source of vorticity at (x, y, z) = (1, 1, 0) and
a point sink of magnetism at (x, y, z) = (−1, −1, 0). Find the scalar potential field and the vorticity
vector field.

A source-sink pair centered at (1, 1, 0) and at (−1, −1, 0) has the scalar potential field that arises
from superposition of two fields:
1 1
Ψ= p − p . (6.328)
4π (x − 1)2 + (y − 1)2 + z 2 4π (x + 1)2 + (y + 1)2 + z 2
| {z } | {z }
source sink

It satisfies
 
T ∂2Ψ ∂2Ψ ∂2Ψ
−∇ · ∇Ψ = − + + = δ(x − 1)δ(y − 1)δ(z) − δ(x + 1)δ(y + 1)δ(z). (6.329)
∂x2 ∂y 2 ∂z 2
Direct substitution of the scalar potential field into Laplace’s equation reveals that as long as (x, y) 6=
(1, 1), (x, y) 6= (−1, −1), the Laplacian is zero. So it is a solution. The vorticity vector field is
 x−1 x+1 
4π((x−1)2 +(y−1)2 +z 2 )3/2
− 4π((x+1)2 +(y+1) 2 +z 2 )3/2
 y−1 y+1 
ω = −∇Ψ =  4π((x−1)2 +(y−1) 2 +z 2 )3/2
− 4π((x+1)2 +(y+1) 2 +z 2 )3/2  . (6.330)
z z
4π((x−1)2 +(y−1)2 +z 2 )3/2
− 4π((x+1)2 +(y+1)2 +z2 )3/2

© 05 December 2025, Joseph M. Powers. All rights reserved.

6.8. FLUID ANALOG TO MAXWELL-HEAVISIDE EQUATIONS 235

2 0.01 0.05

0.05
0.2
1

- 0.02 0.1

y 0
0.02
- 0.1

-1
- 0.2
- 0.05
- 0.05 0
-2 - 0.01

-2 -1 0 1 2
x
a) b)

Figure 6.5: Superposition of a source-sink pair of static vorticity in a three-dimensional

geometry: a) scalar potential contours, Ψ(x, y, z = 0), and vorticity vector field, ω(x, y, z =
0), b) vorticity vector field ω(x, y, z).

Confining attention to the plane of z = 0, a plot of the scalar potential and vorticity vector field is
shown in Fig. 6.5a. We show in Fig. 6.5b the three-dimensional vorticity vector field. We clearly see
the vorticity emanating from the source at (x, y, z) = (1, 1, 0) and flowing into the sink at (x, y, z) =
(−1, −1, 0).

6.8.5 Dynamics with q and J

Let us reformulate the full equations, Eqs. (6.282-6.285), into two uncoupled wave equations
with source terms due to q and J. We will employ useful and non-unique potential functions
to achieve this formulation. This will entail some analysis techniques that require intricate
and subtle use of vector calculus. We will follow the presentation given Feynman, et al.
(1963) for electromagnetics.
First consider Eq. (6.283):
∇T · ω = 0. (6.331)
Now because of Eq. (6.285), ω is not curl-free, and so it is not the gradient of a scalar
potential. But because it is divergence-free, it is solenoidal, and similar to what we did in
Sec. 6.8.3, we can define a vector potential v via
ω = ∇ × v. (6.332)

© 05 December 2025, Joseph M. Powers. All rights reserved.

236 CHAPTER 6. GOVERNING EQUATIONS: SUMMARY AND SPECIAL CASES

Here v is simply the velocity vector. Vector calculus then tells us that ∇T ·ω = ∇T ·(∇×v) =
0 is identically true.
Let us use our definition of ω to recast Eq. (6.284) as
∂
∇×l=− ∇ × v. (6.333)
∂t
Let us interchange the time derivative and curl operators to write this as
 
∂v
∇× l+ = 0. (6.334)
∂t
Therefore, l + ∂v/∂t must be a gradient that is curl-free, and therefore, it must be the
gradient of a potential. Let us then extend our static analysis embodied in Eq. (6.309) to
define the potential Φ such that
∂v
l+ = −∇Φ. (6.335)
∂t
Let us use this to eliminate l in Eq. (6.282) to get
 
T ∂v
∇ · −∇Φ − = q, (6.336)
∂t
∂ T
−∇2 Φ − ∇ · v = q, (6.337)
∂t | {z }
0
−∇2 Φ = q. (6.338)

This Poisson equation relates Φ to q.

Next operate on Eq. (6.285):
1 1 ∂l
∇×ω = 2
J+ 2 , (6.339)
v v ∂t
1 ∂l 1
∇×ω− 2 = 2 J. (6.340)
v ∂t v
Now eliminate ω and l by using Eqs. (6.332) and (6.335):
 
1 ∂ ∂v 1
∇ × (∇ × v) − 2 −∇Φ − = 2 J. (6.341)
| {z } v ∂t ∂t v
ω | {z }
l

This relates Φ and v in a coupled fashion to our source J. Use a standard vector calculus
identity to expand the term ∇ × (∇ × v) to get

1 ∂ 1 ∂2v 1
∇(∇T · v) − ∇2 v + 2 ∇Φ + 2 2 = 2 J, (6.342)
| {z } v ∂t v ∂t v
∇×(∇×v)

© 05 December 2025, Joseph M. Powers. All rights reserved.

6.8. FLUID ANALOG TO MAXWELL-HEAVISIDE EQUATIONS 237
!
T 1 ∂Φ 2 1 ∂2v 1
∇ ∇
| {z· v} + v 2 ∂t − ∇ v + 2 2
= 2 J, (6.343)
v ∂t v
0
 
1 ∂Φ 1 ∂2v 1
∇ − ∇2 v + = 2 J, (6.344)
v 2 ∂t 2
v ∂t 2 v
2
1 ∂ v 1 1 ∂
−∇2 v + 2 2 = 2 J − 2 ∇Φ, (6.345)
v ∂t v v ∂t
Now Eqs. (6.338) and (6.345) have removed the fields l and ω in favor of the potentials Φ
and v. They are driven by q and J. Rearrange to get
∂2v ∂
2
= v 2 ∇2 v + J − ∇Φ. (6.346)
∂t ∂t
This is a type of inhomogeneous wave equation with forcing due to J and q. However the
wave speed v is not constant:

6.8.6 Dynamics with no q or J

We now consider time-dependent dynamics in a space containing with q = 0 and J = 0.
Thus limited, Eqs. (6.282-6.285) reduce to

∇T · l = 0, (6.347)
∇T · ω = 0, (6.348)
∂ω
∇×l = − , (6.349)
∂t
1 ∂l
∇×ω = 2 . (6.350)
v ∂t
Let us now take the curl of Eq. (6.349) to get
∂
(∇ × ω) = −∇ × ∇ × l. (6.351)
∂t
Now use Eq. (6.350) and the vector identity ∇ × ∇ × l = ∇(∇T · l) − ∇2 l to reduce to
1 ∂2l
= −∇(∇T · l) + ∇2 l. (6.352)
v 2 ∂t2
Then employ Eq. (6.347) to get
∂2l
= v 2 ∇2 l. (6.353)
∂t2
Let us now take the curl of Eq. (6.350) to get
1 ∂
(∇ × l) = ∇ × ∇ × ω. (6.354)
v 2 ∂t

© 05 December 2025, Joseph M. Powers. All rights reserved.

238 CHAPTER 6. GOVERNING EQUATIONS: SUMMARY AND SPECIAL CASES

Now use Eq. (6.349) and the same well-known vector identity ∇ × ∇ × ω = ∇(∇T · ω) − ∇2 ω
to reduce to
1 ∂2ω
− 2 2
= ∇(∇T · ω) − ∇2 ω. (6.355)
v ∂t
Then employ Eq. (6.348) to get

∂2ω
2
= v 2 ∇2 ω. (6.356)
∂t
Note that Eqs. (6.353, 6.356) are similar to the wave equations for propagating electrical
and magnetic field vectors. Electromagnetic theory replaces the fluid velocity magnitude v
with the speed of light. And because the speed of light is a constant, electromagnetic theory
yields linear wave equations with d’Alembert16 solutions. In contrast, our fluid Maxwell-
Heaviside equations are nonlinear, as v is not a constant. The possibility of traveling wave
solutions being admitted by the fluid Maxwell-Heaviside equations is remarkable and unusual
as these are admitted by a theory that is both incompressible and viscous. In fluid mechanics,
we typically study waves in the context of compressible inviscid flows in which the speed of
propagation is related to the sound speed of the fluid.

16
Jean le Rond d’Alembert, 1717-1783, French mathematician.

© 05 December 2025, Joseph M. Powers. All rights reserved.

 Part II

Governing equations: solutions

239
Chapter 7

Vortex dynamics

see Powers, Chapter 7,

see Panton, Chapter 13,
see Yih, Chapter 2,
see Kuthe and Chow, Chapter 5,
see Lamb, Chapter 7,
see Saffman.

In this chapter we will consider in detail the kinematics and dynamics of rotating fluids,
sometimes called vortex dynamics. The two most common quantities that are used to char-
acterize rotating fluids are the vorticity vector, Eq. (3.112):

ω = ∇ × v, (7.1)

and a new scaler quantity we define as the circulation, Γ:

I
Γ= vT · dr. (7.2)
C
H
Here C is the integral about a closed contour C. Both concepts will be important here.
Although it is possible to use Cartesian index notation to describe a rotating fluid, some
of the ideas are better conveyed in a non-Cartesian system, such as the cylindrical coordinate
system. For that reason, and for the sake of giving the student more experience with the
other common notation, the Gibbs notation will often be used in this chapter.

7.1 Transformations to cylindrical coordinates

The rotation of a fluid about an axis induces an acceleration in that a fluid particle’s velocity
vector is certainly changing with respect to time. Such a motion is most easily described
with a set of cylindrical coordinates. The transformation and inverse transformation to and

241
242 CHAPTER 7. VORTEX DYNAMICS

r
j

i x

Figure 7.1: Representation of a point in Cartesian and cylindrical coordinates along with
unit vectors for both systems.

from cylindrical (r, θ, ẑ) coordinates to Cartesian (x, y, z) is given by the familiar
p
x = r cos θ, r=x2 + y 2, (7.3)
y
y = r sin θ, θ = arctan , (7.4)
x
z = ẑ, ẑ = z. (7.5)

Most of the distinctions between the two systems can be understood by considering two-
dimensional geometries. The representation of an arbitrary point in both two-dimensional
(x, y) Cartesian and two-dimensional (r, θ) cylindrical coordinate systems along with the
unit basis vectors for both systems, i, j, and er , eθ , is sketched in Fig. 7.1. Often a
pure two-dimensional representation is called “polar;” while “cylindrical” is reserved for
the three-dimensional extension. We will typically use “cylindrical” for either two- or three-
dimensional systems.

7.1.1 Centripetal and Coriolis accelerations

The fact that a point in motion is accompanied by changes in the basis vectors with respect
to time in the cylindrical representation, but not for Cartesian basis vectors, accounts for
the most striking differences in the formulations of the governing equations, namely the
appearance of so-called

• centripetal acceleration, and

• Coriolis acceleration,

in the cylindrical representation. These were briefly mentioned earlier in Ch. [Link]. Con-
sider the representations of a two-dimensional velocity vector v in both Cartesian and polar

© 05 December 2025, Joseph M. Powers. All rights reserved.

7.1. TRANSFORMATIONS TO CYLINDRICAL COORDINATES 243

u v

Figure 7.2: Geometrical representation of a velocity vector v in terms both Cartesian and
polar basis vectors.

systems:

v = ui + vj, or (7.6)
v = vr er + vθ eθ . (7.7)

We sketch a velocity vector with a representation in both Cartesian and polar coordinate
systems in Fig. 7.2.
Now the unsteady (as opposed to the advective) part of the acceleration vector of a
particle is simply the partial derivative of the velocity vector with respect to time. Now
formally, we must allow for variations of the unit basis vectors as well as the components
themselves so that
∂v ∂u ∂i ∂v ∂j
= i+u + j+v , (7.8)
∂t ∂t ∂t
|{z} ∂t ∂t
|{z}
=0 =0
∂v ∂vr ∂er ∂vθ ∂eθ
= er + vr + eθ + vθ . (7.9)
∂t ∂t ∂t ∂t ∂t
Now the time derivatives of the Cartesian basis vectors are zero, as they are defined not to
change with the position of the particle. Hence for a Cartesian representation, we have for
the unsteady component of acceleration the familiar:
∂v ∂u ∂v
= i + j. (7.10)
∂t ∂t ∂t
However the time derivative of the cylindrical basis vectors does change with time for
particles in motion! To see this, let us first relate er and eθ to i and j. From the sketch of
Fig. 7.3, it is clear that

© 05 December 2025, Joseph M. Powers. All rights reserved.

244 CHAPTER 7. VORTEX DYNAMICS

Figure 7.3: Geometrical representation of polar unit vectors in terms of Cartesian unit
vectors.

er = cos θi + sin θj, (7.11)

eθ = − sin θi + cos θj. (7.12)

This is a linear system of equations:

    
er cos θ sin θ i
= . (7.13)
eθ − sin θ cos θ j

The coefficient matrix has determinant of unity along with orthonormal columns and is a
rotation matrix. Its inverse is its transpose. We can also use Cramer’s rule to invert to find

i = cos θer − sin θeθ , (7.14)

j = sin θer + cos θeθ . (7.15)

Now, examining time derivatives of the unit vectors, we see that

∂er ∂θ ∂θ
= − sin θ i + cos θ j, (7.16)
∂t ∂t ∂t
∂θ
= eθ , (7.17)
∂t
and
∂eθ ∂θ ∂θ
= − cos θ i − sin θ j, (7.18)
∂t ∂t ∂t
∂θ
= − er . (7.19)
∂t
so there is a formal variation of the unit vectors with respect to time as long as the angular
velocity ∂θ/∂t 6= 0. So the acceleration vector is

∂v ∂vr ∂θ ∂vθ ∂θ
= er + vr eθ + eθ − vθ er , (7.20)
∂t ∂t ∂t  ∂t
 ∂t 
∂vr ∂θ ∂vθ ∂θ
= − vθ er + + vr eθ . (7.21)
∂t ∂t ∂t ∂t

Now from basic geometry, as sketched in Fig. 7.4, we have

© 05 December 2025, Joseph M. Powers. All rights reserved.

7.1. TRANSFORMATIONS TO CYLINDRICAL COORDINATES 245

Figure 7.4: Sketch of relation of differential distance ds to velocity in angular direction vθ .

ds = r dθ, (7.22)
vθ dt = r dθ, (7.23)
vθ ∂θ
= . (7.24)
r ∂t
Consequently, we can write the unsteady component of acceleration as
   
∂v  ∂vr vθ 
2
 er +  ∂vθ vr vθ 
=  −   +  eθ . (7.25)
∂t ∂t r
|{z} ∂t r
|{z}
centripetal Coriolis

Two, apparently new, accelerations have appeared as a consequence of the transformation:

centripetal acceleration, vθ2 /r, directed towards the center, and Coriolis acceleration, vr vθ /r,
directed in the direction of increasing θ. These terms do not have explicit dependency on
time derivatives of velocity. And yet when the equations are constructed in this coordinate
system, they represent real accelerations, and are consequences of forces. As can be seen
by considering the general theory of non-orthogonal coordinate transformations, terms like
the centripetal and Coriolis acceleration are associated with the Christoffel symbols of the
transformation; see Powers and Sen (2015), Ch. 1.6.
Such terms perhaps contributed to the development of Einstein’s theory of general rel-
ativity as well. Refusing to accept that our typical expression of a body force, mg, was
fundamental, Einstein instead postulated that it was a term that was a relic of a coordinate
transformation. He held that we in fact exist in a more complex geometry than classically
considered. He constructed his theory of general relativity such that no gravitational force
exists, but when coordinate transformations are employed to give us a classical view of the
non-relativistic universe, the term mg appears in much the same way as centripetal and
Coriolis accelerations appear when we transform to cylindrical coordinates.

7.1.2 Grad and div for cylindrical systems

We can use the chain rule to develop expressions for grad and div in cylindrical coordinate
systems. Consider the Cartesian

∂ ∂ ∂
∇= i+ j + k. (7.26)
∂x ∂y ∂z

© 05 December 2025, Joseph M. Powers. All rights reserved.

246 CHAPTER 7. VORTEX DYNAMICS

The chain rule gives us

∂ ∂r ∂ ∂θ ∂ ∂ ẑ ∂
= + + , (7.27)
∂x ∂x ∂r ∂x ∂θ ∂x ∂ ẑ
∂ ∂r ∂ ∂θ ∂ ∂ ẑ ∂
= + + , (7.28)
∂y ∂y ∂r ∂y ∂θ ∂y ∂ ẑ
∂ ∂r ∂ ∂θ ∂ ∂ ẑ ∂
= + + . (7.29)
∂z ∂z ∂r ∂z ∂θ ∂z ∂ ẑ
Now, we have

∂r 2x x
= p = = cos θ, (7.30)
∂x 2 x2 + y 2 r
∂r 2y y
= p = = sin θ, (7.31)
∂y 2 x2 + y 2 r
∂r
= 0, (7.32)
∂z
and
∂θ y r sin θ sin θ
= − 2 = − = − , (7.33)
∂x x + y2 r2 r
∂θ x r cos θ cos θ
= 2 2
= 2
= , (7.34)
∂y x +y r r
∂θ
= 0, (7.35)
∂z
and
∂ ẑ
= 0, (7.36)
∂x
∂ ẑ
= 0, (7.37)
∂y
∂ ẑ
= 1, (7.38)
∂z
so
∂ ∂ sin θ ∂
= cos θ − , (7.39)
∂x ∂r r ∂θ
∂ ∂ cos θ ∂
= sin θ + , (7.40)
∂y ∂r r ∂θ
∂ ∂
= . (7.41)
∂z ∂ ẑ

© 05 December 2025, Joseph M. Powers. All rights reserved.

7.1. TRANSFORMATIONS TO CYLINDRICAL COORDINATES 247

[Link] Grad

So now we are prepared to write an explicit form for ∇ in cylindrical coordinates:

 
∂ sin θ ∂
∇ = cos θ − (cos θer − sin θeθ )
∂r r ∂θ | {z }
| {z } i
∂
∂x
 
∂ cos θ ∂ ∂
+ sin θ + (sin θer + cos θeθ ) + eẑ , (7.42)
∂r r ∂θ | {z } ∂ ẑ
| {z } j
∂
∂y
   
2

∂ 2 sin θ cos θ sin θ cos θ ∂
∇ = cos θ + sin θ + − + er
∂r r r ∂θ
  2  
∂ sin θ cos2 θ ∂
+ (− sin θ cos θ + sin θ cos θ) + + eθ
∂r r r ∂θ
∂
+ eẑ , (7.43)
∂ ẑ
∂ 1 ∂ ∂
∇ = er + eθ + eẑ . (7.44)
∂r r ∂θ ∂ ẑ

We can now write a simple expression for the advective component, vT ·∇, of the acceleration
vector:
∂ vθ ∂ ∂
vT · ∇ = vr + + vẑ . (7.45)
∂r r ∂θ ∂ ẑ

[Link] Div

The divergence is straightforward. In Cartesian coordinates we have

∂u ∂v ∂w
∇T · v = + + . (7.46)
∂x ∂y ∂z

In cylindrical, we replace derivatives with respect to x, y, z with those with respect to r, θ, ẑ,
so
∂u sin θ ∂u ∂v cos θ ∂v ∂w
∇T · v = cos θ − + sin θ + + . (7.47)
∂r r ∂θ ∂r r ∂θ ∂ ẑ
Now u, v and w transform in the same way as x, y, and z, so

u = vr cos θ − vθ sin θ, (7.48)

v = vr sin θ + vθ cos θ, (7.49)
w = vẑ . (7.50)

© 05 December 2025, Joseph M. Powers. All rights reserved.

248 CHAPTER 7. VORTEX DYNAMICS

Substituting and taking partials, we find that

   
 ∂vr ∂vθ  sin θ  ∂vr ∂vθ 
∇T · v = cos θ cos θ − sin θ − cos θ − sin θvr − sin θ − cos θvθ 
∂r | {z }
| {z∂r} r | {z∂θ} ∂θ
C
A
   B 
 ∂vr ∂vθ  cos θ  ∂vr ∂vθ 
+ sin θ sin θ + cos θ + sin θ + cos θvr + cos θ − sin θvθ 
∂r | {z }
| {z∂r} r | {z∂θ} ∂θ
C
A B
∂vẑ
+ . (7.51)
∂ ẑ
When expanded, the terms labeled A, B, and C cancel in this expression. Then using the
trigonometric identity sin2 θ + cos2 θ = 1, we arrive at the simple form
∂vr vr 1 ∂vθ ∂vẑ
∇T · v = + + + , (7.52)
∂r r r ∂θ ∂ ẑ
that is often rewritten as
1 ∂ 1 ∂vθ ∂vẑ
∇T · v = (rvr ) + + . (7.53)
r ∂r r ∂θ ∂ ẑ
Using the same procedure, we can show that the Laplacian operator transforms to
 
2 1 ∂ ∂ 1 ∂2 ∂2
∇ = r + 2 2 + 2. (7.54)
r ∂r ∂r r ∂θ ∂ ẑ

[Link] Alternate derivations

Presented here are brief details of an alternate, more formal mathematical derivation of
the gradient and Laplacian operators in cylindrical coordinates. General background was
presented in Ch. 2.5 and is given in more detail by Powers and Sen (2015), Ch. 1.6. As
before, one can transform from the Cartesian system with (x, y, z) as coordinates to the
cylindrical system with (r, θ, ẑ) as coordinates via x = r cos θ, y = r sin θ, z = ẑ. We will
consider the domain r ∈ [0, ∞), θ ∈ [0, 2π], ẑ ∈ (−∞, ∞). Then, with the exception of the
origin (x, y, z) = (0, 0, 0), every (x, y, z) will map to a unique (r, θ, ẑ).
The Jacobian matrix of the transformation is
 ∂x ∂x ∂x   
∂r ∂θ ∂ ẑ cos θ −r sin θ 0
J =  ∂y ∂r
∂y
∂θ
∂y 
∂ ẑ
=  sin θ r cos θ 0  . (7.55)
∂z ∂z ∂z 0 0 1
∂r ∂θ ∂ ẑ

We have J = |J| = r; this can be shown to tell us that the transformation is singular and thus
non-unique when r = 0. It is orientation-preserving for r > 0, and it is volume-preserving
only for r = 1; thus, in general it does not preserve volume.

© 05 December 2025, Joseph M. Powers. All rights reserved.

7.1. TRANSFORMATIONS TO CYLINDRICAL COORDINATES 249

If we take dx = (dx, dy, dz)T , and dr = (dr, dθ, dẑ)T , we have

dx = J · dr, (7.56)
   ∂x ∂x ∂x
 
dx ∂r ∂θ ∂ ẑ dr
 dy  =  ∂y ∂y ∂y   dθ  . (7.57)
∂r ∂θ ∂ ẑ
dz ∂z ∂z ∂z dẑ
∂r ∂θ ∂ ẑ

Now we want distance to be invariant in either coordinate system. We have the standard
result for Cartesian systems that

ds2 = dxT · dx = dx2 + dy 2 + dz 2 . (7.58)

For invariance of distance in the transformed system, we thus require

ds2 = dxT · dx = (J · dr)T · (J · dr) , (7.59)

= drT · J T
| {z· J} ·dr, (7.60)
G
T
= dr · G · dr. (7.61)

Recall the metric tensor G is defined by Eq. (2.289) as

G = JT · J, (7.62)
  
cos θ sin θ 0 cos θ −r sin θ 0
=  −r sin θ r cos θ 0   sin θ r cos θ 0  , (7.63)
0 0 1 0 0 1
 
1 0 0
=  0 r2 0  . (7.64)
0 0 1

Because G is diagonal, the implication can be shown to be that new coordinates axes are
also orthogonal. So for our system
  
1 0 0 dr
ds2 = ( dr dθ dẑ )  0 r 2 0   dθ  , (7.65)
0 0 1 dẑ
= dr 2 + (r dθ)2 + dẑ 2 . (7.66)

Now the gradient operator in the Cartesian system is related to that of the cylindrical
system via the same analysis used to obtain Eq. (2.286):
 ∂   ∂ 
∂x ∂r
∂
∇ =  ∂y  = (JT )−1  ∂θ
∂ 
, (7.67)
∂ ∂
∂z ∂ ẑ

© 05 December 2025, Joseph M. Powers. All rights reserved.

250 CHAPTER 7. VORTEX DYNAMICS
  ∂ 
cos θ − sinr θ 0 ∂r
=  sin θ cosr θ 0   ∂θ ∂ 
, (7.68)
∂
0 0 1 ∂ ẑ
 ∂ sin θ ∂

cos θ ∂r − r ∂θ
=  sin θ ∂r
∂ ∂ 
+ cosr θ ∂θ . (7.69)
∂
∂ ẑ

Then we find
 
vr cos θ − vθ sin θ
∇T · v = ( cos θ ∂r
∂
− sin θ ∂
r ∂θ
∂
sin θ ∂r + cos θ ∂
r ∂θ
∂
∂ ẑ
)  vr sin θ + vθ cos θ  , (7.70)
vẑ
1 ∂ 1 ∂vθ ∂vẑ
= (rvr ) + + . (7.71)
r ∂r r ∂θ ∂ ẑ
Consider next the Laplacian operator, ∇2 = ∇T · ∇, that is

∇2 = ∇T · ∇, (7.72)
 ∂

cos θ ∂r − sinr θ ∂θ
∂

= ∂
( cos θ ∂r − sin θ ∂
r ∂θ
∂
sin θ ∂r + cos θ ∂
r ∂θ
∂
∂ ẑ
)  sin θ ∂r + r ∂θ  .
∂ cos θ ∂
(7.73)
∂
∂ ẑ

Detailed expansion followed by extensive use of trigonometric identities reveals that this
reduces to
 
T 2 1 ∂ ∂ 1 ∂2 ∂2 ∂2 1 ∂ 1 ∂2 ∂2
∇ ·∇=∇ = r + 2 2+ 2 = 2+ + 2 2 + 2. (7.74)
r ∂r ∂r r ∂θ ∂ ẑ ∂r r ∂r r ∂θ ∂ ẑ

7.1.3 Incompressible Navier-Stokes equations in cylindrical coor-

dinates
Leaving out some additional details of the transformations, we find that the incompressible
Navier-Stokes equations for a Newtonian fluid with constant viscosity and body force confined
to the −ẑ direction are
1 ∂ 1 ∂vθ ∂vẑ
0 = (rvr ) + + , (7.75)
  r ∂r r ∂θ ∂ ẑ 
∂vr vθ2 ∂vr vθ ∂vr ∂vr 1 ∂p vr 2 ∂vθ
− + vr + + vẑ = − + ν ∇2 vr − 2 − 2 ,(7.76)
∂t r ∂r r ∂θ ∂ ẑ ρ ∂r r r ∂θ
   
∂vθ vr vθ ∂vθ vθ ∂vθ ∂vθ 1 1 ∂p 2 2 ∂vr vθ
+ + vr + + vẑ = − + ν ∇ vθ + 2 − 2 ,
∂t r ∂r r ∂θ ∂ ẑ ρ r ∂θ r ∂θ r
(7.77)
∂vẑ ∂vẑ vθ ∂vẑ ∂vẑ 1 ∂p
+ vr + + vẑ = − + ν∇2 vẑ − gẑ . (7.78)
∂t ∂r r ∂θ ∂ ẑ ρ ∂ ẑ

© 05 December 2025, Joseph M. Powers. All rights reserved.

7.2. IDEAL ROTATIONAL VORTEX 251

Figure 7.5: Sketch of a fluid rotating as a pure solid body.

Within the acceleration terms, strictly unsteady terms, advective terms, as well as centripetal
and Coriolis terms appear. The viscous terms have additional complications that we have
not considered in detail but arise because we must transform ∇2 v, and there are many
non-intuitive terms that arise here when expanded in full.

7.2 Ideal rotational vortex

Let us consider the kinematics and dynamics of an ideal rotational vortex, that we define to
be a fluid rotating as a solid body. Let us assume incompressible flow, so ∇T · v = 0, assume
a simple velocity field, and ask what forces could have given rise to that velocity field. We
will simply use z for the azimuthal coordinate instead of ẑ here. Take
ωo r
vr = 0, vθ = , vz = 0. (7.79)
2
This velocity field was also considered in a Cartesian representation in Ch. 3.11.5. The
kinematics of this flow are simple and sketched in Fig. 7.5. Here ωo is a constant. The
velocity is zero at the origin and grows in amplitude with linear distance from the origin.
The flow is steady, and the streamlines are circles centered about the origin. Obviously, as
r → ∞, the theory of relativity would suggest that such a flow would break down as the
velocity approached the speed of light. In fact, one would find as well that as the velocities
approached the sound speed that compressibility effects would become important far before
relativistic effects.
Whatever the case, does this assumed velocity field satisfy incompressible mass conser-

© 05 December 2025, Joseph M. Powers. All rights reserved.

252 CHAPTER 7. VORTEX DYNAMICS

vation?
1 ∂  ωo r  ∂
?
1 ∂ z}|{
(r(0)) + + (0) = 0. (7.80)
r ∂r r |∂θ {z2 } ∂z
=0

Obviously it does.
Next let us consider the acceleration of an element of fluid and the forces that could
give rise to that acceleration. First consider that portion of the acceleration that is neither
centripetal nor Coriolis for this flow:

∂ ∂ vθ ∂ ∂
+ vr + + vz = 0. (7.81)
∂t |{z} ∂r r |{z}
∂θ |{z} ∂z
|{z} =0 =0
=0 =0

As the only non-zero component of velocity, vθ , has no dependency on θ, the unsteady and
advective portions of the acceleration are zero for this flow. And because vr = 0, there is no
Coriolis acceleration. So the only acceleration is centripetal and is −vθ2 /r = −ωo2 r/4.
Consider now the viscous terms for this flow. We recall for an incompressible Newtonian
fluid that

τij = τji = 2µ∂(i vj) + λ ∂k vk δij , (7.82)

|{z}
=0
= µ (∂i vj + ∂j vi ) , (7.83)
∂j τji = µ (∂j ∂i vj + ∂j ∂j vi ) , (7.84)
 

= µ ∂i ∂j vj +∂j ∂j vi  , (7.85)

|{z}
=0
2
= µ∇ v. (7.86)

We also note that

∇×ω = ǫijk ∂j ωk = ǫijk ∂j ǫkmn ∂m vn , (7.87)

= ǫkij ǫkmn ∂j ∂m vn , (7.88)
= (δim δjn − δin δjm ) ∂j ∂m vn , (7.89)
= ∂j ∂i vj − ∂j ∂j vi , (7.90)
= ∂i ∂j vj −∂j ∂j vi , (7.91)
|{z}
=0
= −∂j ∂j vi . (7.92)

Comparing, we see that for this incompressible flow,

T
∇T · τ = −µ(∇ × ω). (7.93)

© 05 December 2025, Joseph M. Powers. All rights reserved.

7.2. IDEAL ROTATIONAL VORTEX 253

Now, using relations that can be developed for the curl in cylindrical coordinates, we have
for this flow that
1 ∂vz ∂vθ
ωr = − = 0, (7.94)
r ∂θ ∂z
∂vr ∂vz
ωθ = − = 0, (7.95)
∂z ∂r
1 ∂ 1 ∂vr
ωz = (rvθ ) − , (7.96)
r ∂r r ∂θ
1 ∂  ωo r 
= r , (7.97)
r ∂r 2
= ωo . (7.98)
So the flow has a constant rotation rate, ωo . Because it is constant, its curl is zero, and

T
we have for this flow that ∇T · τ = 0. We could just as well show for this flow that
τ = 0. That is because the kinematics are those of pure rotation as a solid body with no
deformation. No deformation implies no viscous stress.
Hence, the three linear momenta equations in the cylindrical coordinate system reduce
to the following:
vθ2 1 ∂p
− = − , (7.99)
r ρ ∂r
1 1 ∂p
0 = − , (7.100)
ρ r ∂θ
1 ∂p
0 = − − gz . (7.101)
ρ ∂z
The r momentum equation strikes a balance between centripetal inertia and radial pressure
gradients. The θ momentum equation shows that as there is no acceleration in this direction,
there can be no net pressure force to induce it. The z momentum equation enforces a balance
between pressure forces and gravitational body forces.
If we take p = p(r, θ, z) and p(ro , θ, zo ) = po , then
∂p ∂p ∂p
dp = dr + dθ + dz, (7.102)
∂r ∂θ ∂z
ρvθ2
= dr + 0 dθ − ρgz dz, (7.103)
r
ρωo2 r 2
= dr − ρgz dz, (7.104)
4r
ρωo2 r
= dr − ρgz dz, (7.105)
4
ρωo2 2
p − po = (r − ro2 ) − ρgz (z − zo ), (7.106)
8
ρω 2
p(r, z) = po + o (r 2 − ro2 ) − ρgz (z − zo ). (7.107)
8

© 05 December 2025, Joseph M. Powers. All rights reserved.

254 CHAPTER 7. VORTEX DYNAMICS

Now on a surface of constant pressure we have p(r, z) = p̂. So

ρωo2 2
p̂ = po + (r − ro2 ) − ρgz (z − zo ), (7.108)
8
ρω 2
ρgz (z − zo ) = po − p̂ + o (r 2 − ro2 ), (7.109)
8
po − p̂ ωo2 2
z = zo + + (r − ro2 ). (7.110)
ρgz 8gz
So a surface of constant pressure is a parabola in r with a minimum at r = 0. This is
consistent with what one observes upon spinning a bucket of water.
Now let us rearrange our general equation for the pressure field and eliminate ω using
vθ = ωo r/2 and defining vθo = ωo ro /2:
1 1 2
p − ρvθ2 + ρgz z = po − ρvθo + ρgz zo = C. (7.111)
2 2
This looks similar to the steady irrotational incompressible Bernoulli equation in which
p + 12 ρv 2 + ρgz z = K. But there is a difference in the sign on one of the terms. Now add ρvθ2
to both sides of the equation to get
1
p + ρvθ2 + ρgz z = C + ρvθ2 . (7.112)
2
Now because vθ = ωo r/2, vr = 0, we have lines of constant r as streamlines, and vθ is
constant on those streamlines, so that we get
1
p + ρvθ2 + ρgz z = C ′ , on a streamline. (7.113)
2
Here C ′ varies from streamline to streamline.
We lastly note that the circulation for this system depends on position. If we choose our
contour integral to be a circle of radius a about the origin we find
I
Γ = vT · dr, (7.114)
IC

= vθ eTθ · (a dθ eθ ), (7.115)
C
Z 2π  
1
= ωo a (a dθ), (7.116)
0 2
= πa2 ωo . (7.117)

7.3 Ideal irrotational vortex

Now let us perform a similar analysis for the following velocity field:
Γo
vr = 0, vθ = , vz = 0. (7.118)
2πr

© 05 December 2025, Joseph M. Powers. All rights reserved.

7.3. IDEAL IRROTATIONAL VORTEX 255

Figure 7.6: Sketch of an ideal irrotational point vortex and a circular contour of r = a.

We have considered the same velocity field as represented in Cartesian coordinates, in

Ch. 3.11.10. The kinematics of this flow are also simple and sketched in Fig. 7.6. We
see once again that the streamlines are circles about the origin. But here, as opposed to the
ideal rotational vortex, vθ → 0 as r → ∞ and vθ → ∞ as r → 0. The vorticity vector of this
flow is
1 ∂vz ∂vθ
ωr = − = 0, (7.119)
r ∂θ ∂z
∂vr ∂vz
ωθ = − = 0, (7.120)
∂z ∂r
1 ∂ 1 ∂vr
ωz = (rvθ ) − , (7.121)
r ∂r  r ∂θ
1 ∂ Γo
= r , (7.122)
r ∂r 2πr
 
1 ∂ Γo
= = 0. (7.123)
r ∂r 2π
This flow field, that seems to be the epitome of a rotating flow, is formally irrotational as
it has zero vorticity! What is happening is that a fluid element not at the origin is actually
undergoing severe deformation as it rotates about the origin; however, it does not rotate
about its own center of mass. Therefore, the vorticity vector is zero, except at the origin,
where it is undefined.
The circulation for this flow about a circle of radius a is
I
Γ = vT · dr, (7.124)
C

© 05 December 2025, Joseph M. Powers. All rights reserved.

256 CHAPTER 7. VORTEX DYNAMICS
I
= vθ eTθ · (a dθ eθ ), (7.125)
C
Z 2π
= vθ (a dθ), (7.126)
0
Z 2π
Γo
= a dθ, (7.127)
0 2πa
= Γo . (7.128)
So the circulation is independent of the radius of the closed contour. In fact it can be shown
that as long as the closed contour includes the origin in its interior that any closed contour
will have this same circulation. We call Γo the ideal irrotational vortex strength, in that it
is proportional to the magnitude of the velocity at any radius.
Let us once again consider the forces that could induce the motion of this vortex if the
flow happens to be incompressible with constant properties and in a potential field where the

T
gravitational body force per unit mass is −gz k. Recall again that ∇T · τ = −µ(∇ × ω),

T
and that because ω = 0 that ∇T · τ = 0 for this flow. Because there is deformation here,
τ itself is not zero, but its divergence is zero. For example, if we consider one component of
viscous stress τrθ and use standard relations that can be derived for incompressible Newtonian
fluids, we find that
   
∂  vθ  1 ∂vr ∂ Γo µΓo
τrθ = µ r + = µr = − . (7.129)
∂r r r ∂θ ∂r 2πr 2 πr 2
The equations of motion reduce to the same ones as for the ideal rotational vortex:
vθ2 1 ∂p
− = − , (7.130)
r ρ ∂r
1 1 ∂p
0 = − , (7.131)
ρ r ∂θ
1 ∂p
0 = − − gz . (7.132)
ρ ∂z
Once more we can deduce a pressure field that is consistent with these and the same set of
conditions at r = ro , z = zo , with p = po :
∂p ∂p ∂p
dp = dr + dθ + dz, (7.133)
∂r ∂θ
|{z} ∂z
=0
ρvθ2
= dr − ρgz dz, (7.134)
r
ρΓ2o dr
= − ρgz dz, (7.135)
4π 2 r3

ρΓ2o 1 1
p − po = − 2 − − ρgz (z − zo ), (7.136)
8π r 2 ro2

© 05 December 2025, Joseph M. Powers. All rights reserved.

7.4. HELMHOLTZ VORTICITY TRANSPORT EQUATION 257

ρΓ2o 1 ρΓ2o 1
p+ + ρgz z = po + + ρgz zo , (7.137)
8π 2 r 2 8π 2 ro2
1 1 2
p + ρvθ2 + ρgz z = po + ρv + ρgz zo = C. (7.138)
2 2 θo
This is once again Bernoulli’s equation. Here it is for an irrotational flow field that is also
time-independent, so the Bernoulli constant C is truly constant for the entire flow field and
not just along a streamline.
On isobars we have p = p̂ that gives us
 
ρΓ2o 1 1
p̂ − po = − 2 − − ρgz (z − zo ), (7.139)
8π r 2 ro2
 
po − p̂ Γ2o 1 1
z = zo + + 2 − . (7.140)
ρgz 8π gz r 2 ro2
The pressure goes to negative infinity at the origin. One can show that actual forces, obtained
by integrating pressure over area, are in fact bounded.

7.4 Helmholtz vorticity transport equation

Here we will take the curl of the linear momenta principle to obtain a relationship, the
Helmholtz vorticity transport equation, that shows how the vorticity field evolves in a general
fluid.

7.4.1 General development

First, we recall some useful vector identities:
 T 
T v ·v
(v · ∇)v = ∇ + ω × v, (7.141)
2
∇ × (a × b) = (bT · ∇)a − (aT · ∇)b + a(∇T · b) − b(∇T · a), (7.142)
∇ × (∇φ) = 0, (7.143)
∇ · (∇ × v) = ∇T · ω = 0.
T
(7.144)

The first is equivalent to Eq. (2.263); the others are easily proved.
We start now with the linear momenta principle for a general fluid; we recast Eq. (4.260)
and write
∂v 1 1
T
+ (vT · ∇)v = f − ∇p + ∇T · τ . (7.145)
∂t ρ ρ
We expand the term (vT · ∇)v and then apply the curl operator to both sides to get
  T    
∂v v ·v 1 1 T

T
∇× +∇ + ω × v = ∇ × f − ∇p + ∇ ·τ . (7.146)
∂t 2 ρ ρ

© 05 December 2025, Joseph M. Powers. All rights reserved.

258 CHAPTER 7. VORTEX DYNAMICS

This becomes, via the linearity of the various operators,

  T     
∂ v ·v 1 1 T

T
(∇ × v) + ∇ × ∇ +∇ × ω × v = ∇ × f − ∇ × ∇p + ∇ × ∇ ·τ .
∂t | {z } 2 ρ ρ
ω | {z }
=0
(7.147)
Using our vector identity for the term with two cross products, Eq. (7.142), we get
   
∂ω T T T T 1 1 T

T
+ (v · ∇)ω −(ω ·∇)v+ω( ∇ | {z· ω}) = ∇×f−∇× ρ ∇p +∇× ρ ∇ · τ
| {z· v} )−v(∇ .
|∂t {z } =0
ω = − ρ1 dρ
= ddt dt

(7.148)
Rearranging, we have
   
dω ω dρ T 1 1 T

T
− = (ω · ∇)v + ∇ × f − ∇ × ∇p + ∇ × ∇ ·τ , (7.149)
dt ρ dt ρ ρ
 T     
1 dω ω dρ ω 1 1 1 1 1 T

T
− 2 = ·∇ v+ ∇×f − ∇× ∇p + ∇ × ∇ ·τ , (7.150)
ρ dt ρ dt ρ ρ ρ ρ ρ ρ
   T     
d ω ω 1 1 1 1 1
T
= ·∇ v+ ∇×f − ∇× ∇p + ∇ × ∇T · τ , (7.151)
dt ρ ρ ρ ρ ρ ρ ρ
     
d ω
1 1
T
ρ = ωT · ∇ v + ∇ × f − ∇ × ∇p + ∇ × ∇T · τ . (7.152)
dt ρ ρ ρ

Now consider the term −∇ × ((1/ρ)∇p). In Einstein notation, we have

   
1 1 1
−ǫijk ∂j ∂k p = −ǫijk ∂j ∂k p − 2 (∂j ρ)(∂k p) , (7.153)
ρ ρ ρ
1 1
= − ǫijk ∂j ∂k p + 2 ǫijk (∂j ρ)(∂k p), (7.154)
ρ | {z } ρ
=0
1
= ∇ρ × ∇p. (7.155)
ρ2
We write the final general form of the vorticity transport equation as
   
d ω T

1 1 T

T
ρ = ω · ∇ v + |∇{z × f} + 2 ∇ρ × ∇p + ∇ × ∇ ·τ . (7.156)
dt ρ | {z } ρ ρ
B | {z } | {z }
A
C D
Here we see the evolution of the vorticity scaled by the density is affected by four physical
processes, that we describe in greater detail directly, namely
• A: bending and stretching of vortex tubes,

• B: non-conservative body forces (if f = −∇ϕ, then f is conservative, and ∇ × f =

−∇ × ∇ϕ = 0. For example f = −gz k gives ϕ = gz z),

© 05 December 2025, Joseph M. Powers. All rights reserved.

7.4. HELMHOLTZ VORTICITY TRANSPORT EQUATION 259

• C: non-barotropic, also known as baroclinic, effects (if a fluid is barotropic, then p =

p(ρ) and ∇p = (dp/dρ)∇ρ thus ∇ρ × ∇p = ∇ρ × (dp/dρ) ∇ρ = 0), and
• D: viscous effects.

7.4.2 Incompressible conservative body force limit

The Helmholtz vorticity transport equation (7.156) reduces significantly in special limiting
cases involving incompressible flow in the limit of a conservative body force. In this limit
Eq. (7.156) reduces to the following
dω 1
= (ω T · ∇)v + ∇ × (∇T · τ )T . (7.157)
dt ρ

[Link] Isotropic, Newtonian, constant viscosity

Now if we further require that the fluid be isotropic and Newtonian with constant viscosity,
the viscous term can be written as
∇ × (∇T · τ )T = ǫijk ∂j ∂m (2µ(∂(m vk) − (1/3) ∂l vl δmk )), (7.158)
|{z}
=0
= µǫijk ∂j ∂m (∂m vk + ∂k vm ), (7.159)
= µǫijk ∂j (∂m ∂m vk + ∂m ∂k vm ), (7.160)
= µǫijk ∂j (∂m ∂m vk + ∂k ∂m vm ), (7.161)
| {z }
=0
= µ∂m ∂m ǫijk ∂j vk , (7.162)
| {z }
ω
2
= µ∇ ω. (7.163)
So we get, recalling that kinematic viscosity ν = µ/ρ,
dω
= (ω T · ∇)v + ν∇2 ω. (7.164)
dt

[Link] Two-dimensional, isotropic, Newtonian, constant viscosity

If we further require two-dimensionality, then we have ω = (0, 0, ω3(x1 , x2 ))T , and ∇ =
(∂1 , ∂2 , 0)T , so ω T · ∇ = 0. Thus, we get the simple
 2 
dω3 2 ∂ ω3 ∂ 2 ω3
= ν∇ ω3 = ν + . (7.165)
dt ∂x21 ∂x22
If the flow is further inviscid ν = 0, we get
dω3
= 0, (7.166)
dt
and we find that there is no tendency for vorticity to change along a streamline. If we further
have an initially irrotational state, then we get ω = 0 for all space and time.

© 05 December 2025, Joseph M. Powers. All rights reserved.

260 CHAPTER 7. VORTEX DYNAMICS

Figure 7.7: Isobars and isochores, center of mass G, and center of pressure for a barotropic
fluid.

7.4.3 Physical interpretations

Let us consider how two of the terms in Eq. (7.156) contribute to the generation of vorticity.

[Link] Baroclinic (non-barotropic) effects

If a fluid is barotropic then we can write p = p(ρ), or ρ = ρ(p). As an example, an isentropic

CPIG has p/po = (ρ/ρo )γ , where γ is the ratio of specific heats, and the o subscript indicates
a constant value. Such a gas is barotropic. For such a fluid, we must have by the chain
rule that ∂i p = (dp/dρ)∂i ρ. Hence ∇p and ∇ρ are vectors that point in the same direction.
Moreover, isobars (lines of constant pressure) must be parallel to isochores (lines of constant
density). If, as sketched in Fig. 7.7, we calculate the resultant vector from the net pressure
force, as well as the center of mass for a finite fluid volume, we would see that the resultant
force had no lever arm with the center of gravity. Hence it would generate no torque, and no
tendency for the fluid element to rotate about its center of mass, hence no vorticity would
be generated by this force.
For a baroclinic fluid, we do not have p = p(ρ); hence, we must expect that ∇p points
in a different direction than ∇ρ. If we examine this scenario, as sketched in Fig. 7.8, we
discover that the resultant force from the pressure has a non-zero lever arm with the center
of mass of the fluid element. Hence, it generates a torque, a tendency to rotate the fluid
about G, and vorticity.

© 05 December 2025, Joseph M. Powers. All rights reserved.

7.4. HELMHOLTZ VORTICITY TRANSPORT EQUATION 261

Figure 7.8: Isobars and isochores, center of mass G, and center of pressure for a baroclinic
fluid.

b
w curve everywhere parallel
to vorticity vector

n locally orthogonal
coordinate system
w, n, b

Figure 7.9: Local orthogonal intrinsic coordinate system oriented with local vorticity field.

[Link] Bending and stretching of vortex tubes

Now let us consider generation of vorticity by three-dimensional effects. Such effects are
commonly characterized as the bending and stretching of what is known as vortex tubes.
Here we focus on just the following inviscid equation:
dω
= (ω T · ∇)v. (7.167)
dt
If we consider a coordinate system that is oriented with the vorticity field as sketched in
Fig. 7.9, we will get many simplifications. We take the following directions
• w: the direction parallel to the vorticity vector,

• n: the principal normal direction, pointing towards the center of curvature,

• b: the biorthogonal direction, orthogonal to w and n.

© 05 December 2025, Joseph M. Powers. All rights reserved.

262 CHAPTER 7. VORTEX DYNAMICS

unstretched
vortex tube

stretched
vortex
tube

Figure 7.10: Increase in vorticity due to stretching of a vortex tube.

With this system, we can say that

  ∂

∂w
(ω T · ∇)v = ( ωw 0 0) ∂
∂n
 v, (7.168)
∂
∂b
∂v
= ωw . (7.169)
∂w

So for the inviscid flow we have

dω ∂v
= ωw . (7.170)
dt ∂w
We have in terms of components

dωw ∂vw
= ωw , (7.171)
dt ∂w
dωn ∂vn
= ωw , (7.172)
dt ∂w
dωb ∂vb
= ωw . (7.173)
dt ∂w

The term ∂vw /∂w we know from kinematics represents a local stretching or extension. Just
as a rotating figure skater increases his or her angular velocity by concentrating his or her
mass about a vertical axis, so does a rotating fluid. The first of these expressions says that
the component of rotation aligned with the present increases if there is stretching in that
direction. This is sketched in Fig. 7.10.
The second and third terms enforce that if vn or vb are changing in the w direction,
when accompanied by non-zero ωw , that changes in the non-aligned components of ω are
induced. Hence the previously zero components ωn , ωb acquire non-zero values, and the lines
parallel to the vorticity vector bend. Hence, we have the term, bending of vortex tubes.
It is generally accepted that the bending and stretching of vortex tubes is an important
mechanism in the transition from laminar to turbulent flow.

© 05 December 2025, Joseph M. Powers. All rights reserved.

7.5. KELVIN’S CIRCULATION THEOREM 263

7.5 Kelvin’s circulation theorem

Kelvin’s circulation theorem describes how the circulation of a material region in a fluid
changes with time. We first recall from Eq. (7.2) the definition of circulation Γ:
I
Γ= vT · dx, (7.174)
C

where C is a closed contour. We next take the material derivative of Γ to get

I
dΓ d
= vT · dx, (7.175)
dt dt C
I I
dvT d
= · dx + vT · dx, (7.176)
dt dt
IC T IC  
dv T dx
= · dx + v ·d , (7.177)
C dt C dt
I I
dvT
= · dx + vT · dv, (7.178)
dt
IC T IC  
dv 1 T
= · dx + d v ·v , (7.179)
C dt C 2
| {z }
=0
I  T
dv
= · dx. (7.180)
C dt

Here we note that because we have chosen a material region for our closed contour that
dx/dt must be the fluid particle velocity. This then allows us to write the second term as a
perfect differential, that integrates over the closed contour to be zero. We continue now by
using the linear momenta principle to replace the particle acceleration with density-scaled
forces to arrive at I  
dΓ 1 1 T

T T
= f − ∇p + ∇ ·τ · dx. (7.181)
dt C ρ ρ
If now the fluid is inviscid (τ = 0), the body force is conservative (f = −∇ϕ), and the fluid
is barotropic ((1/ρ)∇p = (1/ρo )∇P), we then have
I  T
dΓ 1
= −∇ϕ − ∇P · dx, (7.182)
dt C ρo
I  
T P
= − ∇ ϕ+ · dx, (7.183)
C ρo
I  
P
= − d ϕ+ . (7.184)
C ρo
| {z }
=0

© 05 December 2025, Joseph M. Powers. All rights reserved.

264 CHAPTER 7. VORTEX DYNAMICS

The integral on the right hand side is zero because the contour is closed; hence, the integral
is path-independent. Consequently, we arrive at the common version of Kelvin’s circulation
theorem that holds that for a fluid that is inviscid, barotropic, and subjected to conservative
body forces, the circulation following a material region does not change with time:
dΓ
= 0. (7.185)
dt
This is similar to the Helmholtz equation, that, when we make the additional stipulation
of two-dimensionality and incompressibility, gives dω/dt = 0. This is not surprising as the
vorticity is closely linked to the circulation via Stokes’ theorem, Eq. (2.262), that states
I Z Z
T T
Γ= v · dx = (∇ × v) · n dA = ω T · n dA. (7.186)
C A A

7.6 Potential flow of ideal point vortices

Consider the fluid motion induced by the simultaneous interaction of a family of ideal irro-
tational point vortices in an incompressible flow field. Because the flow is irrotational and
incompressible, we have the following useful results:
• Because ∇×v = 0, we can write the velocity vector as the gradient of a scalar potential
φ:
v = ∇φ, if irrotational. (7.187)
We call φ the velocity potential.
• Because ∇T · v = 0, we have

∇T · ∇φ = ∇2 φ = 0, (7.188)

or expanding, we have
∂2φ ∂2φ ∂2φ
+ 2 + 2 = 0. (7.189)
∂x2 ∂y ∂z
• We notice that the equation for φ is linear; hence the method of superposition is valid
here for the velocity potential. That is, we can add an arbitrary number of velocity
potentials together and get a viable flow field.
• The irrotational unsteady Bernoulli equation, Eq. (6.196), gives us the time- and space-
dependent pressure field. This equation is not linear, so we do not expect pressures
from elementary solutions to add to form total pressures.
Recalling that the incompressible, three-dimensional constant viscosity Helmholtz equa-
tion can be written via Eq. (7.164) as
dω
= (ω T · ∇)v + ν∇2 ω, (7.190)
dt

© 05 December 2025, Joseph M. Powers. All rights reserved.

7.6. POTENTIAL FLOW OF IDEAL POINT VORTICES 265

1 h 2

Figure 7.11: Sketch of the mutual influence of two ideal point vortices on each other.

G
1 2
hG
h

Figure 7.12: Sketch showing the center of rotation G.

we see that a flow that is initially irrotational everywhere in an unbounded fluid will always
be irrotational, as dω/dt = 0. There is no mechanism to change the vorticity from its
uniform initial value of zero. This even holds for a viscous flow. However, in a bounded
medium, the no-slip boundary condition almost always tends to diffuse vorticity into the
flow as we shall see. Further for inviscid, barotropic flow, from Kelvin’s circulation theorem,
Eq. (7.185), the circulation Γ has no tendency to change following a particle; that is, Γ
advects unchanged along particle pathlines.

7.6.1 Two interacting ideal vortices

Let us apply this notion to two ideal counterrotating vortices 1 and 2, with respective
strengths, Γ1 and Γ2 , as shown in Fig. 7.11. Were it isolated, vortex 1 would have no
tendency to move itself, but would induce a velocity at a distance h away from its center
of Γ1 /(2πh). This induced velocity in fact advects vortex 2, to satisfy Kelvin’s circulation
theorem. Similarly, vortex 2 induces a velocity of vortex 1 of Γ2 /(2πh).
The center of rotation G is the point along the 1-2 axis for which the induced velocity is
zero, as is illustrated in Fig. 7.12. To calculate it we equate the induced velocities of each
vortex
Γ1 Γ2
= , (7.191)
2πhG 2π(h − hG )
(h − hG )Γ1 = hG Γ2 , (7.192)

© 05 December 2025, Joseph M. Powers. All rights reserved.

266 CHAPTER 7. VORTEX DYNAMICS

Figure 7.13: Sketch showing a pair of counterrotating vortices of equal strength.

hΓ1 = hG (Γ1 + Γ2 ), (7.193)

Γ1
hG = h . (7.194)
Γ1 + Γ2
A pair of equal strength counterrotating vortices is illustrated in Fig. 7.13. Such vortices
induce the same velocity in each other, so they will propagate as a pair at a fixed distance
from one another.

7.6.2 Image vortex

If we choose to model the fluid as inviscid, then there is no viscous stress, and we can no
longer enforce the no-slip condition at a wall. However at a slip wall, we must require that the
velocity vector be parallel to the wall. This can be described as a no-penetration condition,
that is looser than a no-slip condition. No-penetration through a surface with outer normal
n simply requires vT · n = 0, and allows no mass to penetrate such a surface, while allowing
slip at the surface because vT · αt 6= 0, with αt as the unit tangent vector to the surface.
We can model the motion of an ideal vortex separated by a distance h from an inviscid slip
wall by placing a so-called image vortex on the other side of the wall. The image vortex
will induce a velocity that when superposed with the original vortex, renders the resultant
velocity to be parallel to the wall. A vortex and its image vortex, that generates a straight
streamline at a wall, is sketched in Fig. 7.14.

7.6.3 Vortex sheets

We can model the slip line between two inviscid fluids moving at different velocities by
what is known as a vortex sheet. A vortex sheet is sketched in Fig. 7.15. Here we have a
distribution of small vortices, each of strength dΓ, on the x axis. Each of these vortices
induces a small velocity dv at an arbitrary point (x̃, ỹ). The influence of the point vortex at
(x, 0) is sketched in the figure. It generates a small velocity with magnitude
dΓ dΓ
d|v| = = p . (7.195)
2πh 2π (x̃ − x)2 + ỹ 2

© 05 December 2025, Joseph M. Powers. All rights reserved.

7.6. POTENTIAL FLOW OF IDEAL POINT VORTICES 267

h
Γ Γ

image vortex main vortex

Figure 7.14: Sketch showing a vortex and its image to simulate an inviscid wall.

dv
dv
du

Figure 7.15: Sketch showing schematic of vortex sheet.

© 05 December 2025, Joseph M. Powers. All rights reserved.

268 CHAPTER 7. VORTEX DYNAMICS

Using basic trigonometry, we can deduce that the influence of the single vortex of differential
strength on each velocity component is

−dΓỹ − dΓ
dx
ỹ
du = = dx, (7.196)
2π ((x̃ − x)2 + ỹ 2) 2π ((x̃ − x)2 + ỹ 2)
dΓ
dΓ(x̃ − x) dx
(x̃ − x)
dv = 2 2
= dx. (7.197)
2π ((x̃ − x) + ỹ ) 2π ((x̃ − x)2 + ỹ 2)

Here dΓ/dx is a measure of the strength of the vortex sheet. Let us account for the effects of
all of the differential vortices by integrating from x = −L to x = L and then letting L → ∞.
We obtain then the total velocity components u and v at each point to be
 
dΓ    
 L − x̃ L + x̃ 
u = lim −  dx
arctan + arctan , (7.198)
L→∞ 2π 
| ỹ ỹ 
{z } | {z }
→ ± π2 → ± π2
 1 dΓ
 − 2 dx , if ỹ > 0,
= (7.199)
 1 dΓ
2 dx
, if ỹ < 0,
dΓ
(L − x̃)2 + ỹ 2
dx
v = lim ln = 0. (7.200)
L→∞ 4π (L + x̃)2 + ỹ 2
So the vortex sheet generates no y component of velocity anywhere in the flow field and two
uniform x components of velocity of opposite sign above and below the x axis.

7.6.4 Potential of an ideal irrotational vortex

Let us calculate the velocity potential function φ associated with a single ideal irrotational
vortex. Consider an ideal irrotational vortex centered at the origin, and represent the velocity
field here in cylindrical coordinates:
Γo
vr = 0, vθ = , vz = 0. (7.201)
2πr
We have considered the same velocity field in Sec. 7.3 and, as represented in Cartesian
coordinates, in Ch. 3.11.10. Now in cylindrical coordinates the gradient operating on a
scalar function gives

∇φ = v, (7.202)
∂φ 1 ∂φ ∂φ Γo
er + eθ + ez = 0er + eθ + 0ez , (7.203)
∂r r ∂θ ∂z 2πr
∂φ
= 0, (7.204)
∂r

© 05 December 2025, Joseph M. Powers. All rights reserved.

7.6. POTENTIAL FLOW OF IDEAL POINT VORTICES 269

Figure 7.16: Lines of constant potential and velocity vector field for an ideal irrotational
vortex.

1 ∂φ Γo Γo
= , so φ = θ + C(r, z), (7.205)
r ∂θ 2πr 2π
∂φ
= 0. (7.206)
∂z
But because the partials of φ with respect to r and z are zero, C(r, z) is at most a constant,
that we can set to zero without losing any information regarding the velocity itself
Γo
φ= θ. (7.207)
2π
In Cartesian coordinates, we have
Γo y
φ= arctan . (7.208)
2π x
Lines of constant potential for the ideal vortex centered at the origin are sketched in Fig. 7.16.
Taking gz = 0, ro = 0, Eq. (7.136) gives for the pressure field
ρΓ2o
p − po = − . (7.209)
8π 2 r 2
The acceleration vector is entirely in the radial direction and is attributable to centripetal
acceleration:
v2 Γ2
a = − θ er = − 2o 3 er . (7.210)
r 4π r
It is easily verified that the linear momentum principle is satisfied in that
dp
ρa = − . (7.211)
dr
Lines of constant potential for the ideal vortex centered at the origin are sketched in Fig. 7.17.

© 05 December 2025, Joseph M. Powers. All rights reserved.

270 CHAPTER 7. VORTEX DYNAMICS

Figure 7.17: Isobars and the acceleration vector field for an ideal irrotational vortex.

7.6.5 Interaction of multiple vortices

Here we will consider the interactions of a large number of vortices by using the method of
superposition for the velocity potentials.
If we have two vortices with strengths Γ1 and Γ2 centered at arbitrary locations (x1 , y1 )
and (x2 , y2 ), as sketched in Fig. 7.18, the potential for each is given by
   
Γ1 y − y1 Γ2 y − y2
φ1 = arctan , φ2 = arctan . (7.212)
2π x − x1 2π x − x2

Because the equation governing the velocity potential, ∇2 φ = 0, is linear we can add the
two potentials and still satisfy the overall equation so that
   
Γ1 y − y1 Γ2 y − y2
φ= arctan + arctan , (7.213)
2π x − x1 2π x − x2
is a legitimate solution. Taking the gradient of φ,
     
Γ1 y − y1 Γ2 y − y2
∇φ = − − i
2π (x − x1 )2 + (y − y1 )2 2π (x − x2 )2 + (y − y2 )2
    
Γ1 x − x1 Γ2 x − x2
+ + j, (7.214)
2π (x − x1 )2 + (y − y1 )2 2π (x − x2 )2 + (y − y2 )2
so that
   
Γ1 y − y1 Γ2 y − y2
u(x, y) = − 2 2
− , (7.215)
2π (x − x1 ) + (y − y1 ) 2π (x − x2 )2 + (y − y2 )2
   
Γ1 x − x1 Γ2 x − x2
v(x, y) = 2 2
+ . (7.216)
2π (x − x1 ) + (y − y1 ) 2π (x − x2 )2 + (y − y2 )2

© 05 December 2025, Joseph M. Powers. All rights reserved.

7.6. POTENTIAL FLOW OF IDEAL POINT VORTICES 271

Figure 7.18: Two vortices at arbitrary locations.

Extending this to a collection of N vortices located at (xi , yi ) at a given time, we have the
following for the velocity field:
N 
X 
Γi y − yi
u(x, y) = − , (7.217)
i=1
2π (x − xi )2 + (y − yi )2
N 
X 
Γi x − xi
v(x, y) = . (7.218)
i=1
2π (x − xi )2 + (y − yi )2

Now to advect (that is, to move) the kth vortex, we move it with the velocity induced
by the other vortices, because vortices advect with the flow. Recalling that the velocity is
the time derivative of the position uk = dxk /dt, vk = dyk /dt, we then get the following 2N
non-linear ordinary differential equations for the 2N unknowns, the x and y positions of each
of the N vortices:
XN  
dxk Γi yk − yi
= − 2 + (y − y )2
, xk (0) = xok , k = 1, . . . , N,(7.219)
dt i=1,i6=k
2π (xk − xi ) k i

XN  
dyk Γi xk − xi
= 2 2
, yk (0) = yko , k = 1, . . . , N. (7.220)
dt i=1,i6=k
2π (xk − xi ) + (yk − yi )

This set of equations, except for three or fewer point vortices, must be integrated numerically.
These ordinary differential equations are highly non-linear and typically give rise to chaotic
motion of the point vortices. It is a similar calculation to the motion of point masses in a
Newtonian gravitational field, except that the essential variation goes as 1/r for vortices and
1/r 2 for Newtonian gravitational fields. Thus, the dynamics are different. Nevertheless just

© 05 December 2025, Joseph M. Powers. All rights reserved.

272 CHAPTER 7. VORTEX DYNAMICS

as calculations for large numbers of celestial bodies can give rise to solar systems, clusters
of planets, and galaxies, similar “galaxies” of vortices can be predicted with the equations
for vortex dynamics.

Example 7.1
Consider two ideal point vortices of equal strength Γ = 1, in a two-dimensional, incompressible,
irrotational flow field. The vortices are initially at (−1, 0)T and (1, 0)T . Find the trajectories of both
vortices as functions of time.

The equations are

dx1 y1 − y2
= − , x1 (0) = −1, (7.221)
dt 2π ((x1 − x2 )2 + (y1 − y2 )2 )
dx2 y2 − y1
= − , x2 (0) = 1, (7.222)
dt 2π ((x2 − x1 )2 + (y2 − y1 )2 )
dy1 x1 − x2
= , y1 (0) = 0, (7.223)
dt 2π ((x1 − x2 )2 + (y1 − y2 )2 )
dy2 x2 − x1
= , y2 (0) = 0. (7.224)
dt 2π ((x2 − x1 )2 + (y2 − y1 )2 )

Divide the equation for x2 evolution by that for x1 evolution, and divide the equation for y2 evolution
by that for y1 evolution to get

dx2 dy2
= −1, = −1. (7.225)
dx1 dy1

Integrate and apply the initial conditions to get

x2 = −x1 , y2 = −y1 . (7.226)

Use this to reduce to two ordinary differential equations:

dx1 y1
= − , x1 (0) = −1, (7.227)
dt π ((2x1 )2 + (2y1 )2 )
dy1 x1
= , y1 (0) = 0. (7.228)
dt π ((2x1 )2 + (2y1 )2 )

Now divide the equation for y1 evolution by that for x1 evolution to get

dy1 x1
=− . (7.229)
dx1 y1

This gives y1 dy1 + x1 dx1 = 0. Integrate and apply the initial conditions to get

x21 + y12 = 1. (7.230)

So the motion of the vortex is a circle of radius unity centered at the origin. We can then get the
variation with time by eliminating x1 to get
p
dy1 ± 1 − y12
= , y1 (0) = 0. (7.231)
dt 4π

© 05 December 2025, Joseph M. Powers. All rights reserved.

7.6. POTENTIAL FLOW OF IDEAL POINT VORTICES 273

The relevant branch is the negative branch. We integrate and get

 
t
y1 (t) = − sin . (7.232)
4π
The other variables are easily obtained:
     
t t t
x1 (t) = − cos , y2 (t) = sin , x2 (t) = cos . (7.233)
4π 4π 4π

7.6.6 Pressure field

We have thus far examined essentially only the kinematics of vortices. We have actually
used dynamics in our incorporation of the Helmholtz equation and Kelvin’s theorem, but
their simple results really only justify the use of a simple kinematics. Dynamics asks what
are the forces that give rise to the motion. Here, we will assume there is no body force, that
the fluid is inviscid, in which case it must be pressure forces that give rise to the motion,
and that the fluid is at rest at infinity. We have the proper conditions for which Bernoulli’s
equation can be used to give the pressure field. We consider two cases, a single stationary
point vortex, and a group of N moving point vortices.

[Link] Single stationary vortex

If we take p = p∞ in the far field and fi = g = 0, this steady flow gives us
1 T p 1 T p∞
v ·v+ = v∞ · v∞ + , (7.234)
2 ρ 2 ρ
 2
1 Γo p p∞
+ = 0+ , (7.235)
2 2πr ρ ρ
ρΓ2o 1
p(r) = p∞ − 2 2 . (7.236)
8π r
The pressure goes to negative infinity at the origin. This is obviously unphysical. It can
be corrected by including viscous effects, that turn out not to substantially alter our main
conclusions.

[Link] Group of N vortices

For a collection of N vortices, the flow is certainly not steady, and we must in general retain
the time-dependent velocity potential in Bernoulli’s equation, Eq. 6.196), yielding
∂φ 1 p
+ (∇φ)T · ∇φ + = f (t). (7.237)
∂t 2 ρ

© 05 December 2025, Joseph M. Powers. All rights reserved.

274 CHAPTER 7. VORTEX DYNAMICS

wall vortex line

wall streamline

x
v

wall streamline
wall vortex line

Figure 7.19: Wall streamlines and vortex lines at wall y = 0.

Now we require that as r → ∞ that p → p∞ . We also know that as r → ∞ that φ → 0,

hence ∇φ → 0 as well. Hence as r → ∞, we have p∞ /ρ = f (t). So our final result is
1 ∂φ
p = p∞ − ρ(∇φ)T · ∇φ − . (7.238)
2 ∂t
So with a knowledge of the velocity field through φ, we can determine the pressure field that
must have given rise to that velocity field.

7.7 Streamlines and vortex lines at walls

It seems odd that a streamline can be defined at a wall where the velocity is formally zero, but
in the neighborhood of the wall, the fluid velocity is small but non-zero. We can extrapolate
the position of streamlines near the wall to the wall to define a wall streamline. We shall
also consider a so-called vortex line, a line everywhere parallel to the vorticity vector, at the
wall.
We consider the geometry sketched in Fig. 7.19. Here the x − z plane is locally attached
to a wall at y = 0, and the y direction is normal to the wall. Wall streamlines and vortex
lines are sketched in the figure.
Because the flow satisfies a no-slip condition, we have at the wall
u(x, y = 0, z) = 0, v(x, y = 0, z) = 0, w(x, y = 0, z) = 0. (7.239)
Because of this, partial derivatives of all velocities with respect to either x or z will also be
zero at y = 0:
∂u ∂u ∂v ∂v ∂w ∂w
= = = = = = 0. (7.240)
∂x y=0 ∂z y=0 ∂x y=0 ∂z y=0 ∂x y=0 ∂z y=0

© 05 December 2025, Joseph M. Powers. All rights reserved.

7.7. STREAMLINES AND VORTEX LINES AT WALLS 275

Near the wall, the velocity is near zero, so the Mach number is small, and the flow is well
modeled as incompressible. So here, the mass conservation equation implies that ∇T · v = 0,
so applying this at the wall, we get
∂u ∂v ∂w
+ + = 0, so (7.241)
∂x ∂y ∂z y=0
| {zy=0} y=0
| {z }
=0 =0
∂v
= 0. (7.242)
∂y y=0

Now let us examine the behavior of u, v, and w, as we leave the wall in the y direction.
Consider a Taylor series of each:
∂u 1 ∂2u
u = u|y=0 + y+ y2 + . . . , (7.243)
| {z } ∂y y=0 2 ∂y 2 y=0
=0
∂v 1 ∂2v
v = v|y=0 + y+ y2 + . . . , (7.244)
| {z } ∂y y=0 2 ∂y 2 y=0
=0 | {z }
=0
∂w 1 ∂2w
w = w|y=0 + y+ y2 + . . . . (7.245)
| {z } ∂y y=0 2 ∂y 2 y=0
=0

So we get
∂u
u = y + ..., (7.246)
∂y y=0
2
1 ∂ v
v = y2 + . . . , (7.247)
2 ∂y 2 y=0
∂w
w = y + .... (7.248)
∂y y=0

Now for streamlines, we must have

dx dy dz
= = . (7.249)
u v w
For the streamline near the wall, consider just dx/u = dz/w, and also tag the streamline as
dzs , so that the slope of the wall streamline, that is the tangent of the angle θ between the
wall streamline and the x axis is
∂w
dzs w ∂y
y=0
tan θ = = lim = . (7.250)
dx y=0
y→0 u ∂u
∂y
y=0

© 05 December 2025, Joseph M. Powers. All rights reserved.

276 CHAPTER 7. VORTEX DYNAMICS

Now consider the vorticity vector evaluated at the wall:

∂w ∂v ∂w
ωx |y=0 = − = , (7.251)
∂y y=0 ∂z y=0 ∂y y=0
| {z }
=0
∂u ∂w
ωy |y=0 = − = 0, (7.252)
∂z y=0 ∂x y=0
| {z } | {z }
=0 =0
∂v ∂u ∂u
ωz |y=0 = − =− . (7.253)
∂x ∂y ∂y
| {zy=0} y=0 y=0

=0

So we see that on the wall at y = 0, the vorticity vector has no component in the y direction.
Hence, it must be parallel to the wall itself. Further, we can then define the slope of the
vortex line, dzv /dx, at the wall in the same fashion as we define a streamline:

∂u
dzv ωz ∂y
y=0 1
= =− = − dzs . (7.254)
dx y=0 ωx ∂w
dx y=0
∂y
y=0

Because the slope of the vortex line is the negative reciprocal of the slope of the streamline,
we have that at a no-slip wall, streamlines are orthogonal to vortex lines. We also note that
streamlines are orthogonal to vortex lines for flow with variation in the x and y directions
only. For general three-dimensional flows away from walls, we do not expect the two lines
to be orthogonal.
This motivates a local coordinate system attached to the wall with the x axis is aligned
with the wall streamline and the z axis aligned with the wall vortex line. As before the y
axis is normal to the wall. The coordinate system aligned with the wall streamlines and
vortex lines is sketched in Fig. 7.20. In the figure we take the direction n to be normal to
the wall.

© 05 December 2025, Joseph M. Powers. All rights reserved.

7.7. STREAMLINES AND VORTEX LINES AT WALLS 277

v
x

Figure 7.20: Coordinate system aligned with wall streamlines and vortex lines.

© 05 December 2025, Joseph M. Powers. All rights reserved.

278 CHAPTER 7. VORTEX DYNAMICS

© 05 December 2025, Joseph M. Powers. All rights reserved.

Chapter 8

Potential flow

see Powers, Chapter 8,

see Panton, Chapter 18,
see Yih, Chapter 4,
see Lamb, Chapter 4,
see Kuethe and Chow, Chapter 4.

This chapter will expand upon potential flow, introduced in Section 7.6, and will mainly be
restricted to steady, two-dimensional planar, incompressible potential flow. Such flows can
be characterized by a scalar potential field. As is typical for such fields, gradients of the
potential induce flow. A beautiful mathematical theory was developed for potential flows
in the nineteenth century and will be described in this chapter. Additionally, potential flow
solutions can be applied in disparate fields, because the equations governing potential flow
of a fluid are identical in form to those governing some forms of energy and mass diffusion,
as well as electro-magnetics.
Despite its beauty, in some ways it is impractical for many engineering applications,
though not all. As the theory necessarily ignores all vorticity generating mechanisms, it
must ignore viscous effects. Consequently, the theory is incapable of predicting drag forces
on solid bodies. Consequently, those who needed to know the drag resorted in the nineteenth
century to more empirically based methods.
In the early twentieth century, Prandtl took steps to reconcile the practical viscous world
of engineering with the more mathematical world of potential flow with his viscous boundary
layer theory. He showed that indeed potential flow solutions could be of value away from
no-slip walls, and provided a recipe to fix the solutions in the neighborhood of the wall. In so
doing, he opened a new field of applied mathematics known as matched asymptotic analysis.
So why study potential flows? The following arguments offer some justification.

• low speed aerodynamics are often well described by potential flow theory,

• portions of many real flow fields are captured by potential theory, and those that are
not can often be remedied by application of a viscous boundary layer theory,

279
280 CHAPTER 8. POTENTIAL FLOW

• study of potential flow solutions can give great insight into fluid behavior and aid in
the honing of a more precise intuition,

• potential flow solutions are useful as test cases for verification of numerical methods,
and

• there is pedantic and historical value in understanding potential flow.

8.1 Stream functions and velocity potentials

We first consider stream functions and velocity potentials. We have seen velocity potentials
before in Ch. 7.6 in study of ideal vortices. In this chapter, we will adopt the same assumption
of irrotationality, and further require that the flow be two-dimensional. Recall if a flow
velocity is confined to the (x, y) plane, then the vorticity vector is confined to the z direction
and takes the form first shown in Eq. (3.206):
 
0
ω= 0 . (8.1)
∂v ∂u
∂x
− ∂y

So if the flow is two-dimensional and irrotational, we have

∂v ∂u
− = 0. (8.2)
∂x ∂y

Moreover, because of irrotationality, we can express the velocity vector v as the gradient of
a potential φ, the velocity potential, as first shown in Eq. (6.191):

v = ∇φ. (8.3)

With this definition, fluid flows from regions of low velocity potential to regions of high
velocity potential. The scalar velocity components are

∂φ ∂φ
u= , v= . (8.4)
∂x ∂y

We see by substitution into the irrotationality condition, Eq. (8.2), that this is true identi-
cally:    
∂v ∂u ∂ ∂φ ∂ ∂φ
− = − = 0. (8.5)
∂x ∂y ∂x ∂y ∂y ∂x
Now for two-dimensional incompressible flows, we have by specializing Eq. (6.49) that

∂u ∂v
+ = 0. (8.6)
∂x ∂y

© 05 December 2025, Joseph M. Powers. All rights reserved.

8.1. STREAM FUNCTIONS AND VELOCITY POTENTIALS 281

Substituting for u and v in favor of φ, we get Laplace’s equation for φ, a special case of that
seen earlier in Eq. (7.188):
   
∂ ∂φ ∂ ∂φ
+ = 0, (8.7)
∂x ∂x ∂y ∂y
∇2 φ = 0. (8.8)

Now if the flow is incompressible, we can also define the stream function ψ as follows:

∂ψ ∂ψ
u= , v=− . (8.9)
∂y ∂x

This the scalar analog of Eq. (6.87) for a two-dimensional flow with ψ = (0, 0, ψ)T . Direct
substitution into the incompressible mass conservation equation, Eq. (8.6), shows that this
yields an identity:    
∂u ∂v ∂ ∂ψ ∂ ∂ψ
+ = + − = 0. (8.10)
∂x ∂y ∂x ∂y ∂y ∂x
Now, in an equation that will be critically important soon, we can set our definitions of u
and v in terms of φ and ψ, Eqs. (8.4, 8.9), equal to each other, as they must be:

∂φ ∂ψ
= , (8.11)
∂x
|{z} ∂y
|{z}
u u
∂φ ∂ψ
= − . (8.12)
∂y ∂x}
| {z
|{z}
v v

Now if we differentiate the first equation with respect to y, and the second with respect to
x we see
∂2φ ∂2ψ
= , (8.13)
∂y∂x ∂y 2
∂2φ ∂2ψ
= − 2. (8.14)
∂x∂y ∂x

Now subtract the second from the first to get Laplace’s equation for ψ:

∂2ψ ∂2ψ
0 = + 2, (8.15)
∂y 2 ∂x
2
∇ ψ = 0. (8.16)

Let us now examine lines of constant φ (equipotential lines) and lines of constant ψ (that
we will see are streamlines). So take φ = C1 , ψ = C2 . Because φ = φ(x, y), we take the total

© 05 December 2025, Joseph M. Powers. All rights reserved.

282 CHAPTER 8. POTENTIAL FLOW

derivative on a curve of constant φ and get

∂φ ∂φ
dφ = dx + dy = 0, (8.17)
∂x ∂y
= u dx + v dy = 0, (8.18)
dy u
= − . (8.19)
dx φ=C1 v

Now for ψ = ψ(x, y) we similarly get on curves of constant ψ that

∂ψ ∂ψ
dψ = dx + dy = 0, (8.20)
∂x ∂y
= −v dx + u dy = 0, (8.21)
dy v
= . (8.22)
dx ψ=C2 u

We note two features:

dy 1
• =− ; hence, lines of constant φ are orthogonal to lines of constant ψ,
dx φ=C1 dy
|
dx ψ=C
2
and

• on ψ = C2 , we see that dx/u = dy/v; hence, by Eq. (3.49), lines of ψ = C2 must be

streamlines.
As an aside, we note that the definition of the stream function u = ∂ψ/∂y, v = −∂ψ/∂x,
can be rewritten as
∂ψ dx ∂ψ dy
= , =− . (8.23)
∂y dt ∂x dt
This is a common form from classical dynamics in which we can interpret ψ as the Hamil-
tonian1 of the system. We shall not pursue this path, but note that a significant literature
exists for Hamiltonian systems; see Powers and Sen, (2015), Ch. 9.6.6, for an introduction
or Goldstein (1950), Ch. 7, for a detailed discussion.
Now the study of φ and ψ is essentially kinematics. The only incursion of dynamics is that
we must have irrotational flow. Recalling the Helmholtz equation, Eq. (7.156), we realize that
we can only have potential flow when the vorticity generating mechanisms (three-dimensional
effects, non-conservative body forces, baroclinic effects, and viscous effects) are suppressed.
In that case, the dynamics, that is the driving force for the fluid motion, can be understood
1
William Rowan Hamilton, 1805-1865, Dublin-based Anglo-Irish mathematician. He discovered the
quaternion group, an extension of complex numbers, while walking along a canal and engraved it in the
Broom Bridge. He invented the notion of the dot and cross products, coined the terms “scalar” and “ten-
sor,” and was the first to use “vector” in the modern sense. He introduced the ∇ operator in 1837, albeit
∇

in a rotated form, , based on the harp. Some believe quaternions were satirized in Lewis Carroll’s Alice in
Wonderland and that the Mad Hatter’s tea party was to argue for a return to Euclidean geometry, rather
than the topsy-turvy universe described by quaternion methods.

© 05 December 2025, Joseph M. Powers. All rights reserved.

8.2. MATHEMATICS OF COMPLEX VARIABLES 283

in the context of the unsteady Bernoulli equation, Eq. (6.196, taken for incompressible flow
and negligible body force, in which limit, Eq. (6.182) can be taken to reduce to P = p:

∂φ 1 T p
+ ∇ φ · ∇φ + = f (t). (8.24)
∂t 2 ρ

We do not have to require steady flow to have a potential flow field. It is also easy to correct
for the presence of a conservative body force.
Now solutions to the two key equations of potential flow ∇2 φ = 0, ∇2 ψ = 0, are most effi-
ciently studied using methods involving complex variables. We will delay discussing solutions
until we have reviewed the necessary mathematics.

8.2 Mathematics of complex variables

Here we briefly introduce relevant elements of complex variable theory. Recall that the
imaginary number i is defined such that
√
i2 = −1, i = −1. (8.25)

8.2.1 Euler’s formula

We can get the useful Euler’s formula by considering the following Taylor expansions of
common functions about t = 0:
1 2 1 3 1 4 1 5
et = 1 + t + t + t + t + t ..., (8.26)
2! 3! 4! 5!
1 2 1 3 1 4 1 5
sin t = 0 + t + 0 t − t + 0 t + t . . . , (8.27)
2! 3! 4! 5!
1 2 1 3 1 4 1
cos t = 1 + 0t − t + 0 t + t + 0 t5 . . . (8.28)
2! 3! 4! 5!
With these expansions, now consider the following combinations: (cos t + i sin t)t=θ and
et |t=iθ :

1 2 1 1 1
cos θ + i sin θ = 1 + iθ − θ − i θ3 + θ4 + i θ5 + . . . , (8.29)
2! 3! 4! 5!
1 1 1 1
eiθ = 1 + iθ + (iθ)2 + (iθ)3 + (iθ)4 + (iθ)5 + . . . , (8.30)
2! 3! 4! 5!
1 2 1 3 1 4 1 5
= 1 + iθ − θ − i θ + θ + i θ + . . . (8.31)
2! 3! 4! 5!
As the two series are identical, we have Euler’s formula:

eiθ = cos θ + i sin θ. (8.32)

© 05 December 2025, Joseph M. Powers. All rights reserved.

284 CHAPTER 8. POTENTIAL FLOW

8.2.2 Polar and Cartesian representations

We take x ∈ R1 , y ∈ R1 and define the complex number z to be

z = x + iy. (8.33)

We say that z ∈ C1 . We define the operator ℜ as selecting the real part of a complex number
and ℑ as selecting the imaginary part of a complex number. For Eq. (8.33), we see

ℜ(z) = x, ℑ(z) = y. (8.34)

p
Both operators ℜ and ℑ take C1 → R1 . We can multiply and divide Eq. (8.33) by x2 + y 2
to obtain  
p  x y 
 
z= x2 +  y2 p +i p . (8.35)
| {z }  x + y
2 2 x +y 
2 2
r | {z } | {z }
cos θ sin θ

Noting the similarities between this and the transformation between Cartesian and polar
coordinates suggests we adopt
p x y
r= x2 + y 2 , cos θ = p , sin θ = p . (8.36)
x2 + y 2 x2 + y 2

Thus, we have

z = r (cos θ + i sin θ) , (8.37)

= reiθ . (8.38)

We often say that a complex number can be characterized by its magnitude |z| and its
argument, θ; we say then

r = |z|, (8.39)
θ = arg z. (8.40)

Here, r ∈ R1 and θ ∈ R1 . Note that |eiθ | = 1. If x > 0, the function arg z is identical to
arctan(y/x) and is suggested by the polar and Cartesian representation of z as shown in
Fig. 8.1. However, we recognize that the ordinary arctan (also known as tan−1 ) function
maps onto the range [−π/2, π/2], while we would like arg to map onto [−π, π]. For example,
to capture the entire unit circle if r = 1, we need θ ∈ [−π, π]. This can be achieved if we
define arg, also known as Tan−1 as follows:
!
y
arg z = arg(x + iy) = Tan−1 (x, y) = 2 arctan p . (8.41)
x + x2 + y 2

© 05 December 2025, Joseph M. Powers. All rights reserved.

8.2. MATHEMATICS OF COMPLEX VARIABLES 285

iy

x
x

Figure 8.1: Polar and Cartesian representation of a complex number z.

Table 8.1: Comparison of the action of arg, Tan−1 , and arctan.

x y arg(x + iy) Tan−1 (x, y) arctan(y/x)

1 1 π/4 π/4 π/4
−1 1 3π/4 3π/4 −π/4
−1 −1 −3π/4 −3π/4 π/4
1 −1 −π/4 −π/4 −π/4

© 05 December 2025, Joseph M. Powers. All rights reserved.

286 CHAPTER 8. POTENTIAL FLOW

y y
x
x

Figure 8.2: Comparison of Tan−1 (x, y) and tan−1 (y/x) .

Iff x > 0, this reduces to the more typical

y  y
−1 −1
arg z = arg(x + iy) = Tan (x, y) = arctan = tan , x > 0. (8.42)
x x
The preferred and more general form is Eq. (8.41). We give simple function evaluations
involving arctan and Tan−1 for selected values of x and y in Table 8.1. Use of Tan−1
effectively captures the correct quadrant of the complex plane corresponding to different
positive and negative values of x and y. The function is sometimes known as Arctan or
atan2. A comparison of Tan−1 (x, y) and tan−1 (y/x) is given in Fig. 8.2.
Now we can define the complex conjugate z as
z = x − iy, (8.43)
!
p x y
= x2 + y 2 p − ip , (8.44)
x2 + y 2 x2 + y 2
= r (cos θ − i sin θ) , (8.45)
= r (cos(−θ) + i sin(−θ)) , (8.46)
= re−iθ . (8.47)
Note now that
zz = (x + iy)(x − iy) = x2 + y 2 = |z|2 , (8.48)
= reiθ re−iθ , (8.49)
= r2, (8.50)
= |z|2 . (8.51)
We also have
eiθ − e−iθ
sin θ = , (8.52)
2i

© 05 December 2025, Joseph M. Powers. All rights reserved.

8.2. MATHEMATICS OF COMPLEX VARIABLES 287

eiθ + e−iθ
cos θ = . (8.53)
2

Example 8.1
Use the polar representation of z to find all roots to the algebraic equation

z 4 = 1. (8.54)

We know that z = reiθ . We also note that the constant 1 can be represented as

1 = e2nπi , n = 0, 1, 2, ... (8.55)

This will be useful in finding all roots to our equation. With this representation, Eq. (8.54) becomes

r4 e4iθ = e2nπi , n = 0, 1, 2, .... (8.56)

We have a solution when

nπ
r = 1, θ= , n = 0, 1, 2, ... (8.57)
2
There are unique solutions for n = 0, 1, 2, 3. For larger n, the solutions repeat. So we have four solutions

z = e0i , z = eiπ/2 , z = eiπ , z = e3iπ/2 . (8.58)

In Cartesian form, the four solutions are

z = ±1, z = ±i. (8.59)

Example 8.2
Find all roots to

z 3 = i. (8.60)

We proceed in a similar fashion as for the previous example. We know that

i = ei(π/2+2nπ) , n = 0, 1, 2, ... (8.61)

Substituting this into Eq. (8.60), we get

r3 e3iθ = ei(π/2+2nπ) , n = 0, 1, 2, ... (8.62)

Solving, we get
π 2nπ
r = 1, θ= + . (8.63)
6 3

© 05 December 2025, Joseph M. Powers. All rights reserved.

288 CHAPTER 8. POTENTIAL FLOW

iy iy

x x

Figure 8.3: Sketch of solutions to z 4 = 1 and z 3 = i in the complex plane.

There are only three unique values of θ, those being θ = π/6, θ = 5π/6, θ = 3π/2. So the three roots
are
z = eiπ/6 , z = e5iπ/6 , z = e3iπ/2 . (8.64)
In Cartesian form these roots are
√ √
3+i − 3+i
z= , z= , z = −i. (8.65)
2 2
Sketches of the solutions to this and the previous example are shown in Fig. 8.3. For both examples,
the roots are uniformly distributed about the unit circle, with four roots for the quartic equation and
three for the cubic.

8.2.3 Cauchy-Riemann equations

Now it is possible to define complex functions of complex variables W (z). For example, take
a complex function to be defined as
W (z) = z 2 + z, (8.66)
= (x + iy)2 + (x + iy), (8.67)
= x2 + 2xyi − y 2 + x + iy, (8.68)


= x2 + x − y 2 + i (2xy + y) . (8.69)
In general, we can say
W (z) = φ(x, y) + iψ(x, y). (8.70)
Here φ and ψ are real functions of real variables.
Now W (z) is defined as analytic at zo if dW/dz exists at zo and is independent of the
direction in which it was calculated. That is, using the definition of the derivative
dW W (zo + ∆z) − W (zo )
= . (8.71)
dz zo ∆z

© 05 December 2025, Joseph M. Powers. All rights reserved.

8.2. MATHEMATICS OF COMPLEX VARIABLES 289

Now there are many paths that we can choose to evaluate the derivative. Let us consider
two distinct paths, y = C1 and x = C2 . We will get a result that can be shown to be valid
for arbitrary paths. For y = C1 , we have ∆z = ∆x, so

dW W (xo + iyo + ∆x) − W (xo + iyo ) ∂W

= = . (8.72)
dz zo ∆x ∂x y

For x = C2 , we have ∆z = i∆y, so

dW W (xo + iyo + i∆y) − W (xo + iyo ) 1 ∂W ∂W

= = = −i . (8.73)
dz zo i∆y i ∂y x ∂y x

Now for an analytic function, we need

∂W ∂W
= −i , (8.74)
∂x y ∂y x

or, expanding, we need

 
∂φ ∂ψ ∂φ ∂ψ
+i = −i +i , (8.75)
∂x ∂x ∂y ∂y
∂ψ ∂φ
= −i . (8.76)
∂y ∂y

For equality, and thus path independence of the derivative, we require

∂φ ∂ψ ∂φ ∂ψ
= , =− . (8.77)
∂x ∂y ∂y ∂x

These are the well known Cauchy2 -Riemann equations for analytic functions of complex
variables. They are identical to our kinematic equations, Eqs. (8.11, 8.12), for incompressible
irrotational fluid mechanics. Consequently, any analytic complex function is guaranteed to
be a physical solution. There are an infinite number of functions from which to choose!
We define the complex velocity potential as

W (z) = φ(x, y) + iψ(x, y), (8.78)

and taking a derivative of this potential, we have

dW ∂φ ∂ψ
= +i , (8.79)
dz ∂x ∂x
= u − iv. (8.80)
2
Augustin-Louis Cauchy, 1789-1857, French mathematician and military engineer, worked in complex
analysis, optics, and theory of elasticity.

© 05 December 2025, Joseph M. Powers. All rights reserved.

290 CHAPTER 8. POTENTIAL FLOW

Because the direction of the derivative does not matter, we can equivalently say
   
dW ∂φ ∂ψ ∂ψ ∂φ
= −i +i = −i = u − iv. (8.81)
dz ∂y ∂y ∂y ∂y

We can associate the velocity magnitude with the magnitude of dW/dz:

2
dW dW dW
= , (8.82)
dz dz dz
= (u + iv)(u − iv), (8.83)
= u2 + v 2 , (8.84)
dW √
= u2 + v 2 . (8.85)
dz

Now most common functions are easily shown to be analytic. For example for the function
of Eq. (8.66)
W (z) = z 2 + z, (8.86)
that we have seen can be expressed as

W (z) = (x2 + x − y 2) + i(2xy + y), (8.87)

we have

φ(x, y) = x2 + x − y 2 , ψ(x, y) = 2xy + y, (8.88)

∂φ ∂ψ
= 2x + 1, = 2y, (8.89)
∂x ∂x
∂φ ∂ψ
= −2y, = 2x + 1. (8.90)
∂y ∂y

The fields of ψ and φ are plotted in Fig. 8.4. The Cauchy-Riemann equations are satisfied
because ∂φ/∂x = ∂ψ/∂y and ∂φ/∂y = −∂ψ/∂x. Moreover,

∂ ∂
∇2 φ = (2x + 1) + (−2y) = 2 − 2 = 0, (8.91)
∂x ∂y
∂ ∂
∇2 ψ = (2y) + (2x + 1) = 0 + 0 = 0, (8.92)
∂x ∂y
   
u 2x + 1
v = = . (8.93)
v −2y

So the derivative is independent of direction, and we can say

dW ∂W
= = (2x + 1) + i(2y) = 2(x + iy) + 1 = 2z + 1. (8.94)
dz ∂x y

© 05 December 2025, Joseph M. Powers. All rights reserved.

8.3. ELEMENTARY COMPLEX POTENTIALS 291

y 0

- 1

- 2
- 2 - 1 0 1 2
x

Figure 8.4: Fields of ψ(x, y), φ(x, y) corresponding to the complex potential W (z) = z 2 + z.

We could get this result by ordinary rules of derivatives for real functions.
For an example of a non-analytic function consider W (z) = z. Thus,

W (z) = x − iy. (8.95)

So φ = x and ψ = −y, ∂φ/∂x = 1, ∂φ/∂y = 0, and ∂ψ/∂x = 0, ∂ψ/∂y = −1. Because

∂φ/∂x 6= ∂ψ/∂y, the Cauchy-Riemann equations are not satisfied, and the derivative depends
on direction.

8.3 Elementary complex potentials

Let us examine some simple analytic functions and see the fluid mechanics to which they
correspond.

8.3.1 Uniform flow

Take
W (z) = Bz, with B ∈ C1 . (8.96)
Then
dW
= B = u − iv. (8.97)
dz
Because B is complex, we can say

B = Uo e−iα = Uo cos α − iUo sin α. (8.98)

© 05 December 2025, Joseph M. Powers. All rights reserved.

292 CHAPTER 8. POTENTIAL FLOW

iy
2

0 x

−1

−2
−2 −1 0 1 2

Figure 8.5: Streamlines and equipotential lines for uniform flow.

Thus, we get
u = Uo cos α, v = Uo sin α. (8.99)
This represents a spatially uniform flow with streamlines inclined at angle α to the x axis.
We also see that because W = Bz = (Uo cos α − iUo sin α)(x + iy) that

φ = Uo (x cos α + y sin α), ψ = Uo (y cos α − x sin α). (8.100)

The flow is sketched in Fig. 8.5.

8.3.2 Sources and sinks

Take
W (z) = B ln z, with B ∈ R1 . (8.101)
With z = reiθ we have ln z = ln r + iθ. So

W (z) = B ln r + iBθ. (8.102)

Consequently, we have for the velocity potential and stream function

φ = B ln r, ψ = Bθ. (8.103)

Now v = ∇φ, so
∂φ B 1 ∂φ
vr = = , vθ = = 0. (8.104)
∂r r r ∂θ

© 05 December 2025, Joseph M. Powers. All rights reserved.

8.3. ELEMENTARY COMPLEX POTENTIALS 293

iy

Figure 8.6: Streamlines and equipotential lines for source flow.

So the velocity is all radial, and becomes infinite at r = 0. We can show that the volume
flow rate per unit depth is bounded, and is in fact a constant. For this two-dimensional flow,
we really want to consider the volumetric flow rate per unit depth. We shall call this Q and
recognize it has units of m2 /s.3 Then with dA as a differential area per unit depth, we have
dA = r dθ, and the volume flow rate per unit depth Q through a surface is
Z Z 2π Z 2π
T B
Q= v · n dA = vr r dθ = r dθ = 2πB. (8.105)
A 0 0 r
The volume flow rate is a constant. If B > 0, we have a source. If B < 0, we have a sink.
The potential for a source/sink is often written as
Q
W (z) = ln z. (8.106)
2π
For a source located at a point zo that is not at the origin, we can say
Q
W (z) = ln(z − zo ). (8.107)
2π
The flow is sketched in Fig. 8.6.

8.3.3 Point vortices

For an ideal point vortex, identical to what we studied in Ch. 7.6.4, we have

W (z) = iB ln z, with B ∈ R1 . (8.108)

3
It is also common to interpret Q as a volume flow rate with units of m3 /s. Our notation would have to
be adjusted if we took this interpretation.

© 05 December 2025, Joseph M. Powers. All rights reserved.

294 CHAPTER 8. POTENTIAL FLOW

iy

Figure 8.7: Streamlines, equipotential, and velocity vectors lines for a point vortex.

So
W (z) = iB (ln r + iθ) = −Bθ + iB ln r. (8.109)
Consequently,
φ = −Bθ, ψ = B ln r. (8.110)
We get the velocity field from

∂φ 1 ∂φ B
vr = = 0, vθ = =− . (8.111)
∂r r ∂θ r
So we see that the streamlines are circles about the origin, and there is no radial component
of velocity. Consider the circulation of this flow, Eq. (7.2):
I Z 2π
T B
Γ= v · dr = − r dθ = −2πB. (8.112)
C 0 r

So we often write the complex potential in terms of the ideal vortex strength Γo :

iΓo
W (z) = − ln z. (8.113)
2π
For an ideal vortex not at z = zo , we say

iΓo
W (z) = − ln(z − zo ). (8.114)
2π
The point vortex flow is sketched in Fig. 8.7.

© 05 December 2025, Joseph M. Powers. All rights reserved.

8.3. ELEMENTARY COMPLEX POTENTIALS 295

iy

Q x Q
a a

Figure 8.8: Sketch for source-wall interaction.

8.3.4 Superposition of sources

Because the equation for velocity potential is linear, we can use the method of superposition
to create new solutions as summations of elementary solutions. Say we want to model the
effect of a wall on a source as sketched in Fig. 8.8. At the wall we want u(0, y) = 0. That is
 
dW
ℜ = ℜ {u − iv} = 0, on z = iy. (8.115)
dz

Now let us place a source at z = a and superpose a source at z = −a, where a is a real
scalar; a ∈ R1 . So we have for the complex potential
Q Q
W (z) = ln(z − a) + ln(z + a), (8.116)
|2π {z } |2π {z }
original image
Q
= (ln(z − a) + ln(z + a)) , (8.117)
2π
Q
= ln ((z − a)(z + a)) , (8.118)
2π
Q
= ln(z 2 − a2 ), (8.119)
2π
dW Q 2z
= . (8.120)
dz 2π z 2 − a2
Now on z = iy, that is the location of the wall, we have
 
dW Q 2iy
= = u − iv. (8.121)
dz 2π −y 2 − a2

The term is purely imaginary; hence, the real part is zero, and we have u = 0 on the wall,
as desired.

© 05 December 2025, Joseph M. Powers. All rights reserved.

296 CHAPTER 8. POTENTIAL FLOW

On the wall we do have a non-zero y component of velocity. Hence the wall is not a
no-slip wall. On the wall we have then

Q y
v= . (8.122)
π y 2 + a2

We find the location on the wall of the maximum v velocity by setting the derivative with
respect to y to be zero,
∂v Q (y 2 + a2 ) − y(2y)
= = 0. (8.123)
∂y π (y 2 + a2 )2
Solving, we find critical points at y = ±a. It can be shown that v is a local maximum at
y = a and a local minimum at y = −a. So on the wall we have

1 2 1 Q2 y2
(u + v 2 ) = . (8.124)
2 2 π 2 (y 2 + a2 )2

We can use Bernoulli’s equation to find the pressure field, assuming steady flow and that
p → po as r → ∞. So Bernoulli’s equation in this limit

1 T p po
∇ φ · ∇φ + = , (8.125)
2 ρ ρ

reduces to
1 Q2 y2
p = po − ρ 2 . (8.126)
2 π (y 2 + a2 )2
The pressure is po at y = 0 and is po as y → ∞. By integrating the pressure over the wall
surface, one would find the net force on the wall induced by the source.

8.3.5 Flow in corners

Flow in or around a corner can be modeled by the complex potential

W (z) = Bz n , with B ∈ R1 , (8.127)

n
= B reiθ , (8.128)
= Br n einθ , (8.129)
= Br n (cos(nθ) + i sin(nθ)). (8.130)

So we have
φ = Br n cos nθ, ψ = Br n sin nθ. (8.131)
Now recall that lines on which ψ is constant are streamlines. Examining the stream function,
we obviously have streamlines when ψ = 0 that occurs whenever θ = 0 or θ = π/n.

© 05 December 2025, Joseph M. Powers. All rights reserved.

8.3. ELEMENTARY COMPLEX POTENTIALS 297

iy iy iy
ψ ψ
4

x 0 x x
φ φ
-2 -2

-4 -4
φ φ
-4 -2 0 2 -44 -2 0 2 4

Figure 8.9: Sketch for impingement flow, stagnation flow, and flow in a corner, n = 2.

For example if n = 2, we model a stream striking a flat wall; kinematics of this have been
previously described in Ch. 3.11.8. For this flow, we have

W (z) = Bz 2 , (8.132)
= B(x + iy)2 , (8.133)
= B((x2 − y 2) + i(2xy)), (8.134)
φ = B(x2 − y 2 ), ψ = B(2xy). (8.135)

So the streamlines are hyperbolas. For the velocity field, we take

dW
= 2Bz = 2B(x + iy) = u − iv, (8.136)
dz
u = 2Bx, v = −2By. (8.137)

This flow actually represents flow in a corner formed by a right angle or flow striking a flat
plate, or the impingement of two streams. For n = 2, streamlines are sketched in in Fig. 8.9.

Example 8.3
Explore kinematics and dynamics of the flow field induced by the complex potential W (z) = z 2 .

We have

W (z) = z 2, (8.138)
2
= (x + iy) , (8.139)
= (x2 − y 2 ) + 2xyi. (8.140)

Thus, we have

φ(x, y) = x2 − y 2 , ψ(x, y) = 2xy. (8.141)

© 05 December 2025, Joseph M. Powers. All rights reserved.

298 CHAPTER 8. POTENTIAL FLOW

We see that the velocity field is given by

∂φ ∂φ
u= = 2x, v= = −2y. (8.142)
∂x ∂y
We can also get it from the stream function
∂ψ ∂ψ
u= = 2x, v=− = −2y. (8.143)
∂y ∂x
We also can get the velocity field directly via
dW
= 2z = u − iv. (8.144)
dz
Thus

2(x + iy) = u − iv. (8.145)

Comparing, we see that

u = 2x, v = −2y. (8.146)

So the velocity vector is

 
2x
v= . (8.147)
−2y

The velocity vector is zero at the origin, that is a stagnation point. We also see the Laplacian equations
are satisfied because
∂2φ ∂2φ
+ 2 =2−2 = 0, (8.148)
∂x2 ∂y
2
∂ ψ ∂2ψ
+ =0+0 = 0. (8.149)
∂x2 ∂y 2
Incompressibility is satisfied as
∂u ∂v
∇·v = + = 2 − 2 = 0. (8.150)
∂x ∂y
Irrotationality is satisfied because
∂v ∂u
ωz = − = 0 − 0 = 0. (8.151)
∂x ∂y
Let us examine the deformation tensor, that is the symmetric part of the velocity gradient tensor:
  
∂u 1 ∂u
+ ∂v  
D= 
∂x
 2 ∂y ∂x  = 2 0 . (8.152)
1 ∂u ∂v ∂v 0 −2
2 ∂y + ∂x ∂y

The deformation tensor is already in diagonal form, so there is no necessity to rotate the axes to identify
the principal axes. The principal axes are the eigenvectors associated with the tensor. We can normalize
them and take them to be the unit vectors i and j. There is positive extensional strain aligned with
i. This extensional strain is exactly counterbalanced by negative extensional strain aligned with j. So

© 05 December 2025, Joseph M. Powers. All rights reserved.

8.3. ELEMENTARY COMPLEX POTENTIALS 299

the volume is preserved of the deforming fluid particle, as required by incompressibility. For the fluid
element aligned with the coordinate axes, there is no shear deformation.
The local acceleration vector of a fluid particle is given by
 ∂u ∂u ∂u   
dv ∂t + u ∂x + v ∂y 4x
a= = ∂v ∂v ∂v = . (8.153)
dt ∂t + u ∂x + v ∂y
4y

The acceleration vector is zero at the origin, and points outward from the origin when away from the
origin. A fluid particle on a stagnation streamline has no curvature, and its acceleration vector is
parallel to its velocity vector. Streamlines that are not stagnation streamlines have non-zero curvature.
The acceleration of such a fluid particle has a centripetal component that points towards the local
instantaneous center of curvature.
Dynamics tells us that acceleration vectors must be induced by net force. In our problem, in which
we neglect viscous and body forces, the only net force in play is that induced by the gradient of pressure.
Let us find the pressure field associated with this flow field. Bernoulli’s equation gives us
1
po = p + ρ(u2 + v 2 ), (8.154)
2
1
= p + ρ(4x2 + 4y 2 ), (8.155)
2
p − po = −2ρ(x2 + y 2 ). (8.156)

By inspection isobars are circles. The peak pressure is at the origin; thus the pressure decreases with
the square of the distance from the origin. Fluid accelerates from regions of high pressure to regions
of low pressure, and the acceleration vector points in the opposite direction of the pressure gradient
vector:

∇p = −4ρ (xi + yj) , (8.157)

= −ρa. (8.158)

Thus we recover our linear momenta equation

dv
ρ = −∇p. (8.159)
dt
A sketch of streamlines, equipotential lines, isobars, velocity vectors, and acceleration vectors is given
in Fig. 8.10
Let us focus on one particular streamline, that for which ψ = 2. The equation of this streamline is
1
y= . (8.160)
x
We can use standard notions from calculus to define the curvature, κ. The general formula for curvature
is
d2 y
dx2
κ=  . (8.161)
 2 3/2
dy
1 + dx

For this particular ψ = 2 streamline, we find

2
κ=
. (8.162)
1 3/2
x3 1+ x4

© 05 December 2025, Joseph M. Powers. All rights reserved.

300 CHAPTER 8. POTENTIAL FLOW

a p
1 a a
p
v p

a a
y 0
v v

v
a a
- 1

- 2
- 2 - 1 0 1 2
x

Figure 8.10: Sketch for W (z) = z 2 of streamlines, equipotential lines, isobars, velocity
vectors, acceleration vectors, pressure gradient vectors.

Taylor series reveals that

lim κ = 0, lim κ = 0. (8.163)
x→0 x→∞

This makes sense as the streamline ψ = 2 in the first quadrant approaches the x axis for large x and
the y axis for small positive x. These axes have no curvature. For x ∈ [0, +∞), it is easy to show with
calculus that κ has a maximum value given by
√
2
κmax = κ(x = 1) = . (8.164)
2
For this streamline, when x = 1, we have y = 1/x = 1/1 = 1. So the point of maximum streamline
curvature is at an angle of π/4 from the x axis.
The velocity magnitude is given by
p
||v|| = 4x2 + 4y 2 . (8.165)
For the y = 1/x streamline, this gives
r
4
||v|| = 4x2 + . (8.166)
x2
√
Calculus reveals this has a local minimum of 2 2 at the point (x, y) = (1, 1).
From this analysis, it is easy to see the following are true.
• On the x axis, with y = 0 and for x → ∞, v and a are parallel and point in positive x direction. The
fluid acceleration is parallel to the pressure gradient vector, and both are parallel to the streamwise
direction.
• On the y axis, with x = 0 and for y → ∞, v and a are parallel and point opposite directions. The
fluid acceleration is parallel to the pressure gradient vector, and both are parallel to the streamwise
direction.

© 05 December 2025, Joseph M. Powers. All rights reserved.

8.3. ELEMENTARY COMPLEX POTENTIALS 301

4
a
y

v a
v
y 2
p
v v v
a
1
v a
a
v x
0 v
0 1 2 3 4
a) x b)

Figure 8.11: Kinematics and dynamics on the streamline y = 1/x for the flow field defined
by the complex potential W (z) = z 2 : a) planar rendering, b) three-dimensional rendering.

• For points on the curve θ = π/4, v is orthogonal to a. The fluid acceleration is parallel to the pressure
gradient vector, and both are in the stream-normal direction. Here the acceleration is all centripetal
and due to streamline curvature. And at such points the velocity magnitude has a local minimum.

• For intermediate points, v is neither parallel nor orthogonal to a. The acceleration is parallel to the
pressure gradient vector, and there are non-zero components of acceleration in the streamwise and
stream-normal directions.

We can get the streamlines by direct integration. Because we know the velocity field, u = 2x,
v = −2y, we then have the system of differential equations for the streamlines, pathlines, and streaklines
for this steady flow. Let us find the streamline that passes through (x, y)T = (1, 1)T . We have the
kinematics cast as a dynamical system as discussed in Ch. 3.13.

dx
= 2x, x(0) = 1, (8.167)
dt
dy
= −2y, y(0) = 1. (8.168)
dt
Integrating, we get

x(t) = e2t , (8.169)

−2t
y(t) = e . (8.170)

This is a parametric solution for the streamline, streakline, and pathline that passes through (1, 1)T at
t = 0. Obviously 1/x = e−2t = y, so this streamline is given, as expected, by

1
y= . (8.171)
x
The streamline y = 1/x along with the local velocity and acceleration vectors and a few pressure
contours are plotted in Fig. 8.10. The unit tangent to the streamline is
2
!
v 2xi − 2yj 2xi − x2 j x2 1 √x
1+x 4
αt = = p q = √ i− √ j= 1 (8.172)
||v|| 4x2 + 4y 2 4x2 + 42 1 + x4 1 + x4 − √1+x 4
x

© 05 December 2025, Joseph M. Powers. All rights reserved.

302 CHAPTER 8. POTENTIAL FLOW

We can specialize Eq. (3.151) to find the stretching rate Dt in the streamwise direction along the
streamline y = 1/x. This is
!
  
√x
2
x4 − 1
T x 2
1 2 0 1+x 4
Dt = αt · D · αt = √1+x4 − √1+x4 1 = 2 . (8.173)
0 −2 − √1+x 4
x4 + 1

There is no streamwise stretching at x = 1. As x → ∞, Dt → 2, and as x → 0, Dt → −2.

We see by inspection that the unit vector normal to the streamline αn must be
1
!
√
1+x4
αn = 2 . (8.174)
√x
1+x4

This vector obviously has αTn · αt = 0, and αTn · αn = 1. Moreover αn points toward the center of
curvature of the streamline. We can also find the stretching rate Dn in the stream-normal direction
along the streamline y = 1/x. This is
   √1 !
T 1 x 2 2 0 1+x4 1 − x4
Dn = αn · D · αn = √1+x4 √1+x4 x 2 =2 4 . (8.175)
0 −2 √
4
x +1
1+x

There is no stream-normal stretching at x = 1. As x → ∞, Dn → −2, and as x → 0, Dn → 2. Moreover,

because the flow is incompressible and thus volume-preserving, streamwise stretching is balanced by
stream-normal stretching to that overall one has
x4 − 1 1 − x4
Dt + Dn = 2 4
+2 4 = 0. (8.176)
x +1 x +1
Now let us perform an analysis for extreme values for D similar to that performed in generating
Fig. 3.4. For a general direction α, we have
  
T 2 0 α1
D = α · D · α = ( α1 α2 ) = 2α21 − 2α22 . (8.177)
0 −2 α2

As before, we could plot contours for which D is constant in the (α1 , α2 ) plane and get an infinite family
of curves. In contrast to the ellipses of Fig. 3.4, here we have hyperbolas as contours. Once more, we
also have the constraint of α21 + α22 = 1. And once more, the eigenvalues of D are the special contours
of D, suggesting we examine the two special contours

2 = 2α21 − 2α22 , (8.178)

−2 = 2α21 − 2α22 . (8.179)

These two curves, along with the unit circle α21 + α22 = 1 are plotted in Fig. 8.12 An infinite family of
contours of D exist. Many of them will also intersect the unit circle, and so are candidate solutions.
However, the special contours we selected are extreme values. For intersection with the unit circle, we
require

D ∈ [λmin , λmax ], (8.180)

∈ [−2, 2]. (8.181)

Because D is already diagonal, the eigenvectors are aligned with the unrotated coordinate axes. It is
straightforward to show for D = −2, that α = (0, 1)T and for D = 2, that α = (1, 0)T .

© 05 December 2025, Joseph M. Powers. All rights reserved.

8.3. ELEMENTARY COMPLEX POTENTIALS 303

1.5

1.0

0.5

0.0

- 0.5

- 1.0

- 1.5
- 1.5 - 1.0 - 0.5 0.0 0.5 1.0 1.5

Figure 8.12: Two special contours of D along with the unit circle illustrating the extreme
values of D.

iy

Q −Q
x

source sink

Figure 8.13: Source sink pair.

8.3.6 Doublets
We can form what is known as a doublet flow by considering the superposition of a source
and sink and let the two approach each other. Consider a source and sink of equal and
opposite strength straddling the y axis, each separated from the origin by a distance ǫ as
sketched in Fig. 8.13. The complex velocity potential is

Q Q
W (z) = ln(z + ǫ) − ln(z − ǫ), (8.182)
2π   2π
Q z+ǫ
= ln . (8.183)
2π z−ǫ

It can be shown by synthetic division that as ǫ → 0, that

z+ǫ 2 2
= 1 + ǫ + ǫ2 2 + . . . . (8.184)
z−ǫ z z

© 05 December 2025, Joseph M. Powers. All rights reserved.

304 CHAPTER 8. POTENTIAL FLOW

So the potential approaches

 
Q 2 2 2
W (z) ∼ ln 1 + ǫ + ǫ 2 + . . . . (8.185)
2π z z

Now because ln(1 + x) → x as x → 0, we get for small ǫ that

Q 2 Qǫ
W (z) ∼ ǫ ∼ . (8.186)
2π z πz
Now if we require that
Qǫ
lim → µ, (8.187)
ǫ→0 π

we have
µ µ x − iy µ(x − iy)
W (z) = = = 2 . (8.188)
z x + iy x − iy x + y2
So
x y
φ(x, y) = µ , ψ(x, y) = −µ . (8.189)
x2 + y2 x2 + y2
In polar coordinates, we then say

cos θ sin θ
φ=µ , ψ = −µ . (8.190)
r r
Streamlines and equipotential lines for a doublet are plotted in Fig. 8.14.

8.3.7 Rankine half body

Now consider the superposition of a uniform stream and a source, that we define to be a
Rankine half body:

Q
W (z) = Uz + ln z, with U, Q ∈ R1 , (8.191)
2π
iθ Q
= Ure + (ln r + iθ), (8.192)
2π
Q
= Ur(cos θ + i sin θ) + (ln r + iθ), (8.193)
 2π  
Q Q
= Ur cos θ + ln r + i Ur sin θ + θ . (8.194)
2π 2π

So
Q Q
φ = Ur cos θ + ln r, ψ = Ur sin θ + θ. (8.195)
2π 2π
Streamlines for a Rankine half body are plotted in Fig. 8.15. Now for the Rankine half body,

© 05 December 2025, Joseph M. Powers. All rights reserved.

8.3. ELEMENTARY COMPLEX POTENTIALS 305

iy
4

0 x

−2

−4
−4 −2 0 2 4

Figure 8.14: Streamlines and equipotential lines for a doublet. Notice because the sink is
infinitesimally to the right of the source, there exists a directionality. This can be considered
a type of dipole moment; in this case, the direction of the dipole is −i.

iy
4

0 x

−2

−4
−4 −2 0 2 4

Figure 8.15: Streamlines for a Rankine half body.

© 05 December 2025, Joseph M. Powers. All rights reserved.

306 CHAPTER 8. POTENTIAL FLOW

it is clear that there is a stagnation point somewhere on the x axis, along θ = π. With the
velocity given by
dW Q
=U+ = u − iv, (8.196)
dz 2πz
we get

Q 1 −iθ
U+ e = u − iv, (8.197)
2π r
Q1
U+ (cos θ − i sin θ) = u − iv, (8.198)
2π r
Q Q
u=U+ cos θ, v = sin θ. (8.199)
2πr 2πr
When θ = π, we get u = 0 when;

Q
0 = U+ (−1), (8.200)
2πr
Q
r = . (8.201)
2πU

8.3.8 Flow over a cylinder

We can model flow past a cylinder without circulation by superposing a uniform flow with
a doublet. Defining a2 = µ/U, we write
 
µ a2
W (z) = Uz + = U z + , (8.202)
z z
 
iθ a2
= U re + iθ , (8.203)
re
 
a2
= U r(cos θ + i sin θ) + (cos θ − i sin θ) , (8.204)
r
 2
  
a a2
= U r cos θ + cos θ + i r sin θ − sin θ , (8.205)
r r
    
a2 a2
= Ur cos θ 1 + 2 + i sin θ 1 − 2 . (8.206)
r r

So    
a2 a2
φ = Ur cos θ 1 + 2 , ψ = Ur sin θ 1 − 2 . (8.207)
r r
Now on r = a, we have ψ = 0. Because the stream function is constant here, the curve r = a,
a circle, must be a streamline through which no mass can pass. A sketch of the streamlines
and equipotential lines is plotted in Fig. 8.16.

© 05 December 2025, Joseph M. Powers. All rights reserved.

8.3. ELEMENTARY COMPLEX POTENTIALS 307

iy
2

0 x

−1

−2
−2 −1 0 1 2

Figure 8.16: Streamlines and equipotential lines for flow over a cylinder without circulation.

For the velocities, we have

   
∂φ a2 a2
vr = = U cos θ 1 + 2 + Ur cos θ −2 3 , (8.208)
∂r r r
 2

a
= U cos θ 1 − 2 , (8.209)
r
 
1 ∂φ a2
vθ = = −U sin θ 1 + 2 . (8.210)
r ∂θ r
So on r = a, we have vr = 0, and vθ = −2U sin θ. Thus on the surface, we have
∇T φ · ∇φ = 4U 2 sin2 θ. (8.211)
Bernoulli’s equation, Eq. (6.211), for a steady flow with p → p∞ as r → ∞ then gives
p 1 T p∞ U 2
+ ∇ φ · ∇φ = + , (8.212)
ρ 2 ρ 2
1
p = p∞ + ρU 2 (1 − 4 sin2 θ). (8.213)
2
The pressure coefficient Cp , defined below, then is
p − p∞
Cp ≡ 1 2
= 1 − 4 sin2 θ. (8.214)
2
ρU
A sketch of the pressure distribution, both predicted and experimentally observed, is plot-
ted in Fig. 8.17. We note that the potential theory predicts the pressure well on the front

© 05 December 2025, Joseph M. Powers. All rights reserved.

308 CHAPTER 8. POTENTIAL FLOW

1 2 3
experiment
-1
potential
-2 theory

-3

pressure
distribution
on cylinder
surface from
potential theory

Figure 8.17: Pressure distribution for ideal flow over a cylinder without circulation.

© 05 December 2025, Joseph M. Powers. All rights reserved.

8.3. ELEMENTARY COMPLEX POTENTIALS 309

surface of the cylinder, but not so well on the back surface. This is because in most real
fluids, a phenomenon known as flow separation manifests itself in regions of negative pres-
sure gradients. Correct modeling of separation events requires a re-introduction of viscous
stresses. A potential theory cannot predict separation.

Example 8.4
For a cylinder of radius c at rest in an accelerating potential flow field with a far field velocity of
U = a + bt, find the pressure on the stagnation point of the cylinder.

The velocity potential and velocities for this flow are

 
c2
φ(r, θ, t) = (a + bt)r cos θ 1 + 2 , (8.215)
r
 
∂φ c2
vr = = (a + bt) cos θ 1 − 2 , (8.216)
∂r r
 
1 ∂φ c2
vθ = = −(a + bt) sin θ 1 + 2 , (8.217)
r ∂θ r
 
2 2
 2 !
1 T 1 2 2 c 2 c2
∇ φ · ∇φ = (a + bt) cos θ 1 − 2 + sin θ 1 + 2 , (8.218)
2 2 r r
 
1 2 c4 2c2 2 2


= (a + bt) 1 + 4 + 2 sin θ − cos θ . (8.219)
2 r r

Also, because the flow is unsteady, we will need ∂φ/∂t:

 
∂φ c2
= br cos θ 1 + 2 . (8.220)
∂t r

Now we note in the limit as r → ∞ that

∂φ 1 T 1
→ br cos θ, ∇ φ · ∇φ → (a + bt)2 . (8.221)
∂t 2 2
We also note that on the surface of the cylinder

vr (r = c, θ, t) = 0. (8.222)

The unsteady Bernoulli equation, the incompressible, zero-body force version of Eq. (6.196), gives us
∂φ 1 T p
+ ∇ φ · ∇φ + = f (t). (8.223)
∂t 2 ρ
We use the far field behavior to evaluate f (t):
1 p
br cos θ + (a + bt)2 + = f (t). (8.224)
2 ρ

Now if we make the non-intuitive choice of f (t) = 12 (a + bt)2 + po /ρ, we get

1 p 1 po
br cos θ + (a + bt)2 + = (a + bt)2 + . (8.225)
2 ρ 2 ρ

© 05 December 2025, Joseph M. Powers. All rights reserved.

310 CHAPTER 8. POTENTIAL FLOW

So
p = po − ρbr cos θ = po − ρbx. (8.226)
Because the flow at infinity is accelerating, there must be a far-field pressure gradient to induce this
acceleration. Consider the x momentum equation in the far field
du ∂p
ρ = − , (8.227)
dt ∂x
ρ(b) = −(−ρb). (8.228)
So for the pressure field, we have
   
c2 1 c4 2c2 p 1 po
2 2
br cos θ 1 + 2 + (a + bt) 1 + 4 + 2 (sin θ − cos θ) + = (a + bt)2 + ,
2
(8.229)
r 2 r r ρ 2 ρ
| {z } | {z } | {z }
∂φ/∂t ∇T φ·∇φ/2 f (t)

that can be solved for p to yield

   4 
c2 1 2 c 2c2 2 2
p(r, θ, t) = po − ρbr cos θ 1 + 2 − ρ(a + bt) + 2 (sin θ − cos θ) . (8.230)
r 2 r4 r
For the stagnation point, we evaluate as
1
p(c, π, t) = po − ρbc(−1) (1 + 1) − ρ(a + bt)2 (1 + 2(1)(0 − 1)) , (8.231)
2
1
= po + ρ(a + bt)2 + 2ρbc, (8.232)
2
1 2
= po + ρU + 2ρbc. (8.233)
2
The first two terms would be predicted by a naı̈ve extension of the steady Bernoulli’s equation,
Eq. (6.211). The final term however is not intuitive and is a purely unsteady effect.

The general method of defining a complex potential W (z) can be used to easily generate
some remarkable plots of harmonic functions. We give a few in Fig. 8.18.

8.4 Forces induced by potential flow

There are more basic ways to describe the force on bodies using complex variables directly.
We shall give those methods, but first a discussion of the motivating complex variable theory
is necessary.

8.4.1 Contour integrals

Consider the closed contour integral of a complex function in the complex plane. For such
integrals, we have a useful theory that we will not prove, but will demonstrate here. Consider
contour integrals enclosing the origin with a circle in the complex plane for four functions.
The contour in each is
C : z = R̂eiθ , θ ∈ [0, 2π). (8.234)

© 05 December 2025, Joseph M. Powers. All rights reserved.

8.4. FORCES INDUCED BY POTENTIAL FLOW 311

2 2 2

1 1 1

0 0 0

-1 -1 -1

-2 -2 -2
-2 -1 0 1 2
a) b)- 2 -1 0 1 2
c)- 2 -1 0 1 2

2 2 2

1 1 1

0 0 0

-1 -1 -1

-2 -2 -2
-2 -1 0 1 2
d) -2 -1 0 1 2
e) - 2 -1 0 1 2
f)2
2 2

1 1 1

0 0 0

-1 -1 -1

-2 -2 -2
-2 -1 0 1 2 -2 -1 0 1 2
g) h)
-2 -1 0 1 2
i)

Figure 8.18: Harmonic functions: streamlines and equipotential curves for a) W (z) = z, b)
W (z) = z 3 , c) W (z) = z + 1/z, d) W (z) = z 2 + 1/z 2 , e) W (z) = z 6 + z 2 , f) W (z) = sin z,
g) W (z) = sin z 2 , h) W (z) = 1/(z + 1) − 1/(z − 1), i) W (z) = (1 + i)z.

© 05 December 2025, Joseph M. Powers. All rights reserved.

312 CHAPTER 8. POTENTIAL FLOW

For such a contour,

dz = iR̂eiθ dθ. (8.235)

[Link] Simple pole

We describe a simple pole with the complex potential
a
W (z) = . (8.236)
z
and the contour integral is
I I Z θ=2π
a a
W (z) dz = dz = iR̂eiθ dθ, (8.237)
C C z θ=0 R̂e iθ
Z 2π
= ai dθ = 2πia. (8.238)
0

[Link] Constant potential

We describe a constant with the complex potential
W (z) = b. (8.239)
and the contour integral is
I I Z θ=2π
W (z) dz = b dz = biR̂eiθ dθ, (8.240)
C C θ=0
2π
biR̂ iθ
= e = 0, (8.241)
i
0

because e0i = e2πi = 1.

[Link] Uniform flow

We describe a constant with the complex potential
W (z) = cz. (8.242)
and the contour integral is
I I Z θ=2π
W (z) dz = cz dz = cR̂eiθ iR̂eiθ dθ, (8.243)
C C θ=0
Z 2π
2π
icR̂2 2iθ
= icR̂2 e2iθ dθ = e = 0. (8.244)
0 2i
0
0i 4πi
because e = e = 1.

© 05 December 2025, Joseph M. Powers. All rights reserved.

8.4. FORCES INDUCED BY POTENTIAL FLOW 313

[Link] Quadrupole
A quadrupole potential is described by
k
W (z) = . (8.245)
z2
Taking the contour integral, we find
I Z 2π
k iR̂eiθ
2
dz = k dθ, (8.246)
C z 0 R̂2 e2iθ
Z 2π
ki 2π −iθ ki 1 −iθ
= e dθ = e = 0. (8.247)
R̂ 0 R̂ −i 0

So the only non-zero contour integral is for functions of the form W (z) = a/z. We find all
polynomial powers of z have a zero contour integral about the origin for arbitrary contours
except this special one.

8.4.2 Laurent series

Now it can be shown that any function can be expanded, much as for a Taylor series, as a
Laurent series:4

W (z) = . . . + C−2 (z − zo )−2 + C−1 (z − zo )−1 + C0 (z − zo )0 + C1 (z − zo )1 + C2 (z − zo )2 + . . . .

(8.248)
In compact summation notation, we can say
n=∞
X
W (z) = Cn (z − zo )n . (8.249)
n=−∞

Taking the contour integral of both sides we get

I I n=∞
X
W (z) dz = Cn (z − zo )n dz, (8.250)
C C n=−∞
n=∞
X I
= Cn (z − zo )n dz. (8.251)
n=−∞ C

From our just completed analysis, this has value 2πi only when n = −1, so
I
W (z) dz = C−1 2πi. (8.252)
C
4
Pierre Alphonse Laurent, 1813-1854, Parisian engineer who worked on port expansion in Le Harve,
submitted his work on Laurent series for a Grand Prize in 1842, with the recommendation of Cauchy, but
was rejected because of a late submission.

© 05 December 2025, Joseph M. Powers. All rights reserved.

314 CHAPTER 8. POTENTIAL FLOW

Here C−1 is known as the residue of the Laurent series. In general we have the Cauchy
integral theorem that holds that if W (z) is analytic within and on a closed curve C except
for a finite number of singular points, then
I X
W (z) dz = 2πi residues. (8.253)
C

Let us get a simple formula for Cn . We first exchange m for n in Eq. (8.249) and say
m=∞
X
W (z) = Cm (z − zo )m . (8.254)
m=−∞

Then we operate as follows:

m=∞
X
W (z)
= Cm (z − zo )m−n−1 , (8.255)
(z − zo )n+1 m=−∞
I I m=∞
X
W (z)
dz = Cm (z − zo )m−n−1 dz, (8.256)
C (z − zo )n+1 C m=−∞
m=∞
X I
= Cm (z − zo )m−n−1 dz. (8.257)
m=−∞ C

Here C is any closed contour that has zo in its interior. The contour integral on the right
side only has a non-zero value when n = m. Let us then insist that n = m, giving
I I
W (z)
n+1
dz = Cn (z − zo )−1 dz . (8.258)
C (z − zo ) C
| {z }
=2πi

We know from earlier analysis that the contour integral enclosing a simple pole such as found
on the right side has a value of 2πi. Solving, we find then that
I
1 W (z)
Cn = dz. (8.259)
2πi C (z − zo )n+1
If the closed contour C encloses no poles, then
I
W (z) dz = 0. (8.260)
C

8.4.3 Pressure distribution for steady flow

For steady, irrotational, incompressible flow with no body force present, we have the Bernoulli
equation. We recast Eq. (6.211) as
p 1 T p∞ 1 2
+ ∇ φ · ∇φ = + U∞ . (8.261)
ρ 2 ρ 2

© 05 December 2025, Joseph M. Powers. All rights reserved.

8.4. FORCES INDUCED BY POTENTIAL FLOW 315

We can write this in terms of the complex potential in a simple fashion. First, recall that

∇T φ · ∇φ = u2 + v 2 . (8.262)

We also have dW/dz = u − iv, so dW/dz = u + iv. Consequently,

dW dW
= u2 + v 2 = ∇T φ · ∇φ. (8.263)
dz dz
So we get the pressure field from Bernoulli’s equation to be
 
1 2 dW dW
p = p∞ + ρ U∞ − . (8.264)
2 dz dz

The pressure coefficient Cp is

p − p∞ 1 dW dW
Cp = 1 2
= 1 − 2 dz dz
. (8.265)
2
ρU∞ U∞

8.4.4 Blasius force theorem

For steady flows, we can find the net contribution of a pressure force on an arbitrary shaped
solid body with the Blasius5 force theorem. Consider the geometry sketched in Fig. 8.19.
The surface of the arbitrarily shaped body is described by Ab , and C is a closed contour
containing Ab . First consider the linear momenta equation for steady flow, no body forces,
and no viscous forces,


ρ vT · ∇ v = −∇p, add mass to get conservative form, (8.266)

T
∇T · (ρvvT ) = −∇p, integrate over V , (8.267)
Z Z

T
∇T · (ρvvT ) dV = − ∇p dV, use Gauss, (8.268)
V Z Z V
T
ρv(v · n) dA = − pn dA. (8.269)
A A

Now the surface integral here is really a line integral with unit depth b, dA = b ds. Moreover
the surface enclosing the fluid has an inner contour Ab and an outer contour C. Now on C,
that we prescribe, we will know x(s) and y(s), where s is arc length. So on C we also get
the unit tangent α and unit outward normal n:
 dx   dy 
α = dy ,ds n= ds , on C. (8.270)
dx
ds
− ds
5
Paul Richard Heinrich Blasius, 1883-1970, student of Ludwig Prandtl and long time teacher at the tech-
nical college of Hamburg whose 1907 Ph.D. thesis gave mathematical description of similarity solution to
the boundary layer problem.

© 05 December 2025, Joseph M. Powers. All rights reserved.

316 CHAPTER 8. POTENTIAL FLOW

iy

Figure 8.19: Potential flow about arbitrarily shaped two-dimensional body with fluid control
volume indicated.

A loose analysis that can be verified more rigorously shows

p α and n are p
unit vectors. Stan-
dard geometry tells us ds = dx + dy . So α = (dx/ dx + dy , dy/ dx2 + dy 2 )T . By
2 2 2 2 2

inspection ||α|| = 1. A similar result holds for n.

On Ab we have, because it is a solid surface

vT · n = 0, on Ab . (8.271)

Now let the force on the body due to fluid pressure be F:

Z
pn dA = F. (8.272)
Ab

Now return to our linear momentum equation

Z Z
T
ρvv · n dA = − pn dA. (8.273)
A A

Break this up to get

I I I I
T T
ρv |v {z· n} dA + ρvv · n dA = − pn dA − pn dA, (8.274)
Ab C A C
=0 | {z }
b

I =IF
ρvvT · n dA = −F − pn dA. (8.275)
C C

© 05 December 2025, Joseph M. Powers. All rights reserved.

8.4. FORCES INDUCED BY POTENTIAL FLOW 317

We can break this into x and y components:

I   I
dy dx dy
ρu u − v b ds = −Fx − p b ds, (8.276)
C ds ds ds
C |{z}
| {z }
nx
vT ·n
I   I  
dy dx dx
ρv u − v b ds = −Fy − p − b ds. (8.277)
C ds ds C ds
| {z } | {z }
vT ·n ny

Solving for Fx and Fy per unit depth, we get

I
Fx
= −p dy − ρu2 dy + ρuv dx, (8.278)
b
IC
Fy
= p dx + ρv 2 dx − ρuv dy. (8.279)
b C

Now Bernoulli gives us p = po − (1/2)ρ(u2 + v 2 ), where po is some constant. So the x force

per unit depth becomes
I
Fx 1
= −po dy + ρ(u2 + v 2 ) dy − ρu2 dy + ρuv dx. (8.280)
b C 2

Because the integral over a closed contour of a constant po is zero, we get

I
Fx 1
= ρ(−u2 + v 2 ) dy + ρuv dx, (8.281)
b C 2
I
1
= ρ (−u2 + v 2 ) dy + 2uv dx. (8.282)
2 C

Similarly for the y direction, we get

I
Fy 1
= po dx − ρ(u2 + v 2 ) dx + ρv 2 dx − ρuv dy, (8.283)
b C I 2
1
= ρ (−u2 + v 2 ) dx − 2uv dy. (8.284)
2 C

Now consider the linear combination (Fx − iFy )/b:

I
Fx − iFy 1
= ρ (−u2 + v 2 ) dy + 2uv dx − (−u2 + v 2 )i dx + 2uvi dy, (8.285)
b 2 C
I
1
= ρ (i(u2 − v 2 ) + 2uv) dx + ((−u2 + v 2 ) + 2uvi) dy, (8.286)
2 C
I
1
= ρ (i(u2 − v 2 ) + 2uv) dx + (i(u2 − v 2 ) + 2uv)i dy, (8.287)
2 C

© 05 December 2025, Joseph M. Powers. All rights reserved.

318 CHAPTER 8. POTENTIAL FLOW

iy

Q
x

Figure 8.20: Potential flow about arbitrarily shaped two-dimensional body with distribution
of sources, sinks, vortices, and dipoles.
I
1
= ρ (i(u2 − v 2 ) + 2uv)(dx + i dy), (8.288)
2 C
I
1
= ρ i(u − iv)2 (dx + i dy), (8.289)
2 C
I  2
1 dW
= ρi dz. (8.290)
2 C dz

So if we have the complex potential, we can easily get the force on a body.

8.4.5 Kutta-Zhukovsky lift theorem

Consider the geometry sketched in Fig. 8.20. Here we consider a flow with a freestream
constant velocity of Uo . We take an arbitrary body shape to enclose a distribution of cancel-
ing source sink pairs, doublets, point vortices, quadruples, and any other non-mass adding
potential flow term. This combination gives rise to some surface that is a streamline.
Now far from the body surface a contour sees all of these features as effectively concen-
trated at the origin. Then, the potential can be written as

Q Q iΓ µ
W (z) ∼ Uz
|{z} + ln z − ln z + ln z + +... (8.291)
uniform flow |2π {z 2π } |2π{z } z
|{z}
canceling source sink pair clockwise! vortex doublet

The sign convention for Γ has been violated here, by tradition. Now let us take D to be the
so-called drag force per unit depth and L to be the so-called lift force per unit depth, so in

© 05 December 2025, Joseph M. Powers. All rights reserved.

8.4. FORCES INDUCED BY POTENTIAL FLOW 319

terms of Fx and Fy , we have

Fx Fy
= D, = L. (8.292)
b b
Now by the Blasius force theorem, we have
I  2
1 dW
D − iL = ρi dz, (8.293)
2 C dz
I  2
1 iΓ µ
= ρi U+ − + . . . dz, (8.294)
2 C 2πz z 2
I   2  
1 2 iΓU 1 Γ
= ρi U + − 2 + 2Uµ + . . . dz. (8.295)
2 C πz z 4π 2
P
Now the Cauchy integral theorem, Eq. (8.253), gives us the contour integral to be 2πi residues.
Here the residue is iΓU/π. So we get
  
1 iΓU
D − iL = ρi 2πi , (8.296)
2 π
= −iρΓU. (8.297)

We see that

D = 0, (8.298)
L = ρUΓ. (8.299)

This is a remarkably simple and elegant result! Note that

• Γ is associated with clockwise circulation here. This is something of a tradition in

aerodynamics.

• Because for airfoils Γ ∼ U, we get the lift force L ∼ ρU 2 ,

• For steady inviscid flow, there is no drag. Consideration of either unsteady or viscous
effects would lead to a non-zero x component of force.

Example 8.5
Consider the flow over a cylinder of radius a with clockwise circulation Γ.

To do so, we can superpose a point vortex onto the potential for flow over a cylinder in the following
fashion:
 
a2 iΓ  z 
W (z) = U z + + ln . (8.300)
z 2π a

© 05 December 2025, Joseph M. Powers. All rights reserved.

320 CHAPTER 8. POTENTIAL FLOW

Breaking this up as before into real and complex parts, we get

     
a2 a2 iΓ   r  
W (z) = U r cos θ 1 + 2 + i U r sin θ 1 − 2 + ln + iθ . (8.301)
r r 2π a
So, we find
  r
a2 Γ
ψ = ℑ(W (z)) = U r sin θ 1 − 2 + ln . (8.302)
r 2π a
On r = a, we find that ψ = 0, so the addition of the circulation in the way we have proposed maintains
the cylinder surface to be a streamline. It is important to note that this is valid for arbitrary Γ. That
is the potential flow solution for flow over a cylinder is non-unique. In aerodynamics, this is used to
advantage to add just enough circulation to enforce the so-called Kutta condition.6 The Kutta condition
is an experimentally observed fact that for a steady flow, the trailing edge of an airfoil is a stagnation
point.
The Kutta-Zhukovsky7 lift theorem tells us whenever we add circulation, that a lift force L = ρU Γ
is induced. This is consistent with the phenomena observed in baseball that the “fastball” rises. The
fastball leaves the pitcher’s hand traveling towards the batter and rotating towards the pitcher. The
induced aerodynamic force is opposite to the force of gravity.
Let us get the lift force the hard way and verify the Kutta-Zhukovsky theorem. We can easily get
the velocity field from the velocity potential:
 
a2 Γθ
φ = ℜ(W (z)) = U r cos θ 1 + 2 − . (8.303)
r 2π
Thus, we differentiate φ appropriately to find vr and vθ :
   
∂φ 2a2 a2
vr = = U r cos θ − 3 + U cos θ 1 + 2 , (8.304)
∂r r r
 3 2

2a a
vr |r=a = U cos θ − 3 + 1 + 2 , (8.305)
a a
= 0, (8.306)
   
1 ∂φ 1 a2 Γ
vθ = = −U r sin θ 1 + 2 − , (8.307)
r ∂θ r r 2π
 
a2 Γ
vθ |r=a = −U sin θ 1 + 2 − , (8.308)
a 2πa
Γ
= −2U sin θ − . (8.309)
2πa
We get the pressure on the cylinder surface from Bernoulli’s equation:
1 1 T
p = p∞ + ρU 2 − ρ∇ φ · ∇φ, (8.310)
2 2
 2
1 1 Γ
= p∞ + ρU 2 − ρ −2U sin θ − . (8.311)
2 2 2πa
Now for a small element of the cylinder at r = a, the surface area is dA = br dθ = ba dθ. This is
sketched in Fig. 8.21. We also note that the x and y forces depend on the orientation of the element,
6
Martin Wilhelm Kutta, 1867-1944, Silesian-born German mechanician, studied at Breslau, taught
mainly at Stuttgart, co-developer of Runge-Kutta method for integrating ordinary differential equations.
7
Nikolai Egorovich Zhukovsky, 1847-1921, Russian applied mathematician and mechanician, father of
Russian aviation, purchased glider from Lilienthal, developed lift theorem independently of Kutta, organized
Central Aerohydrodynamic Institute in 1918.

© 05 December 2025, Joseph M. Powers. All rights reserved.

8.5. HEAT TRANSFER ANALOG 321

ce
for
iy

re
ssu
pre al
loc
x

Figure 8.21: Pressure force on a differential area element of cylindrical surface.

given by θ. Elementary trigonometry shows that the elemental x and y forces per depth are
dFx
= −p(cos θ)a dθ, (8.312)
b
dFy
= −p(sin θ)a dθ. (8.313)
b
So integrating over the entire cylinder, we obtain,
Z 2π  2 !
Fx 1 2 1 Γ
= − p∞ + ρU − ρ −2U sin θ − (cos θ)a dθ, (8.314)
b 0 2 2 2πa
| {z }
p
Z 2π  2 !
Fy 1 2 1 Γ
= − p∞ + ρU − ρ −2U sin θ − (sin θ)a dθ. (8.315)
b 0 2 2 2πa
| {z }
p

Integration via computer algebra gives

Fx
= 0, (8.316)
b
Fy
= ρU Γ. (8.317)
b
This is identical to the result we expect from the Kutta-Zhukovsky lift theorem, Eq. (8.299).

8.5 Heat transfer analog

Let us next analyze in a similar fashion Laplace’s equation for steady state two-dimensional
heat transfer:
∂2T ∂2T
+ = 0. (8.318)
∂x2 ∂y 2

© 05 December 2025, Joseph M. Powers. All rights reserved.

322 CHAPTER 8. POTENTIAL FLOW

Here the independent variables are x and y and the dependent variable is the temperature
T.

8.5.1 Characteristics
Now our Laplace’s equation arose from the two-dimensional time-independent form of the
energy equation for v = 0, that is

∂e
ρ = −∇T · q, (8.319)
∂t
|{z}
=0
T
∇ · q = 0, (8.320)
∂qx ∂qy
+ = 0. (8.321)
∂x ∂y

This is our first first-order partial differential equation. To aid this analysis, let us recall
that
∂T ∂T
qx = −k , qy = −k . (8.322)
∂x ∂y

We then see that

∂qx ∂2T ∂qy ∂2T
= −k , = −k . (8.323)
∂y ∂y∂x ∂x ∂x∂y

Equating the mixed second partial derivatives, we get our second first-order partial differen-
tial equation:

∂qx ∂qy
= . (8.324)
∂y ∂x

Equations (8.321) and (8.324) form the system

∂qx ∂qy
+ = 0, (8.325)
∂x ∂y
∂qy ∂qx
− = 0. (8.326)
∂x ∂y

Equations (8.325) and (8.326) can be recast as

   ∂qx     ∂qx   
1 0 ∂x 0 1 ∂y 0
∂qy + ∂qy = . (8.327)
0 1 ∂x
−1 0 ∂y
0
| {z } | {z } | {z } | {z } | {z }
A ∂u B ∂u c
∂x ∂y

© 05 December 2025, Joseph M. Powers. All rights reserved.

8.5. HEAT TRANSFER ANALOG 323

Here our vector u is

 
qx
u= . (8.328)
qy

The associated eigenvalue problem, described in more detail in Sec. 9.6, is

λ −1
det (λA − B) = = 0. (8.329)
1 λ

Solving gives

λ2 + 1 = 0, (8.330)
λ = ±i. (8.331)

The eigenvalues are distinct but not real. Presence of complex eigenvalues indicates the
equation cannot be written in characteristic form, and that finite speed signaling phenomena
are not present in the solution. Because its eigenvalues are imaginary, Laplace’s equation is
elliptic.

8.5.2 Solutions
General solutions for Laplace’s equation may be considered, and they are straightforward if
one is willing to consider complex variables. Generalizing the real T by the complex T, we
can rewrite Laplace’s equation, Eq. (8.318), as

∂2T ∂2T
=− 2. (8.332)
∂y 2 ∂x

We think of T as a complex variable with a real part that is T , that is ℜ(T) = T . Eq. (8.332)
is fully equivalent to the wave equation if we let y be represented as T, t be represented
√ as
2
y, and the square of the wave speed a = −1. With the complex number i = −1, we can
recast Laplace’s equation as

∂2T 2
2∂ T
= i . (8.333)
∂y 2 ∂x2

We thus have the d’Alembert solution with the wave speed a being the imaginary number i:

T(x, y) = f (x + iy) + g(x − iy). (8.334)

Here f and g are arbitrary functions. This generality is remarkable! On lines on which

dy
= λ = ±i, (8.335)
dx

© 05 December 2025, Joseph M. Powers. All rights reserved.

324 CHAPTER 8. POTENTIAL FLOW

we can expect T(x, y) to be a constant. This will be true then on lines such that

y = ±ix + C, (8.336)
±iy = −x ± iC, (8.337)
x ± iy = Ĉ, with Ĉ = ±iC. (8.338)

This induces T(x, y) = f (x + iy) + g(x − iy).

Let us show this satisfies Laplace’s equation. We have the following partial derivatives:
∂T ∂2T ∂T ∂2T
= f ′ + g′, = f ′′ + g ′′, = if ′ − ig ′ , = −f ′′ − g ′′ . (8.339)
∂x ∂x2 ∂y ∂y 2
Obviously Laplace’s equation is satisfied for arbitrary f and g:
∂2T ∂2T
+ 2 = f ′′ + g ′′ − f ′′ − g ′′ = 0. (8.340)
∂x2 ∂y
With a complex variable:

z = x + iy, x ∈ R1 , y ∈ R1 , z ∈ C1 , (8.341)

and z = x − iy, the d’Alembert solution to Laplace’s equation is

T(z) = f (z) + g(z). (8.342)

Thus an arbitrary function of either z or z satisfies Laplace’s equation. We will see that
the curves on which the real part of T is constant are isotherms and curves on which the
imaginary part of T is constant are adiabats. An adiabat is defined as a curve on which
no thermal energy can cross, and so it is tangent to the heat flux vector q. Also because
q = −k∇T ,
• adiabats are orthogonal to isotherms.

Example 8.6
Examine the solution of Laplace’s equation

T(z) = z 2 . (8.343)

Consider the associated heat flux vector for k = 1, q = −∇T.

Let us expand and examine our proposed solution. We have T(z) = z 2 , so

T(x, y) = (x + iy)2 . (8.344)

Expanding, we get

T(x, y) = x2 − y 2 + 2xy i. (8.345)

| {z } |{z}
ℜ(T(z))=T (x,y) ℑ(T(z))=ψ(x,y)

© 05 December 2025, Joseph M. Powers. All rights reserved.

8.5. HEAT TRANSFER ANALOG 325

We consider ℜ(T(z)) to be the harmonic conjugate of ℑ(T(z)). Both functions are harmonic, but they
also arise from the same source T(x, y). Now Laplace’s equation is linear, so the method of superposition
applies. Because of linearity, we can separately examine the solutions. Let us consider the real part:
ℜ(T(z)) = T (x, y) = x2 − y 2 . (8.346)
Clearly, it satisfies Laplace’s equation as
∂2T ∂2T
+ = 2 − 2 = 0. (8.347)
∂x2 ∂y 2
For this field, the heat flux vector is
 ∂T   
∂x −2x
q = −∇T = − ∂T = . (8.348)
∂y 2y

Remarkably, curves on which the imaginary part, ℑ(T), is constant:

2xy = ψ, (8.349)
are parallel to q; thus, they are adiabats. Parameterizing these curves, we can say
ψ
x = s, y= . (8.350)
2s
So
dx dy ψ
= 1, = − 2. (8.351)
ds ds 2s
The tangent vector t is then
dx dy ψ y
t= i+ j = i − 2 j = i − j. (8.352)
ds ds 2s x
If t is parallel to q, then the cross product of the two must be zero. Let us check.
i j k
t×q = 1 −y/x 0 = (2y − 2y)k = 0. (8.353)
−2x 2y 0
Curves of constant T along with the heat flux vector field q for k = 1 are shown in Fig. 8.22. Note also
that the imaginary part of T(z) satisfies Laplace’s equation. With ℑ(T(z)) = ψ(x, y) = 2xy, we see
∂2ψ ∂2ψ ∂2ψ ∂2ψ
= 0, = 0, + = 0. (8.354)
∂x2 ∂y 2 ∂x2 ∂y 2
We also see that
d
− T(z) = qx − iqy , (8.355)
dz
d
− z 2 = qx − iqy , (8.356)
dz
−2z = qx − iqy , (8.357)
−2x − 2yi = qx − iqy . (8.358)
So
qx = −2x, qy = 2y. (8.359)

© 05 December 2025, Joseph M. Powers. All rights reserved.

326 CHAPTER 8. POTENTIAL FLOW

2.0 - 2

- 3
0
1.5
2

y 1.0- 1

0.5

0.0 1 3
0.0 0.5 1.0 1.5 2.0
x

Figure 8.22: A fundamental solution, T (x, y) = x2 − y 2, to Laplace’s equation along with

heat flux vector field for k = 1.

Example 8.7
Show that
p 
T (x, y) = C1 ln (x − x0 )2 + (y − y0 )2 + C2 , (8.360)

is a solution to Laplace’s equation.

Taking the partial derivatives, we get

!
∂2T 1 2(x − x0 )2
= C1 − 2 , (8.361)
∂x2 2
(x − x0 ) + (y − y0 )2
((x − x0 )2 + (y − y0 )2 )
!
∂2T 1 2(y − y0 )2
= C1 − 2 . (8.362)
∂y 2 (x − x0 )2 + (y − y0 )2 ((x − x0 )2 + (y − y0 )2 )

Thus,

∂2T ∂2T
2
+ = 0, (8.363)
∂x ∂y 2
and Laplace’s equation is satisfied for arbitrary x0 , y0 . If x0 = y0 = 0, and transforming to polar
coordinates with r2 = x2 + y 2 , our fundamental solution becomes

T (r, θ) = C1 ln r + C2 . (8.364)

© 05 December 2025, Joseph M. Powers. All rights reserved.

8.5. HEAT TRANSFER ANALOG 327

One can show that Laplace’s equation in polar coordinates is represented as

 
1 ∂ ∂T 1 ∂2T
r + 2 2 = 0. (8.365)
r ∂r ∂r r ∂θ
The fundamental solution obviously satisfies this as
 
1 ∂ C1 1 ∂2
r + 2 (C1 ln r + C2 ) = 0. (8.366)
r ∂r r r |∂θ2 {z }
| {z }
=0 =0

In three-dimensional systems, the fundamental solution to Laplace’s equation is

C1
T (x, y, z) = p + C2 . (8.367)
(x − x0 ) + (y − y0 )2 + (z − z0 )2
2

Direct substitution reveals it satisfies

∂2T ∂2T ∂2T
2
+ 2
+ = 0. (8.368)
∂x ∂y ∂z 2
One can also sum solutions centered at different values of x0 , y0 , z0 . If x0 = y0 = z0 = 0, transformation
to spherical coordinates reveals the fundamental solution is
C1
T (r, θ, φ) = + C2 , (8.369)
r
that satisfies Laplace’s equation in spherical coordinates:
   
1 ∂ 2 ∂T 1 ∂2T 1 ∂ ∂T
r + + sin φ = 0. (8.370)
r2 ∂r ∂r r2 sin2 φ ∂θ2 r2 sin φ ∂φ ∂φ
Extending to higher dimensions, we find
C1
T (w, x, y, z) = + C2 , (8.371)
(w − w0 )2 + (x − x0 )2 + (y − y0 )2 + (z − z0 )2
satisfies
∂2T ∂2T ∂2T ∂2T
2
+ 2
+ 2
+ = 0. (8.372)
∂w ∂x ∂y ∂z 2
And for general N -dimensional systems with N ≥ 3 we find

N
! −(N2−2)
X
T (x1 , . . . , xN ) = C1 (xn − xon )2 + C2 , (8.373)
n=1

satisfies
XN
∂2T
= 0. (8.374)
n=1
∂x2n

With
N
!1/2
X
2
r= (xn − xon ) , (8.375)
n=1

© 05 December 2025, Joseph M. Powers. All rights reserved.

328 CHAPTER 8. POTENTIAL FLOW

we can say
C1
T (r) = + C2 , N ≥ 3, (8.376)
rN −2
satisfies the N -dimensional Laplace’s equation.
This is consistent with solutions to the one-dimensional Laplace’s equation for hyperspherical co-
ordinates:
 
1 d N −1 dT
r = 0. (8.377)
rN −1 dr dr
For various values of N , we get fundamental solutions

N = 1, planar, T (r) = C1 r + C2 , (8.378)

N = 2, polar, T (r) = C1 ln r + C2 , (8.379)
C1
N = 3, spherical, T (r) = + C2 , (8.380)
r
C1
N ≥ 3, hyperspherical, T (r) = N
+ C2 . (8.381)
r −2

Example 8.8
Show in one dimension that the fundamental solution
1
T (x) = (1 − |x|) , (8.382)
2
satisfies

−∇T · ∇T = δ(x), T (1) = T (−1) = 0. (8.383)

Here we consider that T and x have been suitably scaled so they are now dimensionless.

Away from the origin, this is Laplace’s equation. Formally it is a Poisson equation because of the
presence of a source term. In general, the Poisson equation takes the form

−∇T · ∇T = f (x). (8.384)

In this case the source term f is a point source. The point source is embodied in δ(x), the Dirac delta
function. Let us integrate both sides over a volume that has x ∈ [−1, 1], y ∈ [−L, L], z ∈ [−L, L] and
use the divergence theorem, giving
Z Z
− ∇T · ∇T dV = δ(x) dV, (8.385)
V V
Z Z L Z L Z 1
T
− n · ∇T dA = δ(x) dx dy dz, (8.386)
A −L −L
| −1 {z }
=1
Z L Z L
= dy dz, (8.387)
−L −L
2
= 4L . (8.388)

Now ∇T has no component in the y or z directions. So for the sides which have n = (0, 0, ±1)T and
(0, ±1, 0)T , we get nT · ∇T = 0. So we only need to consider the portion of the volume whose sides are

© 05 December 2025, Joseph M. Powers. All rights reserved.

8.5. HEAT TRANSFER ANALOG 329

normal to the x axis, n = ∓i. Now we have ∇T = (∂T /∂x)i = ±(1/2)i. So nT · ∇T = −1/2. We also
have dA = dy dz. This gives
Z Z
1 1
− − dA − − dA = 4L2 , (8.389)
A(x=−1) 2 A(x=1) 2
Z Z
1 1
dA + dA = 4L2 , (8.390)
2 A(x=−1) 2 A(x=1)
Z Z Z Z
1 L L 1 L L
dy dz + dy dz = 4L2 , (8.391)
2 −L −L 2 −L −L
4L2 = 4L2 , (8.392)
1 = 1. (8.393)

Alternatively, we could simply operate as follows:

d2 T
− = δ(x), (8.394)
dx2
Z 1 2 Z 1
d T
− 2
dx = δ(x) dx, (8.395)
−1 dx −1
dT dT
− = 1, (8.396)
dx x=−1 dx x=1
 
1 −1
− = 1, (8.397)
2 2
1 = 1. (8.398)

If we were to extend this to a dimensional problem in heat transfer, taking the volumetric thermal
power deposition Q with units of W/m3 , constant cross sectional area in the y − z plane of A with
units of m2 , domain half length L with units of m, and thermal conductivity k with units of W/m/K,
our solution is
QL2  x 
T (x) = 1− . (8.399)
2k L
This solution satisfies the equation and boundary conditions

d2 T QL
− = δ(x), T (−L) = T (L) = 0. (8.400)
dx2 k
Z L 2 Z
d T QL L
− 2
dx = δ(x) dx, (8.401)
−L dx k −L
dT dT QL
− = , (8.402)
dx −L dx L k
 
QL QL QL
− − = , (8.403)
2k 2k k
1 = 1. (8.404)

We can consider QL with units of W/m2 to be the power deposition per unit area; thus, it is an energy
flux. Note that δ(x) must have units of 1/m.

© 05 December 2025, Joseph M. Powers. All rights reserved.

330 CHAPTER 8. POTENTIAL FLOW

Example 8.9
Show in two dimensions that the fundamental solution
1
T (r) = − ln r, (8.405)
2π
satisfies

−∇T · ∇T = δ(x)δ(y). (8.406)

We consider that T and x and y have been suitably scaled so they are now dimensionless.

Away from the origin, this is Laplace’s equation. Formally it is a Poisson equation because of the
presence of a source term. In this case the source term is a point source. Let us integrate both sides
over a long cylindrical volume that encloses the origin and use the divergence theorem, giving
Z Z
− ∇T · ∇T dV = δ(x)δ(y) dV, (8.407)
V V
Z Z L Z Z
− nT · ∇T dA = δ(x)δ(y) dy dx dz, (8.408)
A −L x y
Z L Z Z
= δ(x) δ(y) dy dx dz, (8.409)
−L x y
| {z }
=1
Z L Z
= δ(x) dx dz, (8.410)
−L
| x {z }
=1
= 2L. (8.411)

Now ∇T has no component in the z direction. So for the ends of the cylinder, which have n = (0, 0, 1)T
and (0, 0, −1)T , we get nT · ∇T = 0. So we only need to consider the portion of the cylinder’s surface
area that is not at either end. Now ∇T is such that it is parallel to the unit normal n, but points in
opposite direction; we have ∇T = (∂T /∂r)er = −1/(2πr)er . And n = er . So nT · ∇T = −1/(2πr).
For the cylinder of radius r, we also have dA = r dθ dz. This gives
Z
1
− − dA = 2L, (8.412)
A 2πr
Z L Z 2π
1
r dθ dz = 2L, (8.413)
−L 0 2πr
Z L Z 2π
1
dθ dz = 2L, (8.414)
2π −L 0
1 = 1. (8.415)

If we were to extend this to a dimensional problem in heat transfer, taking the volumetric thermal
power deposition Q with units of W/m3 , constant length in the z axis of L with units of m, and thermal
conductivity k with units of W/m/K, our solution is

QL2 r
T (r) = − ln . (8.416)
2πk L

© 05 December 2025, Joseph M. Powers. All rights reserved.

8.5. HEAT TRANSFER ANALOG 331

This solution satisfies the equation and boundary conditions

QL2
−∇T · ∇T = δ(x)δ(y), T (r → ∞) = 0, (8.417)
k
Z Z
QL2
−∇T · ∇T dV = δ(x)δ(y) dV, (8.418)
V V k
Z Z L Z Z
QL2
−nT · ∇T dA = δ(x)δ(y) dy dx dz. (8.419)
A −L x y k
Once again, the end effects at z = −L and z = L make no contribution, so the integrals reduce to
Z L Z 2π
QL2 QL2
r dθ dz = (2L), (8.420)
−L 0 2πkr k
QL2 QL2
(2L) = (2L), (8.421)
k k
1 = 1. (8.422)
Here QL2 has units of W/m and is the power deposition per unit length.

Example 8.10
Show in three dimensions that the fundamental solution
1
T (r) = , (8.423)
4πr
satisfies
−∇T · ∇T = δ(x)δ(y)δ(z). (8.424)
We consider that T and x, y, and z have been suitably scaled so they are now dimensionless.

Away from the origin, this is Laplace’s equation. Formally it is a Poisson equation because of the
presence of a source term. In this case the source term is a point source. Let us integrate both sides
over a spherical volume that encloses the origin and use the divergence theorem, giving
Z Z
− ∇T · ∇T dV = δ(x)δ(y)δ(z) dV, (8.425)
V V
Z Z Z Z
− nT · ∇T dA = δ(x)δ(y)δ(z) dz dy dx, (8.426)
A x y z
Z 2π Z π
∂T 2
− r sin φ dφ dθ = 1, (8.427)
0 0 ∂r
Z 2π Z π  
−1
− 2
r2 sin φ dφ dθ = 1, (8.428)
0 0 4πr
Z 2π Z π
1
sin φ dφ dθ = 1, (8.429)
4π 0 0
Z 2π
1 π
(− cos φ)|0 dθ = 1, (8.430)
4π 0
Z 2π
1
2 dθ = 1, (8.431)
4π 0
1 = 1. (8.432)

© 05 December 2025, Joseph M. Powers. All rights reserved.

332 CHAPTER 8. POTENTIAL FLOW

T
0.5

0.4
y y
0.3

0.2

0.1
T T

- 1.0 - 0.5 0.0 0.5 1.0

x
x x
a) b) c)

Figure 8.23: Fundamental solutions to Laplace’s equation for a) planar, b) cylindrical, and
c) spherical systems.

If we were to extend this to a dimensional problem in heat transfer, taking the volumetric thermal
power deposition to be Q with units of W/m3 , and thermal conductivity to be k with units of W/m/K,
our solution is
QL3
T (r) = . (8.433)
4πkr
This solution satisfies the equation and boundary conditions

QL3
−∇T · ∇T = δ(x)δ(y)δ(z), T (r → ∞) = 0, (8.434)
k
Z Z 3
QL
− ∇T · ∇T dV = δ(x)δ(y)δ(z) dV, (8.435)
V V k
Z Z Z Z
QL3
− nT · ∇T dA = δ(x)δ(y)δ(z) dz dy dx, (8.436)
A x y z k
Z 2π Z π   Z Z Z
QL3 QL3
− − r2 sin φ dφ dθ = δ(x)δ(y)δ(z) dz dy dx, (8.437)
0 0 4πkr2 x y z k
QL3 QL3
= , (8.438)
k k
1 = 1. (8.439)

Here QL3 is the power deposition in W from the point source.

Plots for the fundamental solutions for planar, cylindrical, and spherical coordinate sys-
tems are shown in Fig. 8.23. Because Laplace’s equation is linear, linear combinations of
fundamental solutions will also satisfy Laplace’s equation. As such, we can distribute point
sources at different points in space.

Example 8.11
Consider a two-dimensional domain with a point source of thermal energy at (x, y) = (1, 1) and a
point sink of thermal energy at (x, y) = (−1, −1). Find the temperature field and the heat flux vector
field.

© 05 December 2025, Joseph M. Powers. All rights reserved.

8.5. HEAT TRANSFER ANALOG 333

2 0.05 0.2

0.2
0.4
1 - 0.05
0.3

y 0
0.1

- 0.2 - 0.3
-1
- 0.4

- 0.2 0
-2 - 0.1

-2 -1 0 1 2
x

Figure 8.24: Temperature contours, T (x, y), and heat flux vector field, q(x, y), for the su-
perposition of a point source and sink of thermal energy in a two-dimensional geometry.

A point source centered at the origin of a two-dimensional field has

1 1 p
T =− ln r = − ln x2 + y 2 . (8.440)
2π 2π
We have seen that this satisfies, Eq. (8.406), −∇T · ∇T = δ(x)δ(y), with the Dirac delta functions each
of unit strength. Now a sink has the opposite sign. A source centered at (1, 1) and a sink centered at
(−1, −1) has the temperature field that arises from superposition of two fields:
1 p 1 p
T =− ln (x − 1)2 + (y − 1)2 + ln (x + 1)2 + (y + 1)2 . (8.441)
| 2π {z } |2π {z }
source sink

It satisfies
 
∂ 2T ∂2T
−∇T · ∇T = − 2
+ = δ(x − 1)δ(y − 1) − δ(x + 1)δ(y + 1). (8.442)
∂x ∂y 2

Direct substitution of the temperature field into Laplace’s equation reveals that as long as (x, y) 6= (1, 1),
(x, y) 6= (−1, −1), the Laplacian is zero. So it is a solution. The heat flux vector is
x2 +2xy−y 2 −2
!
π(x4 +2x2 y 2 −8xy+y 4 +4)
q = −∇T = −x2 +2xy+y 2 −2
. (8.443)
π(x4 +2x2 y 2 −8xy+y 4 +4)

A plot of the temperature and heat flux vector fields is shown in Fig. 8.24. We clearly see the heat
flux emanating from the source at (x, y) = (1, 1) and flowing into the sink at (x, y) = (−1, −1). It is
easy to add sources and sinks. A plot of the temperature and heat flux vector fields for point sources
at (±1, 0) and point sinks at (0, ±1) is given in Fig. 8.25.

© 05 December 2025, Joseph M. Powers. All rights reserved.

334 CHAPTER 8. POTENTIAL FLOW

y 0

-1

-2
-2 -1 0 1 2
x

Figure 8.25: Temperature contours, T (x, y), and heat flux vector field, q(x, y), for the su-
perposition of two point sources and two point sinks in a two-dimensional geometry.

Example 8.12
Consider a three-dimensional domain with a point source of thermal energy at (x, y, z) = (1, 1, 0)
and a point sink of thermal energy at (x, y, z) = (−1, −1, 0). Find the temperature field and the heat
flux vector field.

A point source centered at the origin of a three-dimensional field has

1 1
T = = p . (8.444)
4πr 4π x + y 2 + z 2
2

We have seen that this satisfies, Eq. (8.424), −∇T · ∇T = δ(x)δ(y)δ(z), with the Dirac delta functions
each of unit strength. Now a sink has the opposite sign. A source centered at (1, 1, 0) and a sink
centered at (−1, −1, 0) has the temperature field that arises from superposition of two fields:

1 1
T = p − p . (8.445)
2 2
4π (x − 1) + (y − 1) + z 2 4π (x + 1) + (y + 1)2 + z 2
2
| {z } | {z }
source sink

It satisfies
 
T ∂2T ∂2T ∂2T
−∇ · ∇T = − + + = δ(x − 1)δ(y − 1)δ(z) − δ(x + 1)δ(y + 1)δ(z). (8.446)
∂x2 ∂y 2 ∂z 2

© 05 December 2025, Joseph M. Powers. All rights reserved.

8.5. HEAT TRANSFER ANALOG 335

2 0.01 0.05

0.05
0.2
1

- 0.02 0.1

y 0
0.02
- 0.1

-1
- 0.2
- 0.05
- 0.05 0
-2 - 0.01

-2 -1 0 1 2
x
a) b)

Figure 8.26: Superposition of a point source and sink of thermal energy in a three-dimensional
geometry: a) temperature contours, T (x, y, z = 0), and heat flux vector field, q(x, y, z = 0),
b) heat flux vector field q(x, y, z).

Direct substitution of the temperature field into Laplace’s equation reveals that as long as (x, y) 6= (1, 1),
(x, y) 6= (−1, −1), the Laplacian is zero. So it is a solution. The heat flux vector is
 x−1 x+1 
4π((x−1)2 +(y−1)2 +z 2 )3/2
− 4π((x+1)2 +(y+1) 2 +z 2 )3/2
 y−1 y+1 
q = −∇T =  4π((x−1)2 +(y−1) 2 +z 2 )3/2
− 4π((x+1)2 +(y+1) 2 +z 2 )3/2  . (8.447)
z z
4π((x−1)2 +(y−1)2 +z 2 ) 3/2 − 4π((x+1)2 +(y+1)2 +z 2 ) 3/2

Confining attention to the plane of z = 0, a plot of the temperature and heat flux vector fields is shown in
Fig. 8.26a. We show in Fig. 8.26b the three-dimensional vector field of heat flux. We clearly see the heat
flux emanating from the source at (x, y, z) = (1, 1, 0) and flowing into the sink at (x, y, z) = (−1, −1, 0).

We have developed Laplace’s equation loosely in the context of two-dimensional steady

heat transfer in an isotropic material. One can also consider Laplace’s equation in the context
of two-dimensional, irrotational, incompressible, inviscid, steady fluid mechanics. There are
clear analogs between the two physically motivating problems. They are as follows. We
can consider temperature T (x, y) to be a potential field. In fluid mechanics, we can consider
φ(x, y) to be the equivalent velocity potential field. For isotropic thermal conductivity k = 1,
the negative of the gradient of the temperature field yields the heat flux vector q = −∇T .
In fluid mechanics, the equivalent is that the velocity vector is the gradient of the velocity
potential u = ∇φ. Both the heat flux and velocity vector fields are curl-free because they

© 05 December 2025, Joseph M. Powers. All rights reserved.

336 CHAPTER 8. POTENTIAL FLOW

are gradients of a scalar field: ∇ × q = 0, ∇ × u = 0. The vector fields of both q and

u are divergence-free because of the first law of thermodynamics, and incompressibility,
respectively: ∇T · q = 0, ∇T · u = 0. Both potentials satisfy Laplace’s equation: ∇2 T = 0,
∇2 φ = 0.
Lastly Eq. (8.326) implies the definition of what we will call a thermal stream function,
ψ, defined such that
∂ψ ∂ψ
qx = , qy = − . (8.448)
∂y ∂x
Substituting these into Eq. (8.326) yields
   
∂ ∂ψ ∂ ∂ψ
− − = 0, (8.449)
∂x ∂x ∂y ∂y
∂2ψ ∂2ψ
+ 2 = 0, (8.450)
∂x2 ∂y
∇2 ψ = 0. (8.451)

Moreover, it is in fact the case that

ψ = ℑ(T(z)), (8.452)

where T(z) is the complex temperature field from Eq. (8.342). Recalling that qx = −∂T /∂x
and qy = −∂T /∂y, we also have

∂T ∂ψ ∂T ∂ψ
− = , − =− . (8.453)
∂x ∂y ∂y ∂x
If we take φ = −T , we recover precisely the Cauchy-Riemann equations from complex
variable theory.

8.5.3 Conservation of thermal energy

Consider the general Laplace’s equation:

∇T · ∇T = 0. (8.454)

For this equation, there are no sources of energy, and it must be in equilibrium as there
is no time-dependency. We can think of this as a two- or three-dimensional equation. Let
us simply think of it as three-dimensional. Reduction to two dimensions is straightforward.
Integrate over a volume V that is bounded by a surface A to get
Z Z
T
∇ · ∇T dV = 0 dV . (8.455)
V V
| {z }
=0

© 05 December 2025, Joseph M. Powers. All rights reserved.

8.5. HEAT TRANSFER ANALOG 337

Use the divergence theorem theorem to get

Z
nT · ∇T dA = 0. (8.456)
A

Here n is the outer surface normal vector for A. We recall the heat flux q = −k∇T by
Fourier’s law, so
Z
−1
nT · q dA = 0. (8.457)
k A
The net heat flux entering the volume through the bounding surface A must be zero.

8.5.4 Biharmonic equation

We present brief results for an equation that is an extension of Laplace’s equation: the
biharmonic equation:

(∇T · ∇)(∇T · ∇)ψ = ∇2 (∇2 ψ) = ∇4 ψ = 0. (8.458)

The equation arises in fluid mechanics if inertia is negligible and viscous forces balance pres-
sure forces. It also arises linear elastic models of static solids. All solutions to Laplace’s
equation, which are harmonic solutions, are also solutions to the biharmonic equation. How-
ever, there exist solutions to the biharmonic equation that are not solutions to Laplace’s
equation.
In two-dimensional Cartesian coordinates, the biharmonic equation is
 2  2 
∂ ∂2 ∂ ∂2
+ + ψ = 0, (8.459)
∂x2 ∂y 2 ∂x2 ∂y 2
| {z }| {z }
∇2 ∇2
4 4
∂ ψ ∂ ψ ∂4ψ
+ 2 + = 0. (8.460)
∂x2 ∂x2 ∂y 2 ∂y 4
In Cartesian coordinates, the general solution can be shown to be

ψ(x, y) = xv(x, y) − yu(x, y) + w(x, y). (8.461)

Here u(x, y), v(x, y), and w(x, y) are required to be harmonic functions and so satisfy
Laplace’s equation. Additionally, v(x, y) must be a harmonic conjugate, as introduced in
Sec. 8.5.2, of u(x, y).

Example 8.13
Consider the harmonic conjugate functions introduced in Eq. (8.345), u = x2 − y 2 , v = 2xy, and
the harmonic function w(x, y) = x − y. Build a solution ψ(x, y) to the biharmonic equation.

© 05 December 2025, Joseph M. Powers. All rights reserved.

338 CHAPTER 8. POTENTIAL FLOW

y 0

-1

-2
-2 -1 0 1 2
x

Figure 8.27: Contours of the biharmonic function ψ(x, y) = x2 y + x + y 3 − y.

We first check

∇2 u = ∇2 (x2 − y 2 ) = 2 − 2 = 0, (8.462)
∇2 v = ∇2 (2xy) = 0, (8.463)
∇2 w = ∇2 (x − y) = 0. (8.464)

And we know from Eq. (8.345) that the harmonic conjugates arose from the complex function z 2 . So

ψ(x, y) = xv − yu + w = x(2xy) − y(x2 − y 2 ) + x − y = x2 y + x + y 3 − y. (8.465)

Direct calculation shows that

∇2 ψ = 8y 6= 0, (8.466)

but

∇4 ψ = 0, (8.467)

because ∇2 (8y) = 0. We show in Fig. 8.27 a contour plot of the biharmonic function ψ(x, y).

More generally, for two-dimensional Cartesian coordinates, we can generate a solution

the the biharmonic equation by taking

ψ = ℜ (f (z) + zg(z)) , (8.468)

where z = x + iy and both f and g are arbitrary complex functions of z.

© 05 December 2025, Joseph M. Powers. All rights reserved.

8.5. HEAT TRANSFER ANALOG 339

y 0

-1

-2
-2 -1 0 1 2
x

Figure 8.28: Contours of the biharmonic function ψ(x, y) = x(x2 − 3y 2)(1 + x2 + y 2 ).

Example 8.14
Generate a solution to the biharmonic equation using the functions f (z) = z 3 and g(z) = z 4 .

We get

ψ = ℜ (f (z) + zg(z)) , (8.469)

= ℜ (x + iy)3 + (x − iy)(x + iy)4 , (8.470)
2 2 2 2 2 2 2 2


= ℜ x(x − 3y )(1 + x + y ) + y(3x − y )(1 + x + y )i , (8.471)
2 2 2 2
= x(x − 3y )(1 + x + y ). (8.472)

We check and find

∇2 ψ = 16(x3 − 3xy 2 ), ∇4 ψ = 0, (8.473)

as expected. The harmonic conjugate φ is

φ = ℑ x(x2 − 3y 2 )(1 + x2 + y 2 ) + y(3x2 − y 2 )(1 + x2 + y 2 )i = y(3x2 − y 2 )(1 + x2 + y 2 ). (8.474)

We check and find that φ also satisfies the biharmonic equation and yields

∇2 φ = 48x2 y − 16y 3 , ∇4 φ = 0. (8.475)

We show in Fig. 8.28 a contour plot of the biharmonic function ψ(x, y).

© 05 December 2025, Joseph M. Powers. All rights reserved.

340 CHAPTER 8. POTENTIAL FLOW

In plane polar coordinates, neglecting all variation in θ, the biharmonic equation reduces
to
   
4 1 d d 1 d dψ
∇ ψ= r r = 0. (8.476)
r dr dr r dr dr

This has general solution

ψ(r) = c1 + c2 ln r + c3 r 2 ln r + c4 r 2 . (8.477)

The fundamental solutions of ∇4 ψ = 0 are a constant, r 2 , ln r, and r 2 ln r. Linear combi-

nations of these are solutions. We can also displace the origin and form solutions and their
linear combinations.

Example 8.15
Build a solution to the two-dimensional biharmonic equation that is some linear combination of
fundamental solutions centered at (x, y) = (−1, 0), (0, 0), and (1, 0).

We have a wide variety of choices. Let us take r1 to be the distance from (−1, 0), r2 to be the
distance from (0, 0), and r3 to be the distance from (1, 0). Then let us choose

ψ = r12 ln r1 − ln r2 + r32 ln r3 , (8.478)

p p
p
2 2 2 2
= (x + 1) + y ln (x + 1)2 + y 2 − ln x2 + y 2 + (x − 1) + y ln (x − 1)2 + y 2 . (8.479)

Direct calculation reveals that ∇2 ψ 6= 0 but ∇4 ψ = 0. We show in Fig. 8.29 a contour plot of the
biharmonic function ψ(x, y).

© 05 December 2025, Joseph M. Powers. All rights reserved.

8.5. HEAT TRANSFER ANALOG 341

y 0

-1

-2
-2 -1 0 1 2
x

Figure 8.29: Contours of a two-dimensional biharmonic function that satisfies ∇4 ψ(x, y) = 0.

© 05 December 2025, Joseph M. Powers. All rights reserved.

342 CHAPTER 8. POTENTIAL FLOW

© 05 December 2025, Joseph M. Powers. All rights reserved.

Chapter 9

One-dimensional compressible flow

see Powers, Chapter 9,

see Thompson,
see Shapiro, Chapters 4-8,
see Anderson (2021), Chapters 3, 5, 7,
see Kuethe and Chow, Chapter 9,
see Yih, Chapter 6,
see Liepmann and Roshko, Chapter 2,
see Whitaker, Chapter 10,
see Courant and Friedrichs,
see White (1986), Chapters 7-9.

This chapter will focus on one-dimensional flow of a compressible fluid. The following topics
will be covered:

• thermodynamics of general compressible fluids,

• development of generalized one-dimensional flow equations,

• isentropic flow with area change,

• flow with normal shock waves, and

• the method of characteristics.

We will assume for this chapter:

• v ≡ 0, w ≡ 0, ∂/∂y ≡ 0, ∂/∂z ≡ 0; one-dimensional flow.

Friction and heat transfer will not be modeled rigorously. Instead, they will be modeled in
a fashion that loosely captures the relevant physics and retains analytic tractability. Math-
ematically, we will not model friction and heat transfer as a classical diffusion processes;
consequently, we will consider µ ≡ 0 and k ≡ 0. However we will introduce simpler, less

343
344 CHAPTER 9. ONE-DIMENSIONAL COMPRESSIBLE FLOW

rigorous, new terms to model friction and heat transfer. They will have a different mathe-
matical character. As a consequence, our solutions will not represent rational limiting cases
of the more fundamental Navier-Stokes equations. Direct comparison of results using our
modeling approximations will never completely agree with equivalent (and expensive) pre-
dictions of compressible Navier-Stokes equations. Further, we will ignore the influences of
an external body force, fi = 0. Our model will best be seen as an adaptation of the Euler
equations of Ch. 6.5. It will have the advantage of yielding rapid and non-intuitive insight
into how actual fluids behave under the extreme conditions of flow near or above the speed
of sound.

9.1 Thermodynamics of general compressible fluids

Here let us briefly consider compressible fluids with general equations of state. The topic
is broad, but we will restrict attention to the speed of sound as well as how a typical
thermodynamic analysis is affected by non-ideal equations of state.

9.1.1 Maxwell relation

Let us develop a necessary Maxwell1 relation. The topic of Maxwell relations is non-trivial
and draws upon a systematic knowledge of thermodynamic potentials. Additional details
are given by Powers (2016). We will only develop here what is necessary for our purposes.
Consider first the Helmholtz free energy, â, defined in Eq. (5.225) in terms of other thermo-
dynamic variables as

â = e − T s. (9.1)

Then differentiating, we get

dâ = de − T ds − s dT. (9.2)

Now use the Gibbs equation, Eq. (4.163), to eliminate de to get

dâ = (−p dv̂ + T ds) −T ds − s dT, (9.3)

| {z }
de
= −p dv̂ − s dT. (9.4)

We can think of â as â = â(v̂, T ), because intensive thermodynamic properties are functions

of two independent state variables for simple compressible substances. Then calculus tells
us
∂â ∂â
dâ = dv̂ + dT. (9.5)
∂v̂ T ∂T v̂
1
James Clerk Maxwell, 1831-1879, Scottish mathematical physicist.

© 05 December 2025, Joseph M. Powers. All rights reserved.

9.1. THERMODYNAMICS OF GENERAL COMPRESSIBLE FLUIDS 345

Comparing to Eq. (9.4), we see that we must have what is the equivalent of Eqs. (5.237):

∂â ∂â
−p = , −s = . (9.6)
∂v̂ T ∂T v̂

Differentiating the first with respect to T and the second with respect to v̂ gives
∂p ∂ 2 â ∂s ∂ 2 â
− = , − = . (9.7)
∂T v̂ ∂T ∂v̂ ∂v̂ T ∂v̂∂T
Assuming â is continuous and sufficiently differentiable, the order of differentiation of the
mixed second partials does not matter, thus giving the Maxwell relation
∂p ∂s
= . (9.8)
∂T v̂ ∂v̂ T

This is useful because ∂p/∂T |v̂ is available from the thermal equation of state, and it will
be required in analysis of the next section where we find caloric equations of state that are
consistent with a given thermal equation of state.

9.1.2 Internal energy from thermal equation of state

Let us first find the internal energy e(T, v̂) for a general material whose thermal equation of
state, p = p(v̂, T ), is known. Starting then with e, we have

e = e(T, v̂), (9.9)

∂e ∂e
de = dT + dv̂. (9.10)
∂T v̂ ∂v̂ T

Specific heat at constant volume, cv , for a general material is defined as

∂e
cv = . (9.11)
∂T v̂

Using this, Eq. (9.10) becomes

∂e
de = cv dT + dv̂. (9.12)
∂v̂ T

Now from the Gibbs equation, Eq. (4.163), we find

de = T ds − p dv̂, (9.13)
de ds
= T − p, (9.14)
dv̂ dv̂
∂e ∂s
= T − p. (9.15)
∂v̂ T ∂v̂ T

© 05 December 2025, Joseph M. Powers. All rights reserved.

346 CHAPTER 9. ONE-DIMENSIONAL COMPRESSIBLE FLOW

Substitute from the Maxwell relation, Eq. (9.8), to get:

∂e ∂p
=T − p. (9.16)
∂v̂ T ∂T v̂

Thus, Eq. (9.12) can be rewritten as

 
∂p
de = cv dT + T − p dv̂. (9.17)
∂T v̂
| {z }
∂e
|
∂ v̂ T

Integrating, we then get the caloric equation of state:

Z e Z T Z v̂  
∂p
dê = cv (T̂ ) dT̂ + T − p dṽ, (9.18)
eo To v̂o ∂T ṽ
Z T Z v̂  
∂p
e(T, v̂) = eo + cv (T̂ ) dT̂ + T − p dṽ. (9.19)
To v̂o ∂T ṽ

This is the caloric equation of state that is thermodynamically consistent with the given
thermal equation of state.

Example 9.1
Find a general expression for e(T, v̂) if we have an ideal gas:

RT
p(T, v̂) = . (9.20)
v̂

Proceed as follows:
∂p R
= , (9.21)
∂T v v̂
∂p RT
T −p = − p, (9.22)
∂T v̂ v̂
RT RT
= − = 0. (9.23)
v̂ v̂
Thus, e is
Z T
e(T ) = eo + cv (T̂ ) dT̂ . (9.24)
To

Iff cv is a constant, then we have CPIG, and the caloric equation of state is

e(T ) = eo + cv (T − To ). (9.25)

© 05 December 2025, Joseph M. Powers. All rights reserved.

9.1. THERMODYNAMICS OF GENERAL COMPRESSIBLE FLUIDS 347

Example 9.2
Find a general expression for e(T, v̂) for a van der Waals2 gas:
RT a
p(T, v̂) = − . (9.26)
v̂ − b v̂ 2

Proceed as before:
∂p R
= , (9.27)
∂T v̂ v̂ − b
∂p RT
T −p = − p, (9.28)
∂T v̂ v̂ − b
 
RT RT a
= − − , (9.29)
v̂ − b v̂ − b v̂ 2
a
= . (9.30)
v̂ 2
Thus, the caloric equation of state for e is
Z T Z v̂
a
e(T, v̂) = eo + cv (T̂ ) dT̂ + 2
dṽ, (9.31)
To v̂o ṽ
Z T  
1 1
= eo + cv (T̂ ) dT̂ + a − . (9.32)
To v̂o v̂

If cv is constant, the caloric equation of state for the van der Waals gas reduces to
 
1 1
e(T, v̂) = eo + cv (T − To ) + a − . (9.33)
v̂o v̂

Example 9.3
A van der Waals gas with R = 200 J/kg/K, a = 150 Pa m6 /kg2 , b = 0.001 m3 /kg, cv =
(350 + 0.2(T − 300 K)) J/kg/K begins at T1 = 300 K, p1 = 105 Pa. It is isothermally compressed
to state 2 where p2 = 106 Pa. It is then isochorically heated to state 3 where T3 = 1000 K. Find w13 ,
q13 , and s3 − s1 . Assume the surroundings are at 1000 K.

Recall
RT a
p= − 2. (9.34)
v̂ − b v̂
2
Johannes Diderik van der Waals, 1837-1923, Dutch thermodynamicist and 1910 Nobel laureate in
physics for his work in developing his celebrated equation of state.

© 05 December 2025, Joseph M. Powers. All rights reserved.

348 CHAPTER 9. ONE-DIMENSIONAL COMPRESSIBLE FLOW

So at state 1
 
J
200 kg K (300 K) 150 Pa m6 /kg2
105 Pa = − . (9.35)
v̂1 − 0.001 m3
kg
v̂12

or expanding, and ignoring the units

−0.15 + 150v̂ − 60100v̂ 2 + 100000v̂ 3 = 0. (9.36)

This is a cubic equation, and it thus has three solutions:

m3
v̂1 = 0.598 , (9.37)
kg
m3
v̂1 = 0.00125 − 0.0097i , not physical, (9.38)
kg
m3
v̂1 = 0.00125 + 0.0097i , not physical. (9.39)
kg
Now at state 2 we know p2 and T2 so we can determine v̂2
 
200 kgJ K (300 K) 150 Pa m6 /kg2
106 Pa = − . (9.40)
v̂2 − 0.001 m
kg
3
v̂22

The physical solution is v̂2 = 0.0585 m3 /kg. Now at state 3 we know v̂3 = v̂2 and T3 . Determine p3 :
 
200 kgJ K (1000 K) 150 Pa m6 /kg2 6
p3 = 3 3 −  2 = 3434430 Pa = 3.43 × 10 Pa. (9.41)
0.0585 m m
kg − 0.001 kg 0.0585 m3
kg

R2 R3 R2
Now w13 = w12 + w23 = 1
p dv̂ + 2
p dv̂ = 1
p dv̂ because 2 − 3 is at constant volume. So
Z v̂2  
RT a
w13 = − dv̂, (9.42)
v̂1 v̂ − b v̂ 2
Z v̂2 Z v̂2
dv̂ dv̂
= RT1 −a 2
, (9.43)
v̂1 v̂ − b v̂1 v̂
   
v̂2 − b 1 1
= RT1 ln +a − , (9.44)
v̂1 − b v̂2 v̂1
   
0.0585 − 0.001 1 1
= 200 × 300 ln + 150 − , (9.45)
0.598 − 0.001 0.0585 0.598
J J
= −140408 + 2313 , (9.46)
kg kg
J
= −138095 , (9.47)
kg
kJ
= −138 . (9.48)
kg
The gas is compressed, so the work is negative. Because e is a state property:
Z T3  
1 1
e3 − e1 = cv (T ) dT + a − . (9.49)
T1 v̂1 v̂3

© 05 December 2025, Joseph M. Powers. All rights reserved.

9.1. THERMODYNAMICS OF GENERAL COMPRESSIBLE FLUIDS 349

Now
1
cv = 350 + 0.2(T − 300) = 290 + T, (9.50)
5
so
Z 
T3   
1 1 1
e3 − e1 = 290 + T dT + a − , (9.51)
T1 5 v̂1 v̂3
 
1 2 2

1 1
= 290 (T3 − T1 ) + T − T1 + a − , (9.52)
10 3 v̂1 v̂3
 
1 2 2

1 1
= 290 (1000 − 300) + 1000 − 300 + 150 − , (9.53)
10 0.598 0.0585
= 203000 + 91000 − 2313, (9.54)
J
= 291687 , (9.55)
kg
kJ
= 292 . (9.56)
kg
Now from the first law, we have
e3 − e1 = q13 − w13 , (9.57)
q13 = e3 − e1 + w13 , (9.58)
= 292 − 138, (9.59)
kJ
= 154 . (9.60)
kg
The heat transfer is positive as heat was added to the system.
Now find the entropy change. Manipulate the Gibbs equation, Eq. (4.164):
T ds = de + p dv̂, (9.61)
1 p
ds = de + dv̂, (9.62)
T T
1  a  p
= cv (T ) dT + 2 dv̂ + dv̂, (9.63)
T v̂ T  
1  a  1 RT a
= cv (T ) dT + 2 dv̂ + − dv̂, (9.64)
T v̂ T v̂ − b v̂ 2
cv (T ) R
= dT + dv̂, (9.65)
T v̂ − b
Z T3
cv (T ) v̂3 − b
s3 − s1 = dT + R ln , (9.66)
T1 T v̂1 − b
Z 1000  
290 1 v̂3 − b
= + dT + R ln , (9.67)
300 T 5 v̂1 − b
1000 1 0.0585 − 0.001
= 290 ln + (1000 − 300) + 200 ln , (9.68)
300 5 0.598 − 0.001
= 349 + 140 − 468, (9.69)
J kJ
= 21 = 0.021 . (9.70)
kg K kg K
Is the second law satisfied for each portion of the process? First look at 1 → 2:
e2 − e1 = q12 − w12 , (9.71)

© 05 December 2025, Joseph M. Powers. All rights reserved.

350 CHAPTER 9. ONE-DIMENSIONAL COMPRESSIBLE FLOW

q12 = e2 − e1 + w12 , (9.72)

Z T2  !     
1 1 v̂2 − b 1 1
= cv (T ) dT + a − + RT1 ln +a − . (9.73)
T1 v̂1 v̂2 v̂1 − b v̂2 v̂1
Because T1 = T2 and that fact that we can cancel the terms in a, we get
   
v̂2 − b 0.0585 − 0.001 J
q12 = RT1 ln = 200 × 300 ln = −140408 . (9.74)
v̂1 − b 0.598 − 0.001 kg
Because the process is isothermal, we find
 
v̂2 − b
s2 − s1 = R ln , (9.75)
v̂1 − b
 
0.0585 − 0.001
= 200 ln , (9.76)
0.598 − 0.001
J
= −468.0 . (9.77)
kg K
Entropy drops because heat was transferred out of the system.
Check the second law. Note that in this portion of the process in which the heat is transferred out
of the system, that the surroundings must have Tsurr ≤ 300 K. For this portion of the process let us
take Tsurr = 300 K.
q12
s2 − s1 ≥ ? (9.78)
T
J
J −140408 kg
−468.0 ≥ , (9.79)
kg K 300 K
J J
−468.0 ≥ −468.0 , ok. (9.80)
kg K kg K
Next look at 2 → 3
q23 = e3 − e2 + w23 , (9.81)
Z T3  ! Z v̂3
!
1 1
= cv (T ) dT + a − + p dv̂ . (9.82)
T2 v̂2 v̂3 v̂2

Because the process is isochoric, we have

Z T3
q23 = cv (T ) dT, (9.83)
T2
Z 1000  
T
= 290 + dT, (9.84)
300 5
J
= 294000 . (9.85)
kg
Now look at the entropy change for the isochoric process:
Z T3
cv (T )
s3 − s2 = dT , (9.86)
T2 T
Z T3  
290 1
= + dT , (9.87)
T2 T 5
1000 1
= 290 ln + (1000 − 300), (9.88)
300 5
J
= 489 . (9.89)
kg K

© 05 December 2025, Joseph M. Powers. All rights reserved.

9.1. THERMODYNAMICS OF GENERAL COMPRESSIBLE FLUIDS 351

Entropy rises because heat was transferred into the system.

In order to transfer heat into the system we must have a different thermal reservoir. This one must
have Tsurr ≥ 1000 K. Assume here that the heat transfer was from a reservoir held at 1000 K to assess
the influence of the second law.
q23
s3 − s2 ≥ ? (9.90)
T
J
J 294000 kg
489 ≥ , (9.91)
kg K 1000 K
J
≥ 294 , ok. (9.92)
kg K

9.1.3 Sound speed

Let us find the sound speed c(T, ρ) for a general material with known thermal equation of
state p(ρ, T ). Later in Sec. 9.2.3, we will see how one can obtain the sound speed from a
general caloric equation of state e(p, ρ). At this point, c is best thought of as a thermody-
namic property. Later in Sec. 9.4.6, we will see how it represents the speed of propagation
of small acoustic disturbances. Let us define c as
s
∂p ∂p
c= , c2 = . (9.93)
∂ρ s ∂ρ s
Use the Gibbs relation, Eq. (4.163):
T ds = de + p dv̂. (9.94)
Now use Eq. (9.17) to eliminate de:
   
∂p
T ds = cv dT + T −p dv̂ +p dv̂, (9.95)
∂T v̂
| {z }
de
∂p
= cv dT + T dv̂, (9.96)
∂T v̂
T ∂p
= cv dT − 2 dρ. (9.97)
ρ ∂T ρ
Because p = p(T, v̂), we can also say p = p(T, ρ), and then we get by calculus
∂p ∂p
dp = dT + dρ, (9.98)
∂T ρ ∂ρ T
∂p
dp − ∂ρ
dρ
T
dT = ∂p
. (9.99)
∂T ρ

© 05 December 2025, Joseph M. Powers. All rights reserved.

352 CHAPTER 9. ONE-DIMENSIONAL COMPRESSIBLE FLOW

Thus substituting for dT in Eq. (9.97), we find

 ∂p

dp − dρ
 − T ∂p
∂ρ
T ds = cv  ∂p
T
dρ. (9.100)
∂T ρ
ρ2 ∂T ρ

Grouping terms in dp and dρ, we get

!  ∂p

cv ∂ρ T ∂p 
T ds = ∂p
dp − cv ∂p
T
+ dρ. (9.101)
∂T ρ ∂T ρ
ρ2 ∂T ρ

For the isentropic sound speed, we must have ds ≡ 0; we thus obtain

 ∂p

∂p 1 ∂p cv
∂ρ T ∂p 
c2 = = ∂p
T
+ , (9.102)
∂ρ s cv ∂T ρ ∂T ρ
ρ2 ∂T ρ
!2
∂p T ∂p
= + . (9.103)
∂ρ T cv ρ2 ∂T ρ

So the isentropic sound speed for a general thermal equation of state is

s v !2
u
∂p u ∂p T ∂p
c(T, ρ) = =t + . (9.104)
∂ρ s ∂ρ T cv ρ2 ∂T ρ

Without the benefit of modern understanding

p of thermodynamics, in 1687 Newton3 ef-
fectively concluded the speed of sound was p ∂p/∂ρ|T . So with benefit of the known Boyle’s4
Law, Newton could predict a sound speed of p/ρ. We might call thatpthe isothermal sound
speed as opposed to the more accurate isentropic sound speed, that is γp/ρ. He then went
to considerable effort to measure the sound speed, but could not reconcile the discrepancy of
his measurements with his isothermal theory. Newton’s approach was corrected by Laplace
in 1816, who generated what amounts to our adiabatic prediction, long before notions of
thermodynamics were settled. Laplace’s notions rested on an uncertain theoretical founda-
tion; he in fact adjusted his theory often, and it was not until thermodynamics was well
established several decades later that our understanding of sound waves was clarified. The
interested reader can consult Finn5 .
3
I. Newton, 1934, Principia, Cajori’s revised translation of Motte’s 1729 translation, U. California Press,
Berkeley.
4
Robert Boyle, 1627-1691, Anglo-Irish natural philosopher and landlord of the author’s County Waterford
ancestors.
5
B. S. Finn, 1964, “Laplace and the speed of sound,” Isis, 55(1): 7-19.

© 05 December 2025, Joseph M. Powers. All rights reserved.

9.1. THERMODYNAMICS OF GENERAL COMPRESSIBLE FLUIDS 353

Example 9.4
Find the sound speed for an ideal gas:
p(T, ρ) = ρRT. (9.105)

The necessary partials are

∂p ∂p
= RT, = ρR. (9.106)
∂ρ T ∂T ρ
so
s
T 2
c(T, ρ) = RT + (ρR) , (9.107)
c v ρ2
s
R2 T
= RT + , (9.108)
cv
s  
R
= RT 1+ , (9.109)
cv
s  
cP − cv
= RT 1 + , (9.110)
cv
s  
cv + cP − cv
= RT , (9.111)
cv
p
= γRT . (9.112)
Sound speed depends on temperature alone for the ideal gas.

Example 9.5
Find the sound speed of a so-called virial gas:
p(T, ρ) = ρRT (1 + bρ) . (9.113)

The necessary partials are

∂p ∂p
= RT + 2bρRT, = ρR (1 + bρ) . (9.114)
∂ρ T ∂T ρ

Thus,
s
T 2
c(T, ρ) = RT + 2bρRT +(ρR (1 + bρ)) , (9.115)
c v ρ2
s  
R 2
= RT 1 + 2bρ + (1 + bρ) . (9.116)
cv
The sound speed of a virial gas depends on both temperature and density.

© 05 December 2025, Joseph M. Powers. All rights reserved.

354 CHAPTER 9. ONE-DIMENSIONAL COMPRESSIBLE FLOW

In Sec. 9.6.1, we shall need to consider p = p(ρ, s), taking advantage of the fact that in
thermodynamics, one can cast any intensive thermodynamic variable in terms of two other
independent intensive thermodynamic variables.

Example 9.6
For a CPIG, find p = p(ρ, s).

Start with the Gibbs equation, Eq. (4.163), T ds = de + p dv̂. For a CPIG, we have de = cv dT ,
p = RT /v̂, so the Gibbs equation reduces to
dT dv̂
ds = cv +R , (9.117)
T v̂
1
dT − ρ2 dρ
= cv +R 1 , (9.118)
T ρ
dT dρ
= cv −R . (9.119)
T ρ
Now because p = ρRT , we also have

dp = ρR dT + RT dρ, (9.120)
dp dT dρ
= + , (9.121)
p T ρ
dT dp dρ
= − . (9.122)
T p ρ
Substitute this into Eq. (9.119) to get
 
dp dρ dρ
ds = cv − −R , (9.123)
p ρ ρ
dp dρ
= cv − (cv + R) , (9.124)
p ρ
dp dρ
= cv − cp , (9.125)
p ρ
ds dp dρ
= −γ , (9.126)
cv p ρ
s − so p ρ
= ln − γ ln , (9.127)
cv po ρo
 γ
p ρ s − so
ln = ln + , (9.128)
po ρo cv
 γ  
ρ s − so
p(ρ, s) = po exp . (9.129)
ρo cv

© 05 December 2025, Joseph M. Powers. All rights reserved.

9.1. THERMODYNAMICS OF GENERAL COMPRESSIBLE FLUIDS 355

9.1.4 Adiabatic gamma

Here we consider the so-called adiabatic gamma, γa . It is closely related to the ratio of
specific heats γ, but appeals to energy conservation. In this discussion, we build on the
analysis of Davis.6
Let us begin with the Gibbs equation, Eq. (4.163): T ds = de + p dv̂, and take it in the
isentropic limit for which ds = 0:
de + p dv̂ = 0. (9.130)
For a general caloric equation of state, we can expect
e = e(p, v̂). (9.131)
Thus
∂e ∂e
de = dp + dv̂. (9.132)
∂p v̂ ∂v̂ p

Substitute Eq. (9.132) into Eq. (9.130) so as to eliminate de to get

∂e ∂e
dp + dv̂ + p dv̂ = 0, (9.133)
∂p v̂ ∂v̂ p
∂e
∂ v̂ p
+p
dp + dv̂ = 0, (9.134)
∂e
∂p
 v̂
∂e
v̂ +p p
∂ v̂ p
dp + dv̂ = 0. (9.135)
∂e
p ∂p v̂
| {z v̂ }
≡γa

We now define the adiabatic gamma, γa , as

 
∂e
v̂ ∂ v̂ p
+p
γa (p, v̂) ≡ . (9.136)
∂e
p ∂p
v̂

For a general material, γa is a thermodynamic variable that is a function of two independent

variables, e.g. γa = γa (p, v̂). It is easy to show for a CPIG that γa = cp /cv and is a constant.
So for a CPIG, γa = γ, the ratio of specific heats.
With this definition of γa , the first law of thermodynamics, Eq. (9.135), becomes
p
dp + γa dv̂ = 0. (9.137)
v̂
6
W. C. Davis, 1985, “Equation of state for detonation products,” Eighth International Detonation Sym-
posium, ed. J. Short, Naval Surface Weapons Center, White Oak, Silver Spring, MD, pp. 785-795.

© 05 December 2025, Joseph M. Powers. All rights reserved.

356 CHAPTER 9. ONE-DIMENSIONAL COMPRESSIBLE FLOW

We could also say

dp dv̂
= −γa (p, v̂) , (9.138)
p v̂
Z
p dv̂
ln = − γa (p, v̂) . (9.139)
po v̂

This is as far as we can go for a general equation of state with γa = γa (p, v̂). However, in the
case that γa is a constant, which is the case for a CPIG, we get ln(p/po ) = −γa ln(v̂/v̂o ) =
ln(v̂o /v̂)γa . This gives pv̂ γa = po v̂oγa , our polytropic relation for a CPIG.
Recall that Eq. (9.137) is an isentropic relation by our assumptions. So we can rearrange
that equation to say

v̂ ∂p ∂ ln p ρ ∂p ∂ ln p
γa (p, v̂) = − =− = = . (9.140)
p ∂v̂ s ∂ ln v̂ s p ∂ρ s ∂ ln ρ s

Also, because v̂ = 1/ρ, dv̂ = −1/ρ2 dρ, so Eq. (9.137) is restated as

p
dp − γa dρ = 0, (9.141)
ρ

giving

∂p p
= γa = c2 . (9.142)
∂ρ s ρ

We will see that c2 is the adiabatic sound speed. With the equation of state for e as
e = e(v, s), it is possible to develop the relations

∂e ∂e
= −p, = T. (9.143)
∂v̂ s ∂s v̂

Using these, we can say

∂2e
v̂ ∂ 2 e ∂v2
s
γa = = −v̂ ∂e
. (9.144)
p ∂v̂ 2 s ∂ v̂ s

For non-ideal gases, it can be shown that the ratio of specific heats γ must be

cp ∂v̂ ∂p
γ= = . (9.145)
cv ∂p T ∂v̂ s

In order for γ = γa , we must have −v̂/p = ∂v̂/∂p|T . This is the case for ideal gases, but not
for general non-ideal gases.

© 05 December 2025, Joseph M. Powers. All rights reserved.

9.2. GENERALIZED ONE-DIMENSIONAL EQUATIONS 357

9.2 Generalized one-dimensional equations

Now let us introduce the conservation principles of mass, momentum, and energy to our
thermodynamics to better understand the dynamics of compressible flow. Here we will
derive in a conventional way the one-dimensional equations of flow with area change. The
development is guided by Shapiro (1953), Ch. 8. Although for the geometry we use, it
will appear that we should be using at least two-dimensional equations, our results will be
approximately correct when we interpret them as an average value at a given x location.
Our results will be valid as long as the area changes slowly relative to how fast the flow can
adjust to area changes.
We can be guided by our equations from Ch. 6.2. However, the ad hoc nature of friction
and heat transfer commonly we will here employ makes a fresh derivation essential. The flow
we wish to consider, flow with area change, heat transfer, and wall friction, is illustrated by
the following sketch of a control volume, Fig. 9.1.

Figure 9.1: Control volume sketch for one-dimensional compressible flow with area change,
heat transfer, and wall friction.

For this flow, we will adopt the following conventions

• surface 1 and 2 are open and allow fluxes of mass, momentum, and energy,

• surface w is a closed wall; no mass flux through the wall is allowed,

• external heat flux qw (energy/area/time: W/m2 ) through the wall is allowed: qw is a

known, fixed parameter,

• diffusive, longitudinal heat transfer is ignored, qx = 0; thus, thermal conductivity

k = 0,

• wall shear τw (force/area: N/m2 ) is allowed: τw is a known, fixed parameter,

© 05 December 2025, Joseph M. Powers. All rights reserved.

358 CHAPTER 9. ONE-DIMENSIONAL COMPRESSIBLE FLOW

• diffusive viscous stress is not allowed, τxx = 0; thus, viscosity µ = 0, and

• cross-sectional area is a known, fixed function: A(x).

9.2.1 Mass
Take the over-bar notation to indicate a volume-averaged quantity. The amount of mass in
a control volume after a time increment ∆t is equal to the original amount of mass plus that
which came in minus that which left:

ρ̄Ā∆x t+∆t
= ρ̄Ā∆x t + ρ1 A1 (u1 ∆t) − ρ2 A2 (u2 ∆t) . (9.146)

Rearrange and divide by ∆x∆t:

ρ̄Ā t+∆t
− ρ̄Ā t ρ2 A2 u2 − ρ1 A1 u1
+ = 0. (9.147)
∆t ∆x

Taking the limit as ∆t → 0, ∆x → 0, we get

∂ ∂
(ρA) + (ρAu) = 0. (9.148)
∂t ∂x

If the flow is steady, then

d
(ρAu) = 0, (9.149)
dx
dρ dA du
Au + ρu + ρA = 0, (9.150)
dx dx dx
1 dρ 1 dA 1 du
+ + = 0. (9.151)
ρ dx A dx u dx

Now integrate Eq. (9.149) from x1 to x2 to get

Z x2 Z x2
d
(ρAu) dx = 0 dx, (9.152)
x1 dx x1
Z 2
d (ρAu) = 0, (9.153)
1
ρ2 u2 A2 − ρ1 u1 A1 = 0, (9.154)
ρ2 u2 A2 = ρ1 u1 A1 ≡ ṁ. (9.155)

Here ṁ is the mass flux with units kg/s. For steady flow, it is a constant.

© 05 December 2025, Joseph M. Powers. All rights reserved.

9.2. GENERALIZED ONE-DIMENSIONAL EQUATIONS 359

9.2.2 Linear momentum

Newton’s second law of motion says the time rate of change of linear momentum of a body
equals the sum of the forces acting on the body. In the x direction this is roughly as follows,
for a system:
d X
(mu) = Fx . (9.156)
dt
In discrete form this would be
mu|t+∆t − mu|t X
= Fx , (9.157)
∆t X 
mu|t+∆t = mu|t + Fx ∆t. (9.158)

For a control volume containing fluid, we must also account for the momentum that
enters and leaves the control volume. The amount of momentum in a control volume after
a time increment ∆t is equal to the original amount of momentum plus that which came in
minus that which left plus that introduced by the forces acting on the control volume. Note
that the
• pressure force at surface 1 pushes the fluid,

• pressure force at surface 2 restrains the fluid,

• force due to the reaction of the wall to the pressure force pushes the fluid if the area
change is positive, and

• force due to the reaction of the wall to the shear force restrains the fluid.
We write the linear momentum principle as



ρ̄Ā∆x ū t+∆t = ρ̄Ā∆x ū t
+ (ρ1 A1 (u1 ∆t)) u1
− (ρ2 A2 (u2 ∆t)) u2
+ (p1 A1 ) ∆t − (p2 A2 ) ∆t
+ (p̄ (A2 − A1 )) ∆t


− τw L¯p ∆x ∆t. (9.159)

Rearrange and divide by ∆x∆t to get

ρ̄Āū − ρ̄Āū ρ2 A2 u22 − ρ1 A1 u21 p2 A2 − p1 A1 A2 − A1

t+∆t t
+ =− + p̄ − τw L¯p . (9.160)
∆t ∆x ∆x ∆x
In the limit ∆x → 0, ∆t → 0 we get
∂ ∂
∂ ∂A
(ρAu) + ρAu2 = − (pA) + p − τw Lp . (9.161)
∂t ∂x ∂x ∂x

© 05 December 2025, Joseph M. Powers. All rights reserved.

360 CHAPTER 9. ONE-DIMENSIONAL COMPRESSIBLE FLOW

In steady state, we find

d
d dA
ρAu2 = − (pA) + p − τw Lp , (9.162)
dx dx dx
du d dA dp dA
ρAu + u (ρAu) = −p −A +p − τw Lp , (9.163)
dx |dx {z } dx dx dx
=0
du dp Lp
ρu = − − τw , (9.164)
dx dx A
Lp
ρu du + dp = −τw dx, (9.165)
A
1 Lp
du + dp = −τw dx, (9.166)
ρu ṁ
 2
u Lp
ρd + dp = −τw dx. (9.167)
2 A
If there is no wall shear, then Eq. (9.167) reduces to
 2
u
dp = −ρ d , no wall shear. (9.168)
2
An increase in velocity magnitude decreases the pressure. We can equivalently rewrite
Eq. (9.165) for τw = 0 as
du dp
ρu + = 0, no wall shear. (9.169)
dx dx
For flow with no area change, dA/dx = 0. For that limit, along with the limit of τw = 0,
Eq. (9.162) reduces to
d
d
ρAu2 = − (pA) , (9.170)
dx dx
d

A ρu2 + p = 0, (9.171)
dx
d

ρu2 + p = 0, (9.172)
dx
ρu2 + p = ρ1 u21 + p1 , no wall shear, no area change. (9.173)

9.2.3 Energy
The first law of thermodynamics states that the change of total energy of a body equals the
heat transferred to the body minus the work done by the body:

E2 − E1 = Q − W, (9.174)
E2 = E1 + Q − W. (9.175)

© 05 December 2025, Joseph M. Powers. All rights reserved.

9.2. GENERALIZED ONE-DIMENSIONAL EQUATIONS 361

So for our control volume this becomes the following when we also account for the energy
flux in and out of the control volume in addition to the work and heat transfer:
   

ū2
ū2
ρ̄Ā∆x ē + = ρ̄Ā∆x ē +
2 t+∆t 2 t
   
u21 u22
+ρ1 A1 (u1 ∆t) e1 + − ρ2 A2 (u2 ∆t) e2 +
2 2


¯
+qw Lp ∆x ∆t + (p1 A1 ) (u1 ∆t) − (p2 A2 ) (u2 ∆t) . (9.176)
Note:
• the mean pressure times area difference does no work because it is acting on a stationary
boundary, and
• the work done by the wall shear force is not included.7
Rearrange and divide by ∆t∆x:
       
2 2 u22 u21
ρ̄Ā ē + ū2 − ρ̄Ā ē + ū2 ρ2 A2 u2 e2 + 2
+ p2
ρ2
− ρ1 A1 u1 e1 + 2
+ p1
ρ1
t+∆t t
+
∆t ∆x
= qw L̄p .
(9.177)
In differential form as ∆x → 0, ∆t → 0
     
∂ u2 ∂ u2 p
ρA e + + ρAu e + + = qw Lp . (9.178)
∂t 2 ∂x 2 ρ
In steady state:
  
d u2 p
ρAu e + + = qw Lp , (9.179)
dx 2 ρ
   
d u2 p u2 p d
ρAu e+ + + e+ + (ρAu) = qw Lp , (9.180)
dx 2 ρ 2 ρ |dx {z }
=0
 
d u2 p qw Lp
ρu e+ + = , (9.181)
dx 2 ρ A
 
de du 1 dp p dρ qw Lp
ρu +u + − 2 = . (9.182)
dx dx ρ dx ρ dx A
7
In neglecting work done by the wall shear force, I have taken an approach that is nearly universal, but
fundamentally difficult to defend. At this stage of the development of these notes, I am not ready to enter
into a grand battle with all established authors and probably confuse the student; consequently, results for
flow with friction will be consistent with those of other sources. The argument typically used to justify this
is that the real fluid satisfies no-slip at the boundary; thus, the wall shear actually does no work. However,
one can easily argue that within the context of the one-dimensional model that has been posed that the
shear force behaves as an external force that reduces the fluid’s mechanical energy. Moreover, it is possible
to show that neglect of this term results in the loss of frame invariance,


a serious defect indeed. To model
the work of the wall shear, one would include the term τw L¯p ∆x (ū∆t) in the energy equation.

© 05 December 2025, Joseph M. Powers. All rights reserved.

362 CHAPTER 9. ONE-DIMENSIONAL COMPRESSIBLE FLOW

Now consider the product of velocity and momentum from Eq. (9.164) to get an equation
for the mechanical energy:

du dp τw Lp u
ρu2 +u =− . (9.183)
dx dx A
Subtract this, the mechanical energy, from Eq. (9.182) to get an equation for the thermal
energy

de pu dρ qw Lp τw Lp u
ρu − = + , (9.184)
dx ρ dx A A
de p dρ (qw + τw u) Lp
− 2 = . (9.185)
dx ρ dx ṁ

Because e = e(p, ρ), we have

∂e ∂e
de = dρ + dp, (9.186)
∂ρ p ∂p ρ
de ∂e dρ ∂e dp
= + . (9.187)
dx ∂ρ p dx ∂p ρ dx

so the steady energy equation becomes

∂e dρ ∂e dp p dρ (qw + τw u) Lp
+ − 2 = , (9.188)
∂ρ p dx ∂p ρ dx ρ dx ṁ
| {z }
de
dx
 
p ∂e
−
dp ρ
2 ∂ρ
p dρ (qw + τw u) Lp
−  = . (9.189)
dx ∂e dx ∂e
ṁ ∂p
∂p
ρ ρ
| {z }
≡c2

Now let us consider the term in braces, that we label c2 , in the previous equation. It will
be seen to be the square of the sound speed. We can put that term in a more common form
by considering the Gibbs equation, Eq. (4.164):
p
T ds = de − dρ, (9.190)
ρ2

along with a general caloric equation of state e = e(p, ρ), from which we get

∂e ∂e
de = dp + dρ. (9.191)
∂p ρ ∂ρ p

© 05 December 2025, Joseph M. Powers. All rights reserved.

9.2. GENERALIZED ONE-DIMENSIONAL EQUATIONS 363

Substituting into the Gibbs equation, we get

∂e ∂e p
T ds = dp + dρ − 2 dρ. (9.192)
∂p ρ ∂ρ p ρ
| {z }
de

Taking s to be constant and dividing by dρ, we get

∂e ∂p ∂e p
0= + − . (9.193)
∂p ρ ∂ρ s ∂ρ p ρ2

Rearranging, we get
p ∂e
ρ2
−
∂p ∂ρ
p
= c2 = , (9.194)
∂ρ s
∂e
∂p
ρ

so
dp dρ (qw + τw u) Lp
− c2 = , (9.195)
dx dx ∂e
ṁ ∂p
ρ
dp dρ (qw + τw u) Lp
− c2 = . (9.196)
dx dx ∂e
ρuA ∂p
ρ

Here c is the isentropic sound speed, a thermodynamic property of the material. We shall see
later in Sec. 9.4.6 why it is appropriate to interpret this property as the propagation speed
of small disturbances. At this point, it should simply be thought of as a state property.

Example 9.7
Find the speed of sound for a CPIG.

For a CPIG, we have

p = ρRT, e = cv T + e o , cp − cv = R. (9.197)

So
p
e = cv + eo , (9.198)
ρR
p
= cv + eo , (9.199)
(cp − cv )ρ
1 p
= cp + eo , (9.200)
cv − 1
ρ
1 p
= + eo . (9.201)
γ−1ρ

© 05 December 2025, Joseph M. Powers. All rights reserved.

364 CHAPTER 9. ONE-DIMENSIONAL COMPRESSIBLE FLOW

The relevant partial derivatives are

∂e 1 p ∂e 1 1
=− , = . (9.202)
∂ρ p γ − 1 ρ2 ∂p ρ γ−1ρ

So by Eq. (9.194), we have

 
p 1 p
ρ2 − − γ−1 ρ2
c2 = 1 1 , (9.203)
γ−1 ρ
(γ − 1) ρp2 + ρp2
= 1 , (9.204)
ρ
p
= γ , (9.205)
ρ
ρRT
= γ , (9.206)
ρ
= γRT, (9.207)
p
c = γRT . (9.208)

Example 9.8
For qw = 0, τw = 0, find a relation between p and ρ for the steady flow of a CPIG.

Start with the energy equation, Eq. (9.196), in the limit as qw = 0, τw = 0:

dp dρ
= c2 . (9.209)
dx dx
Now from the previous example, we know for a CPIG that c2 = γp/ρ, so

dp p dρ
= γ , (9.210)
dx ρ dx
1 dp 1 dρ
= γ , (9.211)
p dx ρ dx
dp dρ
= γ , (9.212)
p ρ
p ρ
ln = γ ln , (9.213)
po ρo
 γ
ρ
= ln , (9.214)
ρo
 γ
p ρ
= , (9.215)
po ρo
p po
= . (9.216)
ργ ργo

© 05 December 2025, Joseph M. Powers. All rights reserved.

9.2. GENERALIZED ONE-DIMENSIONAL EQUATIONS 365

This is equivalent to what we have earlier derived in Eq. (6.133). In terms of specific volume, we could
say
pv̂ γ = po v̂oγ . (9.217)
This is the equation for a so-called polytropic process in which the polytropic exponent is γ.

Consider now the special case of flow with no heat transfer qw ≡ 0. We still allow area
change and wall friction is allowed (see earlier footnote, p. 361):
 
d u2 p
ρu e+ + = 0, (9.218)
dx 2 ρ
u2 p u21 p1
e+ + = e1 + + , (9.219)
2 ρ 2 ρ1
u2 u21
h+ = h1 + . (9.220)
2 2

9.2.4 Summary of equations

We can summarize the one-dimensional compressible flow equations in various forms here.
In the equations below, we assume A(x), τw , qw , and Lp are all known.

[Link] Unsteady conservative form

∂ ∂
(ρA) + (ρAu) = 0, (9.221)
∂t ∂x
∂ ∂
∂A
(ρAu) + ρAu2 + pA = p − τw Lp , (9.222)
   ∂t  ∂x  ∂x
∂ u2 ∂ u2 p
ρA e + + ρAu e + + = qw Lp , (9.223)
∂t 2 ∂x 2 ρ
e = e(ρ, p), (9.224)
p = p(ρ, T ). (9.225)

[Link] Unsteady non-conservative form

dρ ρ ∂
= − (Au), (9.226)
dt A ∂x
du ∂p τw Lp
ρ = − − , (9.227)
dt ∂x A
de p dρ (qw + τw u)Lp
ρ − = , (9.228)
dt ρ dt A
e = e(ρ, p), (9.229)
p = p(ρ, T ). (9.230)

© 05 December 2025, Joseph M. Powers. All rights reserved.

366 CHAPTER 9. ONE-DIMENSIONAL COMPRESSIBLE FLOW

[Link] Steady conservative form

d
(ρAu) = 0, (9.231)
dx
d
dA
ρAu2 + pA = p − τw Lp , (9.232)
 dx
  dx
d u2 p
ρAu e + + = qw Lp , (9.233)
dx 2 ρ
e = e(ρ, p), (9.234)
p = p(ρ, T ). (9.235)

[Link] Steady non-conservative form

dρ ρ d
u = − (Au), (9.236)
dx A dx
du dp τw Lp
ρu = − + , (9.237)
dx dx A
de pu dρ (qw + τw u)Lp
ρu − = , (9.238)
dx ρ dx A
e = e(ρ, p), (9.239)
p = p(ρ, T ). (9.240)

In whatever form we consider, we have five equations in five unknown dependent variables:
ρ, u, p, e, and T . We can always use the thermal and caloric state equations to eliminate e
and T to give rise to three equations in three unknowns.

Example 9.9
Let us consider the flow of air with heat addition.
Given: Air initially at p1 = 100 kPa, T1 = 300 K, u1 = 10 m/s flows in a duct of length 100 m.
The duct has a constant circular cross sectional area of A = 0.02 m2 and is isobarically heated with
a constant heat flux qw along the entire surface of the duct. At the end of the duct the flow has
p2 = 100 kPa, T2 = 500 K.
Find: the mass flow rate ṁ, the wall heat flux qw and the entropy change s2 − s1 ; check for
satisfaction of the second law.
Assume: CPIG, R = 0.287 kJ/(kg K), cp = 1.0035 kJ/(kg K).

We begin by considering the geometry.

A = πr2 , (9.241)
r
A
r = , (9.242)
π
√ p
Lp = 2πr = 2 πA = 2 π (0.02 m2 ) = 0.501 m. (9.243)

© 05 December 2025, Joseph M. Powers. All rights reserved.

9.2. GENERALIZED ONE-DIMENSIONAL EQUATIONS 367

Now get the mass flux.

p1 = ρ1 RT1 , (9.244)
p1 100 kPa
ρ1 = =   , (9.245)
RT1 0.287 kgkJK (300 K)
kg
= 1.161 . (9.246)
m3
So
 
kg  m 
kg
ṁ = ρ1 u1 A1 = 1.161 3 10 0.02 m2 = 0.2322 . (9.247)
m s s
Now get the flow variables at state 2:
p2 100 kPa
ρ2 = ,=   , (9.248)
RT2 0.287 kgkJK (500 K)
kg
= 0.6969 , (9.249)
m3
ρ2 u2 A2 = ρ1 u1 A1 , (9.250)
ρ1 u1 A1 ρ1 u 1
u2 = = , (9.251)
ρ A ρ
 2 2  2

kg
1.161 m 3 10 ms m
= kg
= 16.67 . (9.252)
0.6969 m3
s
Now consider the energy equation:
 
d u2 p qw Lp
ρu e+ + = , (9.253)
dx 2 ρ A
 
d u2 qw Lp
h+ = , (9.254)
dx 2 ṁ
Z L  2
 Z L
d u qw Lp
h+ dx = dx, (9.255)
0 dx 2 0 ṁ
u2 u2 qw LLp
h2 + 2 − h1 − 1 = , (9.256)
2 2 ṁ
u22 u21 qw LLp
cp (T2 − T1 ) + − = . (9.257)
2 2 ṁ
Solving for qw , we get
  
ṁ u22 u21
qw = cp (T2 − T1 ) + − , (9.258)
LLp 2 2
!  

!
m 2 m 2
0.2322 kgs J 16.67 s 10 s
= 1003.5 (500 K − 300 K) + − ,(9.259)
(100 m) (0.501 m) kg K 2 2
 
kg J m2
= 0.004635 2 200700 − 88.9 2 , (9.260)
m s kg s
 
kg J J
= 0.004635 2 200700 − 88.9 , (9.261)
m s kg kg
W
= 930 2 . (9.262)
m

© 05 December 2025, Joseph M. Powers. All rights reserved.

368 CHAPTER 9. ONE-DIMENSIONAL COMPRESSIBLE FLOW

The heat flux is positive, that indicates a transfer of thermal energy into the air.
Now find the entropy change.
   
T2 p2
s2 − s1 = cp ln − R ln , (9.263)
T1 p1
       
J 500 K J 100 kPa
= 1003.5 ln − 287 ln , (9.264)
kg K 300 K kg K 100 kPa
J
= 512.6 − 0 = 512.6 . (9.265)
kg K
Is the second law satisfied? Assume the heat transfer takes place from a reservoir held at 500 K. The
reservoir would have to be at least at 500 K in order to bring the fluid to its final state of 500 K. It
could be greater than 500 K and still satisfy the second law.
Q12
S2 − S1 ≥ , (9.266)
T
Q̇12
Ṡ2 − Ṡ1 ≥ , (9.267)
T
Q̇12
ṁ (s2 − s1 ) ≥ , (9.268)
T
qw Atot
≥ , (9.269)
T
qw LLp
≥ , (9.270)
T
qw LLp
s2 − s1 ≥ , (9.271)
ṁT

J 930 s Jm2 (100 m) (0.501 m)
512.6 ≥   , (9.272)
kg K 0.2322 kg (500 K)
s
J
≥ 401.3 . (9.273)
kg K

9.2.5 Influence coefficients

Now, let us uncouple the steady one-dimensional equations. First let us summarize again,
in a slightly different manner than before, by rearranging the mass, Eq. (9.150), momentum,
Eq. (9.164), and energy, Eq. (9.196), equations:
dρ du ρu dA
u +ρ = − , (9.274)
dx dx A dx
du dp τw Lp
ρu + = − , (9.275)
dx dx A
dp dρ (qw + τw u) Lp
− c2 = . (9.276)
dx dx ρuA ∂p∂e
ρ

© 05 December 2025, Joseph M. Powers. All rights reserved.

9.2. GENERALIZED ONE-DIMENSIONAL EQUATIONS 369

In matrix form, these are recast as

   dρ   − ρu dA 
u ρ 0 dx
A dx
 0   du   − τw Lp 
ρu 1 dx = A . (9.277)
dp (qw +τw u)Lp
−c2 0 1 dx ρuA ∂e | ∂p ρ

Use Cramer’s rule to solve for the derivatives. First calculate the determinant of the coeffi-
cient matrix:



u ((ρu)(1) − (1)(0)) − ρ (0)(1) − (−c2 )(1) = ρ u2 − c2 . (9.278)

Implementing Cramer’s rule, we find after detailed calculation:

   
ρu dA

τw Lp (qw +τw u)Lp
ρu − A dx − ρ − A + ρ
dρ ∂e
ρuA ∂p |ρ
= , (9.279)
dx ρ (u2 − c2 )
   
2 ρu dA

τw Lp (qw +τw u)Lp
−c − A dx + u − A − u
du ∂e
ρuA ∂p |ρ
= , (9.280)
dx ρ (u2 − c2 )
   
2 ρu dA

2 τw Lp 2 (qw +τw u)Lp
ρuc − A dx − ρc − A + ρu
dp ρuA ∂p∂e
|ρ
= . (9.281)
dx ρ (u2 − c2 )

Simplify to find
(qw +τw u)Lp
−ρu2 dA + τw Lp +
dρ 1 dx ∂e
ρu ∂p |
ρ
= , (9.282)
dx A u 2 − c2
(qw +τw u)Lp
c2 ρu dA − uτw Lp −
du 1 dx ∂e
ρ ∂p
ρ
|
= , (9.283)
dx A ρ (u2 − c2 )
(qw +τw u)Lp u
−c2 ρu2 dA + c2 τw Lp +
dp 1 dx ∂e
ρ ∂p | ρ
= . (9.284)
dx A u 2 − c2
We have

• a system of coupled non-linear ordinary differential equations,

• in standard form for dynamic system analysis: du/dx = f(u),

• valid for general equations of state, and

• singular when the fluid particle velocity is sonic u = c.

© 05 December 2025, Joseph M. Powers. All rights reserved.

370 CHAPTER 9. ONE-DIMENSIONAL COMPRESSIBLE FLOW

9.3 Flow with area change

This section will consider flow with area change with an emphasis on isentropic flow. Some
problems will involve non-isentropic flow but a detailed discussion of such flows will be
delayed.

9.3.1 Isentropic Mach number relations

Take the special case of
• τw = 0,
• qw = 0,
• CPIG.
Then we can recast Eqs. (9.231-9.233) as
d
(ρuA) = 0, (9.285)
dx
du dp
ρu + = 0, (9.286)
 dx dx
d u2 p
e+ + = 0. (9.287)
dx 2 ρ
Integrate the energy equation using Eq. (4.136), h = e + p/ρ, to get
u2 u2o
h+ = ho + . (9.288)
2 2
If we define the “o” condition to be a condition of rest, then uo ≡ 0. This is a stagnation
condition, as introduced on p. 101 and p. 220. In an unfortunate choice of nomenclature,
properties evaluated for the fluid in motion are named static properties. So
u2
h+ = ho , (9.289)
2
u2
(h − ho ) + = 0. (9.290)
2
Because we have a CPIG,
u2
cp (T − To ) + = 0, (9.291)
2
u2
T − To + = 0, (9.292)
2cp
To u2
1− + = 0. (9.293)
T 2cp T

© 05 December 2025, Joseph M. Powers. All rights reserved.

9.3. FLOW WITH AREA CHANGE 371

Now note that

cp − cv cp cp − cv γR
cp = cp = cp = , (9.294)
cp − cv cv c v − 1 γ−1
so
To γ − 1 u2
1− + = 0, (9.295)
T 2 γRT
To γ − 1 u2
= 1+ . (9.296)
T 2 γRT
Recall the sound speed and Mach number for a CPIG:
c2 = γRT, if p = ρRT, e = cv T + ê, (9.297)
 u 2
M2 ≡ . (9.298)
c
Thus, the ratio of stagnation temperature To to static temperature T is
To γ −1 2
= 1+ M , (9.299)
T 2
 −1
T γ−1 2
= 1+ M . (9.300)
To 2
Now if the flow is isentropic and for a CPIG, it can be inferred from Eq. (9.215) that ratios
of static to stagnation values are
 γ−1   γ−1
T ρ p γ
= = . (9.301)
To ρo po
Thus,
 − γ−1
1
ρ γ−1 2
= 1+ M , (9.302)
ρo 2
 − γ−1
γ
p γ−1 2
= 1+ M . (9.303)
po 2
For air γ = 7/5, so
 −1
T 1 2
= 1+ M , (9.304)
To 5
 − 5
ρ 1 2 2
= 1+ M , (9.305)
ρo 5
 − 7
p 1 2 2
= 1+ M . (9.306)
po 5

© 05 December 2025, Joseph M. Powers. All rights reserved.

372 CHAPTER 9. ONE-DIMENSIONAL COMPRESSIBLE FLOW

Figures 9.2, 9.3, and 9.4 show the variation of T , ρ and p with M 2 for isentropic flow. Other
thermodynamic properties can be determined from these, e.g. the sound speed:

s r  −1/2
c γRT T γ−1 2
= = = 1+ M . (9.307)
co γRTo To 2

T (K)
300

250

200

150

100

50

0 2 4 6 8 10

Figure 9.2: Static temperature versus Mach number squared; calorically perfect ideal gas,
R = 0.287 kJ/(kg K), γ = 7/5, To = 300 K.

p (bar)
1

0.8

0.6

0.4

0.2

0 2 4 6 8 10

Figure 9.3: Static pressure versus Mach number squared; calorically perfect ideal gas, R =
0.287 kJ/(kg K), γ = 7/5, po = 1 bar.

Example 9.10
Show that how the isentropic relation for the ratio of stagnation to static pressures reduces to the
incompressible Bernoulli’s equation in the limit of low Mach number, and show how it deviates from
this as the Mach number rises.

© 05 December 2025, Joseph M. Powers. All rights reserved.

9.3. FLOW WITH AREA CHANGE 373

1.2

0.8

0.6

0.4

0.2

0 2 4 6 8 10

Figure 9.4: Static density versus Mach number squared; calorically perfect ideal gas, R =
0.287 kJ/(kg K), γ = 7/5, ρo = 1.16 kg/m3 .

Begin with the reciprocal of Eq. (9.303):

  γ
po γ − 1 2 γ−1
= 1+ M . (9.308)
p 2
Taylor series of this equation about M = 0 yields
po γ γ
= 1 + M2 + M4 + . . . , (9.309)
p 2 8
 
γ 1
= 1 + M2 1 + M2 + . . . , (9.310)
2 4
2
 
γ ρv 1
= 1+ 1 + M2 + . . . , (9.311)
2 γp 4
2
 
ρv 1 2
= 1+ 1 + M + ... , (9.312)
2p 4
2
 
ρv 1 2
po = p + 1 + M + ... . (9.313)
2 4
For M = 0, we recover the incompressible Bernoulli equation, and the correction at small finite Mach
number to the incompressible Bernoulli equation estimate for stagnation pressure is evident.

Example 9.11
Given: An airplane is flying into still air at u = 200 m/s. The ambient air is at 288 K and 101.3 kPa.
Find: Temperature, pressure, and density at nose of airplane.
Assume: Steady isentropic flow of a CPIG.

Analysis: In the steady wave frame, the ambient conditions are static while the nose conditions are
stagnation.
u u 200 ms
M= = √ = r   = 0.588. (9.314)
c γRT 7 J
5 287 kg K 288 K

© 05 December 2025, Joseph M. Powers. All rights reserved.

374 CHAPTER 9. ONE-DIMENSIONAL COMPRESSIBLE FLOW

so
   
1 1
To = T 1 + M 2 = (288 K) 1 + 0.5882 = 307.9 K, (9.315)
5 5
  52
1 101.3 kPa kg
ρo = ρ 1 + M2 =  
5 = 1.45 , (9.316)
5 0.287 kgkJK (288 K) 1 + 15 0.5882 2 m3
 7   27
1 2 2 1 2
po = p 1+ M = (101.3 kPa) 1 + 0.588 = 128 kPa. (9.317)
5 5

The temperature, pressure, and density all rise in the isentropic process. In this wave frame, the kinetic
energy of the flow is being converted isentropically to thermal energy.

9.3.2 Sonic properties

Let “*” denote a property at the sonic state M 2 ≡ 1. Then we get

 −1
T∗ γ−1 2 2
= 1+ 1 = , (9.318)
To 2 γ+1
 − 1   γ−1
1
ρ∗ γ − 1 2 γ−1 2
= 1+ 1 = , (9.319)
ρo 2 γ+1
 − γ   γ−1
γ
p∗ γ − 1 2 γ−1 2
= 1+ 1 = , (9.320)
po 2 γ+1
 −1/2 r
c∗ γ−1 2 2
= 1+ 1 = , (9.321)
co 2 γ+1
r
p 2γ
u∗ = c∗ = γRT∗ = RTo . (9.322)
γ+1

If the fluid is air, we have γ = 7/5 and

T∗
= 0.8333, (9.323)
To
ρ∗
= 0.6339, (9.324)
ρo
p∗
= 0.5283, (9.325)
po
c∗
= 0.9129. (9.326)
co

© 05 December 2025, Joseph M. Powers. All rights reserved.

9.3. FLOW WITH AREA CHANGE 375

9.3.3 Effect of area change

To understand the effect of area change, the influence of the mass equation must be con-
sidered. So far we have really only looked at energy. In the isentropic limit the mass,
momentum, and energy equations for a CPIG, Eqs. (9.151, 9.169, 9.212), reduce to

dρ du dA
+ + = 0, (9.327)
ρ u A
ρu du + dp = 0, (9.328)
dp dρ
= γ . (9.329)
p ρ

Substitute energy, then mass into momentum:

p
ρu du + γ dρ = 0, (9.330)
ρ
| {z }
dp
p ρ ρ 
ρu du + γ − du − dA = 0, (9.331)
ρ | u {z A }
dρ
 
p 1 1
du + γ − 2 du − dA = 0, (9.332)
ρ u uA
 
γp/ρ p dA
du 1 − 2 = γ , (9.333)
u ρ uA
 
du γp/ρ γp/ρ dA
1− 2 = , (9.334)
u u u2 A
 
du 1 1 dA
1− 2 = , (9.335)
u M M2 A
du
dA
M2 − 1 = , (9.336)
u A
du 1 dA
= 2
. (9.337)
u M −1 A

Figure 9.5 shows the performance of a fluid in a variable area duct. We note

• there is a singularity when M 2 = 1,

• if M 2 = 1, we need dA = 0,

• area minimum necessary to transition from subsonic to supersonic flow,

• it can be shown an area maximum is not relevant.

© 05 December 2025, Joseph M. Powers. All rights reserved.

376 CHAPTER 9. ONE-DIMENSIONAL COMPRESSIBLE FLOW

consider u > 0

subsonic subsonic
diffuser nozzle

dA > 0, M 2 < 1 so dA < 0, M 2 < 1 so

du < 0, flow slows down du > 0, flow speeds up
dp > 0 dp < 0

supersonic supersonic
nozzle diffuser

dA > 0, M 2 > 1 so dA < 0, M 2 > 1 so

du > 0, flow speeds up du < 0, flow slows down
dp < 0 dp > 0

Figure 9.5: Behavior of fluid in sub- and supersonic nozzles and diffusers.

Consider A at a sonic state. From the mass equation:

ρuA = ρ∗ u∗ A∗ , (9.338)
ρuA = ρ∗ c∗ A∗ , (9.339)
√ √
A ρ∗ 1 ρ∗ p 1 ρ∗ γRT∗ γRT
= c∗ = γRT∗ = √ , (9.340)
A∗ ρ u ρ u ρ γRT u
r r
A ρ∗ T∗ 1 ρ∗ ρo T∗ To 1
= = . (9.341)
A∗ ρ T M ρo ρ To T M

Substitute from earlier-developed relations and get

   21 γ−1
γ+1
A 1 2 γ−1 2
= 1+ M . (9.342)
A∗ M γ+1 2

Fig. 9.6 shows the performance of a fluid in a variable area duct.

Note that

• A/A∗ has a minimum value of 1 at M = 1,

• For each A/A∗ > 1, there exist two values of M, and

• A/A∗ → ∞ as M → 0 or M → ∞.

© 05 December 2025, Joseph M. Powers. All rights reserved.

9.3. FLOW WITH AREA CHANGE 377

M
0 0.5 1 1.5 2 2.5 3

Figure 9.6: Area ratio versus Mach number for a calorically perfect ideal gas, R =
0.287 kJ/(kg K), γ = 7/5.

9.3.4 Choking
Consider mass flow rate variation with pressure difference. We have then
• small pressure difference gives small velocity and small mass flow,
• as pressure difference grows, velocity and mass flow rate grow,
• velocity is limited to sonic at a particular duct location,
• this provides fundamental restriction on mass flow rate,
• it can be proven rigorously that sonic condition gives maximum mass flow rate.

ṁmax = ρ∗ u∗ A∗ , (9.343)
  γ−11 r 
2 2γ
if ideal gas: = ρo RTo A∗ , (9.344)
γ+1 γ+1
  γ−11  1/2
2 2 p
= ρo γRTo A∗ , (9.345)
γ+1 γ+1
  21 γ−1
γ+1
2 p
= ρo γRTo A∗ . (9.346)
γ+1
A flow that has a maximum mass flow rate is known as choked flow. Flows will choke at
area minima in a duct.

Example 9.12
Consider an isentropic area change problem with choking.8
8
adopted from White (1986), p. 529, Ex. 9.5.

© 05 December 2025, Joseph M. Powers. All rights reserved.

378 CHAPTER 9. ONE-DIMENSIONAL COMPRESSIBLE FLOW

Given: Air with stagnation conditions po = 200 kPa, To = 500 K flows through a throat to an exit
Mach number of 2.5. The desired mass flow is 3.0 kg/s.
Find: a) throat area, b) exit pressure, c) exit temperature, d) exit velocity, and e) exit area.
Assume: CPIG, isentropic flow, γ = 7/5.

First find the stagnation density via the ideal gas law:

po 200 kPa kg
ρo = =  = 1.394 3 . (9.347)
RTo kJ
0.287 kg K (500 K) m

Because it necessarily flows through a sonic throat:

  12 γ−1
γ+1
2 p
ṁmax = ρo γRTo A∗ , (9.348)
γ+1
ṁmax
A∗ =   21 γ−1
γ+1 , (9.349)
2 √
ρo γ+1 γRTo
3 kg
A∗ =   rs   = 0.008297 m2 . (9.350)
kg J
1.394 m3 (0.5787) 1.4 287 kg K (500 K)

Because we know Me , use isentropic relations to find other exit conditions.

 − γ−1
γ  −3.5
γ−1 2 1
pe = po 1 + Me = (200 kPa) 1 + 2.52 = 11.71 kPa, (9.351)
2 5
 −1  −1
γ−1 2 1 2
Te = To 1 + Me = (500 K) 1 + 2.5 = 222.2 K. (9.352)
2 5

Note
pe 11.71 kPa kg
ρe = =  = 0.1834 3 . (9.353)
RTe 0.287 kgkJK (222.2 K) m

Now the exit velocity is simply

s  
p J m
ue = Me ce = Me γRTe = 2.5 1.4 287 (222.2 K) = 747.0 . (9.354)
kg K s

Now determine the exit area.

   12 γ−1
γ+1
A∗ γ−1 2
2
A = 1+ Me , (9.355)
Me γ+1
2
  3
0.008297 m2 5 1
= 1 + 2.52 = 0.0219 m2 . (9.356)
2.5 6 5

© 05 December 2025, Joseph M. Powers. All rights reserved.

9.4. NORMAL SHOCK WAVES 379

9.4 Normal shock waves

This section will develop relations for normal shock waves in fluids with general equations of
state. It will be specialized to a CPIG to illustrate common features of the waves. Assume
for this section we have
• one-dimensional flow,

• steady flow,

• no area change,

• viscous effects and wall friction do not have time to influence flow, and

• heat conduction and wall heat transfer do not have time to influence flow.
We will consider the problem in the context of the piston problem as sketched in Fig. 9.7.
The physical problem is as follows:

laboratory frame steady wave frame

Figure 9.7: Normal shock sketch.

• drive a piston with known velocity ûp into a fluid at rest (û1 = 0) with known properties,
p1 , ρ1 in the x̂ laboratory frame,

• determine the disturbance speed U,

• determine the disturbance properties û2 , p2 , ρ2 ,

• recognize that in this frame of reference we have an unsteady problem.

The transformed problem is
• use a Galilean transformation x = x̂ − Ut, u = û − U to transform to the frame in
which the wave is at rest, therefore rending the problem steady in this frame,

© 05 December 2025, Joseph M. Powers. All rights reserved.

380 CHAPTER 9. ONE-DIMENSIONAL COMPRESSIBLE FLOW

• solve as though U is known to get downstream “2” conditions: u2 (U), p2 (U), ...,
• invert to solve for U as function of u2 , the transformed piston velocity: U(u2 ),
• back transform to get all variables as function of û2 , the laboratory piston velocity:
U(û2 ), p2 (û2 ), ρ2 (û2 ), ....

9.4.1 Rankine-Hugoniot equations

Under these assumptions, we can recast the conservation principles and equation of state in
the steady frame, Eqs.(9.231-9.235), as follows:
d
(ρu) = 0, (9.357)
dx
d

ρu2 + p = 0, (9.358)
 dx
 
d u2
ρu h + = 0, (9.359)
dx 2
h = h(p, ρ). (9.360)
Upstream conditions are ρ = ρ1 , p = p1 , u = −U. With knowledge of the equation of state,
we get h = h1 . In what is a natural, but in fact naı̈ve, step we can integrate the equations
from upstream to state “2” to give the correct Rankine-Hugoniot jump equations.910
ρ2 u2 = −ρ1 U, (9.361)
ρ2 u22
+ p2 = ρ1 U 2 + p1 , (9.362)
u2 U2
h2 + 2 = h1 + , (9.363)
2 2
h2 = h(p2 , ρ2 ). (9.364)
This analysis is straightforward and yields the correct result. In actuality, however, the
analysis should be more nuanced. We are going to solve these algebraic equations to arrive at
discontinuous shock jumps. Thus, we should be concerned about the validity of of differential
equations in the vicinity of a discontinuity.
As described by LeVeque (1992), the proper way to arrive at the shock jump equations
is to use a more primitive form of the conservation laws, expressed in terms of integrals of
conserved quantities balanced by fluxes of those quantities, e.g. Eq. (4.4). If q is a set of
conserved variables, and f(q) is the flux of q (e.g. for mass conservation, ρ is a conserved
variable and ρu is the flux), then the primitive form of the conservation law can be written
as
Z
d x2
q(x, t) dx = f(q(x1 , t)) − f(q(x2 , t)). (9.365)
dt x1
9
William John Macquorn Rankine, 1820-1872, Scottish engineer and mechanician, pioneer of thermody-
namics and steam engine theory, taught at University of Glasgow, studied fatigue in railway engine axles.
10
Pierre Henri Hugoniot, 1851-1887, French engineer.

© 05 December 2025, Joseph M. Powers. All rights reserved.

9.4. NORMAL SHOCK WAVES 381

Here we have considered flow into and out of a one-dimensional box for x ∈ [x1 , x2 ]. For the
Euler equations, we have
   
ρ ρu
2
q= ρu
 , f(q) =   ρu + p . (9.366)
1 2 1 2 p
ρ e + 2u ρu e + 2 u + ρ

If we assume there is a discontinuity in the region x ∈ [x1 , x2 ] propagating at speed U, we

can break up the integral into the form
Z − Z
d x1 +U t d x2
q(x, t) dx + q(x, t) dx = f(q(x1 , t)) − f(q(x2 , t)). (9.367)
dt x1 dt x1 +U t+
Here x1 + Ut− lies just before the discontinuity and x1 + Ut+ lies just past the discontinuity.
Using Leibniz’s rule, Eq. (2.276), we get
Z x1 +U t− Z x2
− ∂q + ∂q
q(x1 + Ut , t)U + 0 + dx + 0 − q(x1 + Ut , t)U + dx (9.368)
x1 ∂t x1 +U t+ ∂t
= f(q(x1 , t)) − f(q(x2 , t)).
Now if we assume that on either side of the discontinuity the volume of integration is suffi-
ciently small so that the time and space variation of q is negligibly small, we get
q(x1 )U − q(x2 )U = f(q(x1 )) − f(q(x2 )), (9.369)
U (q(x1 ) − q(x2 )) = f(q(x1 )) − f(q(x2 )). (9.370)
Defining next the notation for a jump as
Jq(x)K ≡ q(x2 ) − q(x1 ), (9.371)
the jump conditions are rewritten as
U Jq(x)K = Jf(q(x))K . (9.372)
If U = 0, as is the case when we transform to the frame where the wave is at rest, we simply
recover
0 = f(q(x1 )) − f(q(x2 )), (9.373)
f(q(x1 )) = f(q(x2 )), (9.374)
Jf(q(x))K = 0. (9.375)
That is the fluxes on either side of the discontinuity are equal. This is precisely what we
obtained by our naı̈ve analysis. We also get a more general result for U 6= 0, that is the
well-known
f(q(x2 )) − f(q(x1 )) Jf(q(x))K
U= = . (9.376)
q(x2 ) − q(x1 ) Jq(x)K

© 05 December 2025, Joseph M. Powers. All rights reserved.

382 CHAPTER 9. ONE-DIMENSIONAL COMPRESSIBLE FLOW

The general Rankine-Hugoniot equation then for the one-dimensional Euler equations across
a non-stationary jump is given by
   
ρ2 − ρ1 ρ2 u2 − ρ1 u1
U ρ2 u
2 − ρ1 u1 =  ρ2 u22 + p2 − ρ1 u21 
− p1 .
1 2 1 2

1 2 p 1 2 p1
ρ2 e2 + 2 u2 − ρ1 e1 + 2 u1 ρ2 u2 e2 + 2 u2 + ρ2 − ρ1 u1 e1 + 2 u1 + ρ1
2

(9.377)

9.4.2 Rayleigh line

If we operate on the momentum equation as follows

p2 = p1 + ρ1 U 2 − ρ2 u22 , (9.378)
ρ2 U 2 ρ22 u22
= p1 + 1 − . (9.379)
ρ1 ρ2
Because mass gives us ρ22 u22 = ρ21 U 2 we get an equation for the Rayleigh Line,11 a line in
(p, 1/ρ) space:
 
1 1
p2 = p1 + ρ21 U 2 − . (9.380)
ρ1 ρ2
Note that the Rayleigh line
• passes through ambient state,

• has negative slope,

• has a slope with magnitude proportional to square of the wave speed, and

• is independent of state and energy equations.

9.4.3 Hugoniot curve

Let us now work on the energy equation, using both mass and momentum to eliminate
velocity. First eliminate u2 via the mass equation:
u2 U2
h2 + 2 = h1 + , (9.381)
2 2
 2
1 ρ1 U U2
h2 + = h1 + , (9.382)
2 ρ2 2
11
John William Strutt (Lord Rayleigh), 1842-1919, aristocratic-born English mathematician and physi-
cist, studied at Cambridge, influenced by Stokes, toured the United States rather than the traditional
continent of Europe, described correctly why the sky is blue, appointed Cavendish professor experimental
physics at Cambridge, won the Nobel prize for the discovery of Argon, described traveling waves and solitons.

© 05 December 2025, Joseph M. Powers. All rights reserved.

9.4. NORMAL SHOCK WAVES 383

2  !
U2 ρ1
h2 − h1 + −1 = 0, (9.383)
2 ρ2
 
U 2 ρ21 − ρ22
h2 − h1 + = 0, (9.384)
2 ρ22
 
U 2 (ρ1 − ρ2 ) (ρ1 + ρ2 )
h2 − h1 + = 0. (9.385)
2 ρ22
Now use the Rayleigh line, Eq. (9.380), to eliminate U 2 :
  −1
2 1 1 1
U = (p2 − p1 ) − , (9.386)
ρ21 ρ1 ρ2
  −1
1 ρ2 − ρ1
= (p2 − p1 ) , (9.387)
ρ2 ρρ
 1 1 2 
1 ρ1 ρ2
= (p2 − p1 ) 2
. (9.388)
ρ1 ρ2 − ρ1
So the energy equation becomes
   
1 1 ρ1 ρ2 (ρ1 − ρ2 ) (ρ1 + ρ2 )
h2 − h1 + (p2 − p1 ) = 0, (9.389)
2 ρ21 ρ2 − ρ1 ρ22
  
1 1 ρ1 + ρ2
h2 − h1 − (p2 − p1 ) = 0, (9.390)
2 ρ1 ρ2
 
1 1 1
h2 − h1 − (p2 − p1 ) + = 0. (9.391)
2 ρ2 ρ1
Regrouping to see what induces enthalpy changes, we get
  
1 1 1
h2 − h1 = (p2 − p1 ) + , (9.392)
2 ρ2 ρ1
 
v̂2 + v̂1
h −h = (p2 − p1 ), (9.393)
| 2 {z }1 2 | {z }
∆h | {z } ∆p
v̂mean
∆h = v̂mean ∆p. (9.394)
This equation is the Hugoniot equation. It
• holds that enthalpy change equals the product of the mean volume, and the pressure
difference,12 ,
• is independent of wave speed U and velocity u2 , and
• is independent of the equation of state.
12
Note the similarity here between a common result for reversible thermodynamics. Using the definition
of enthalpy, h = e + pv̂ in the Gibbs equation gives T ds = dh − v̂ dp. For an isentropic change, we get
dh = v̂ dp.

© 05 December 2025, Joseph M. Powers. All rights reserved.

384 CHAPTER 9. ONE-DIMENSIONAL COMPRESSIBLE FLOW

9.4.4 Solution procedure for general equations of state

The shocked state can be determined by the following procedure:

• specify the equation of state h(p, ρ),

• substitute the equation of state into the Hugoniot, Eq. (9.392), to get a second relation
between p2 and ρ2 ,

• use the Rayleigh line, Eq. (9.380), to eliminate p2 in the Hugoniot so that the Hugoniot
is a single equation in ρ2 ,

• solve for ρ2 as functions of “1” and U,

• back substitute to solve for p2 , u2 , h2 , T2 as functions of “1” and U,

• invert to find U as function of “1” state and u2 ,

• back transform to laboratory frame to get U as function of “1” state and piston velocity
û2 = ûp .

9.4.5 Calorically perfect ideal gas solutions

Let us follow this procedure for the special case of a CPIG.

h = cp (T − To ) + ĥ, (9.395)
p = ρRT. (9.396)

Thus,
 
p po
h = cp − + ĥ, (9.397)
Rρ Rρo
 
cp p p o
= − + ĥ, (9.398)
R ρ ρo
 
cp p po
= − + ĥ, (9.399)
cp − cv ρ ρo
 
γ p po
= − + ĥ. (9.400)
γ − 1 ρ ρo

Evaluate at states 1 and 2 and substitute into the Hugoniot equation, Eq. (9.392):
          
γ p2 po γ p1 po 1 1 1
− + ĥ − − + ĥ = (p2 − p1 ) + .
γ − 1 ρ2 ρo γ − 1 ρ1 ρo 2 ρ2 ρ1
(9.401)

© 05 December 2025, Joseph M. Powers. All rights reserved.

9.4. NORMAL SHOCK WAVES 385

Rearranging, we find
    
γ p2 p1 1 1 1
− − (p2 − p1 ) + = 0, (9.402)
γ − 1 ρ2 ρ1 2 ρ2 ρ1
   
γ 1 1 1 γ 1 1 1
p2 − − − p1 − − = 0, (9.403)
γ − 1 ρ2 2ρ2 2ρ1 γ − 1 ρ1 2ρ2 2ρ1
   
γ+1 1 1 γ+1 1 1
p2 − − p1 − = 0, (9.404)
2 (γ − 1) ρ2 2ρ1 2 (γ − 1) ρ1 2ρ2
   
γ+1 1 1 γ+1 1 1
p2 − − p1 − = 0. (9.405)
γ − 1 ρ2 ρ1 γ − 1 ρ1 ρ2

The final form is

γ+1 1 1
γ−1 ρ1
− ρ2
p2 = p1 γ+1 1 1
. (9.406)
γ−1 ρ2
− ρ1

We see the Hugoniot for a CPIG

• is a hyperbola in (p, 1/ρ) space,

• has as 1/ρ2 → (γ − 1)/(γ + 1)(1/ρ1 ) causes p2 → ∞, note for γ = 7/5, we get ρ2 → 6ρ1
for infinite pressure, and

• has as 1/ρ2 → ∞, p2 → −p1 (γ − 1)/(γ + 1); note negative pressure, not physical here.

The Rayleigh line and Hugoniot curve are sketched in Fig. 9.8. Note:

• intersections of the two curves are solutions to the equations,

• the ambient state “1” is one solution,

• the other solution “2” is known as the shock solution,

• the shock solution has higher pressure and higher density,

• higher wave speed implies higher pressure and higher density,

• a minimum wave speed exists, it

– occurs when the Rayleigh line is tangent to the Hugoniot curve,

– occurs for infinitesimally small pressure changes,
– corresponds to a sonic wave speed, and
– has disturbances that are acoustic.

• if pressure increases, it can be shown that entropy increases, and

© 05 December 2025, Joseph M. Powers. All rights reserved.

386 CHAPTER 9. ONE-DIMENSIONAL COMPRESSIBLE FLOW

p (kPa)
500
shocked state

400
excluded zone
slope of Rayleigh line < 0
excluded
300
zone, Rayleigh line, slope
from mass and momentum

200
Hugoniot,
initial state from energy
100
excluded zone, 2nd law violation

2 3 4 5 6 7

excluded zone, negative pressure

Figure 9.8: Rayleigh line and Hugoniot curve for a typical shocked calorically perfect ideal
gas.

• if pressure decreases (for wave speeds that are less than sonic), entropy decreases; this
is non-physical.
Substitute the Rayleigh line into the Hugoniot equation to get a single equation for ρ2 :
  γ+1 1
2 2 1 1 γ−1 ρ1
− ρ12
p1 + ρ1 U − = p1 γ+1 1 . (9.407)
ρ1 ρ2 γ−1 ρ
− ρ1 2 1

This equation is quadratic in 1/ρ2 and factorizable. Use computer algebra to solve and get
two solutions, one ambient 1/ρ2 = 1/ρ1 and one shocked solution:
 
1 1 γ−1 2γ p1
= 1+ . (9.408)
ρ2 ρ1 γ + 1 (γ − 1) U 2 ρ1
The shocked density ρ2 is plotted against wave speed U for CPIG air in Fig. 9.9a. Note
• the density solution allows allows all wave speeds 0 < U < ∞,

• the range, however, is U ∈ [c1 , ∞),

• the Rayleigh line and Hugoniot show U ≥ c1 ,

• the solution for U = U(ûp ), (to be shown), requires U ≥ c1 ,

• strong shock limit: U 2 → ∞, ρ2 → (γ + 1)/(γ − 1),

© 05 December 2025, Joseph M. Powers. All rights reserved.

9.4. NORMAL SHOCK WAVES 387

• acoustic limit: U 2 → γp1 /ρ1 , ρ2 → ρ1 , and

• non-physical limit: U 2 → 0, ρ2 → 0.

strong
shock
limit
7
6
8 x10
6
5 6 x10
6

4 exact ambient =
6
3 solution 4 x10 100000 Pa
exact
2 6 solution and
2 x10
1 strong shock limit

500 1000 1500 2000 2500 3000 500 1000 1500 2000 2500 3000
U (m/s) U (m/s)

a) b)

Figure 9.9: Shock a) density and b) pressure versus shock wave speed for a calorically perfect
ideal gas, R = 0.287 kJ/(kg K), γ = 7/5.

Back substitute into Rayleigh line and mass conservation to solve for the shocked pressure
and the fluid velocity in the shocked wave frame:
2 γ−1
p2 = ρ1 U 2 − p1 , (9.409)
γ+1 γ+1
 
γ−1 2γ p1
u2 = −U 1+ . (9.410)
γ+1 (γ − 1) U 2 ρ1
The shocked pressure p2 is plotted against wave speed U for CPIG air in Fig. 9.9b including
both the exact solution and the solution in the strong shock limit. For these parameters, the
results are indistinguishable. The shocked wave frame fluid particle velocity u2 is plotted
against wave speed U for CPIG air in Fig. 9.10a. The shocked wave frame fluid particle
Mach number, M22 = ρ2 u22 /(γp2 ), is plotted against wave speed U for CPIG air in Fig. 9.10b.
In the steady frame, the Mach number of the
• undisturbed flow is (and must be) > 1: supersonic, and
• shocked flow is (and must be) < 1: subsonic.
Transform back to the laboratory frame u = û − U:
 
γ−1 2γ p1
û2 − U = −U 1+ , (9.411)
γ+1 (γ − 1) U 2 ρ1
 
γ−1 2γ p1
û2 = U − U 1+ . (9.412)
γ+1 (γ − 1) U 2 ρ1

© 05 December 2025, Joseph M. Powers. All rights reserved.

388 CHAPTER 9. ONE-DIMENSIONAL COMPRESSIBLE FLOW

U (m/s)
500 1000 1500 2000 2500 3000
-100 1

-200 strong 0.8 exact

shock solution
-300 limit 0.6

-400 0.4
exact strong shock limit
0.2
-500 solution
U (m/s)
0 500 1000 1500 2000 2500 3000

a) b)

Figure 9.10: a) Shock wave frame fluid particle velocity and b) Mach number squared
of shocked fluid particle versus shock wave speed for a calorically perfect ideal gas, R =
0.287 kJ/(kg K), γ = 7/5.

Manipulate this equation and solve the resulting quadratic equation for U and get
s  2
γ+1 γp1 2 γ +1
U= û2 ± + û2 . (9.413)
4 ρ1 4
Now if û2 > 0, we expect U > 0 so take positive root, also set the velocity equal to the
piston velocity û2 = ûp .
s  2
γ+1 γp1 2
γ+1
U= ûp + + ûp . (9.414)
4 ρ1 4
Note:
• acoustic limit: as ûp → 0, U → c1 ; the shock speed approaches the sound speed, and
• strong shock limit: as ûp → ∞, U → ûp (γ + 1)/2.
The shock speed U is plotted against piston velocity ûp for CPIG air in Fig. 9.11a. Both the
exact solution and strong shock limit are shown. If we define the Mach number of the shock
as
U
Ms ≡ , (9.415)
c1
we get
s  2
γ + 1 ûp û2p γ+1
Ms = √ + 1+ . (9.416)
4 γRT1 γRT1 4

© 05 December 2025, Joseph M. Powers. All rights reserved.

9.4. NORMAL SHOCK WAVES 389

U (m/s)
exact exact
1200 solution solution
1000 3
strong strong
800 shock shock
limit 2 limit
600
acoustic acoustic
limit, 400
limit, 1
200

200 400 600 800 1000 200 400 600 800 1000

a) b)

Figure 9.11: a) Shock speed and b) shock Mach number versus piston velocity for a calorically
perfect ideal gas, R = 0.287 kJ/(kg K), γ = 7/5.

The shock Mach number Ms is plotted against piston velocity ûp for CPIG air in Fig. 9.11b.
Both the exact solution and strong shock limit are shown.
Let us find the entropy change induced by a shock for a CPIG. We first need an expression
for the entropy change. Begin with the Gibbs equation, Eq. (4.163):

T ds = de + p dv̂, (9.417)
de p
ds = + dv̂. (9.418)
T T
Now invoke the CPIG assumption to get

dT dv̂.
ds = cv +R . (9.419)
T v̂
Now for the ideal gas with pv̂ = RT , we get

p dv̂ + v̂ dp = R dT. (9.420)

Divide the left side by pv̂ and the right side by the equivalent RT to get

dv̂ dp dT
+ = . (9.421)
v̂ p T

Use this to eliminate temperature in Eq. (9.419) to get

 
dv̂ dp dv̂.
ds = cv + +R , (9.422)
v̂ p v̂
dv̂ dp
= (cv + R) + cv , (9.423)
v̂ p
dv̂ dp
= (cv + (cp − cv )) + cv , (9.424)
v̂ p

© 05 December 2025, Joseph M. Powers. All rights reserved.

390 CHAPTER 9. ONE-DIMENSIONAL COMPRESSIBLE FLOW

dv̂ dp
= cp + cv , (9.425)
v̂ p
 
dv̂ dp
ds = cv γ + , (9.426)
v̂ p
 
v̂2 p2
s2 − s1 = cv γ ln + ln , (9.427)
v̂1 p1
  γ 
v̂2 p2
= cv ln + ln , (9.428)
v̂1 p1
  γ 
ρ1 p2
= cv ln + ln , (9.429)
ρ2 p1
 γ 
ρ1 p2
= cv ln . (9.430)
ρ2 p1

Then we use Eqs. (9.408, 9.409, and 9.415) to eliminate the pressure and density ratios in
favor of Ms , and follow this with algebraic reduction to arrive at
  (γ−1)M 2 +2 γ 
2
s2 − s1
s
(γ+1)Ms2
(γ (2Ms − 1) + 1)
= ln  . (9.431)
cv γ+1

For γ = 7/5, we plot (s2 − s1 )/cv as a function of Ms in Fig. 9.12. Clearly for Ms = 1, we

1.0

0.5

1 2 3 4 5

- 0.5

- 1.0

Figure 9.12: Scaled entropy jump through a discontinuity as a function of Ms for a CPIG
with γ = 7/5.

have (s2 − s1 )/cv = 0, so the sonic wave is isentropic. And clearly for Ms > 1, the entropy
rises, thus satisfying the second law for what is an adiabatic irreversible compression. For
Ms < 1, the entropy is predicted to fall for an adiabatic expansion. This is not observed
in nature and violates the second law of thermodynamics. So we must have Ms ≥ 1 for a

© 05 December 2025, Joseph M. Powers. All rights reserved.

9.4. NORMAL SHOCK WAVES 391

propagating discontinuity. The equation for entropy jump is complicated. We can better
understand it by performing a Taylor series expansion in the neighborhood of Ms = 1. Doing
so yields
s2 − s1 16γ (γ − 1) 3 4


= (Ms − 1) + O (Ms − 1) . (9.432)
cv 3(γ + 1)2
Clearly for general γ > 1, the entropy rises for Ms > 1 and falls for Ms < 1, and the local
behavior is cubic in the deviation of Ms from unity.

9.4.6 Acoustic limit

Consider that state 2 is a small perturbation of state 1 so that
ρ2 = ρ1 + ∆ρ, (9.433)
u2 = u1 + ∆u, (9.434)
p2 = p1 + ∆p. (9.435)
Substituting into the normal shock equations, we get
(ρ1 + ∆ρ) (u1 + ∆u) = ρ1 u1 , (9.436)
(ρ1 + ∆ρ) (u1 + ∆u)2 + (p1 + ∆p) = ρ1 u1 2 + p1 , (9.437)
γ p1 + ∆p 1 γ p1 1 2
+ (u1 + ∆u)2 = + u1 . (9.438)
γ − 1 ρ1 + ∆ρ 2 γ − 1 ρ1 2
Expanding, we get
ρ1 u1 + u1 (∆ρ) + ρ1 (∆u) + (∆ρ) (∆u) = ρ1 u1 , (9.439)


ρ1 u1 2
+ 2ρ1 u1 (∆u) + u1 (∆ρ) + ρ1 (∆u)2 + 2u1 (∆u) (∆ρ) + (∆ρ) (∆u)2
2

+ (p1 + ∆p) = ρ1 u1 2 + p1 , (9.440)

 
γ p1 1 p1 1

+ ∆p − 2 ∆ρ + ... + u1 2 + 2u1 (∆u) + (∆u)2 (9.441)
γ − 1 ρ1 ρ1 ρ1 2
γ p1 1 2
= + u1 . (9.442)
γ − 1 ρ1 2
Subtracting the base state and eliminating products of small quantities yields
u1 (∆ρ) + ρ1 (∆u) = 0, (9.443)
2ρ u (∆u) + u1 2 (∆ρ) + ∆p = 0, (9.444)
1 1 
γ 1 p1
∆p − 2 ∆ρ + u1 (∆u) = 0. (9.445)
γ − 1 ρ1 ρ1
In matrix form this is
    
u1 ρ1 0 ∆ρ 0
 u1 2 2ρ1 u1 1   ∆u  =  0  . (9.446)
γ p1 γ 1
− γ−1 ρ21
u1 γ−1 ρ1 ∆p 0

© 05 December 2025, Joseph M. Powers. All rights reserved.

392 CHAPTER 9. ONE-DIMENSIONAL COMPRESSIBLE FLOW

As the right hand side is zero, the determinant must be zero and there must be a linear
dependency of the solution. First check the determinant:
   
2γ γ u1 2 γ p1
u1 u1 − u1 − ρ1 + = 0, (9.447)
γ−1 γ − 1 ρ1 γ − 1 ρ21
 
u1 2 1 2 p1
(2γ − (γ − 1)) − γ u1 + γ = 0, (9.448)
γ−1 γ −1 ρ1
 
2 2 p1
u1 (γ + 1) − γu1 + γ = 0, (9.449)
ρ1
p1
u1 2 = γ = c21 . (9.450)
ρ1
So the velocity is necessarily sonic for a small disturbance.
Take ∆u to be known and solve a resulting 2 × 2 system:
    
u1 0 ∆ρ −ρ1 ∆u
γ p1 γ 1 = . (9.451)
− γ−1 ρ21 γ−1 ρ1 ∆p −u1 ∆u

Solving yields
ρ1 ∆u ∆u
∆ρ = − q = −ρ1 , (9.452)
p1
γ ρ1 c1
r
p1
∆p = −ρ1 γ ∆u = −ρ1 c1 ∆u. (9.453)
ρ1

9.5 Flow with area change and normal shocks

This section will consider flow from a reservoir with the fluid at stagnation conditions to a
constant pressure environment. The pressure of the environment is commonly known as the
back pressure: pb .
Generic problem: Given A(x), stagnation conditions and pb , find the pressure, tempera-
ture, density at all points in the duct and the mass flow rate.

9.5.1 Converging nozzle

A converging nozzle operating at several different values of pb is sketched in Fig. 9.13. The
flow through the duct can be solved using the following procedure:

• check if pb ≥ p∗ ,

• if so, set pe = pb ,

• determine Me from isentropic flow relations,

© 05 December 2025, Joseph M. Powers. All rights reserved.

9.5. FLOW WITH AREA CHANGE AND NORMAL SHOCKS 393

e d c
1

0 1

1
a--subsonic exit
b--subsonic exit
c--sonic exit
d--choked, external expansion
e--choked, external expansion

Figure 9.13: Converging nozzle sketch.

• determine A∗ from A/A∗ relation,

• at any point in the flow where A is known, compute A/A∗ and then invert A/A∗ relation
to find local M.

Note:

• These flows are subsonic throughout and correspond to points a and b in Fig. 9.13.

• If pb = p∗ then the flow is sonic at the exit and just choked. This corresponds to point
c in Fig. 9.13.

• If pb < p∗ , then the flow chokes, is sonic at the exit, and continues to expand outside
of the nozzle. This corresponds to points d and e in Fig. 9.13.

9.5.2 Converging-diverging nozzle

A converging-diverging nozzle operating at several different values of pb is sketched in
Fig. 9.14.
The flow through the duct can be solved using the a similar following procedure

• set At = A∗ ,

• with this assumption, calculate Ae /A∗ ,

• determine Mesub , Mesup , both supersonic and subsonic, from A/A∗ relation,

© 05 December 2025, Joseph M. Powers. All rights reserved.

394 CHAPTER 9. ONE-DIMENSIONAL COMPRESSIBLE FLOW

possible
normal
shock

1 a--subsonic exit
b--subsonic exit
c--subsonic design
d--shock in duct
e-shock at end of duct
sonic
throat f--external compression
g--supersonic design
h--external expansion
x

hg f e d c
1

0
1

Figure 9.14: Converging-diverging nozzle sketch.

• determine pesub , pesup, from Mesub , Mesup ; these are the supersonic and subsonic design
pressures,
• if pb > pesub, the flow is subsonic throughout and the throat is not sonic. Use same
procedure as for converging duct: Determine Me by setting pe = pb and using isentropic
relations,
• if pesub > pb > pesup, the procedure is complicated.
– estimate the pressure with a normal shock at the end of the duct, pesh .
– If pb ≥ pesh , there is a normal shock inside the duct,
– If pb < pesh , the duct flow is shockless, and there may be compression outside the
duct.
• if pesup = pb , the flow is at supersonic design conditions and the flow is shockless, and

© 05 December 2025, Joseph M. Powers. All rights reserved.

9.6. METHOD OF CHARACTERISTICS 395

• if pb < pesup , the flow in the duct is isentropic and there is expansion outside the duct.

9.6 Method of characteristics

Here we discuss how to use the so-called method of characteristics to model a one-dimensional
unsteady, inviscid, non-heat conducting fluid. The emphasis will be on rarefaction waves.
This analysis is a good deal more rigorous than much of traditional one-dimensional gas dy-
namics, and draws upon some of the more difficult mathematical methods we will encounter.
In assuming no diffusive transport, we have eliminated all mechanisms for entropy gen-
eration; consequently, we will be able to model the process as isentropic. We note that even
without diffusion, shocks can generate entropy. However, the expansion waves are inherently
continuous, and do remain isentropic. We will consider a general equation of state, and
later specialize to a CPIG. The problem is inherently non-linear and is modeled by partial
differential equations of the type that is known as hyperbolic. Such problems, in contrast to
say Laplace’s equation, that requires boundary conditions, require initial data only, and no
boundary data.

9.6.1 Inviscid one-dimensional equations

The equations to be considered are shown here in non-conservative form. These are one-
dimensional limits of the system developed in Ch. 6.5:

∂ρ ∂ρ ∂u
+u +ρ = 0, (9.454)
∂t ∂x ∂x
∂u ∂u ∂p
ρ + ρu + = 0, (9.455)
∂t ∂x ∂x
∂s ∂s
+u = 0, (9.456)
∂t ∂x
p = p(ρ, s). (9.457)

Here we have written the energy equation in terms of entropy. The development of this was
shown in Ch. [Link]. We have also utilized the general result from thermodynamics that any
intensive property can be written as a function of two other independent thermodynamic
properties. Here we have chosen to write pressure as a function of density and entropy, as
we did in Eq. (9.129) for a special case. Thus, we have four equations for the four unknowns,
ρ, u, p, s.
Now we note that
∂p ∂p
dp = dρ + ds, so, (9.458)
∂ρ s ∂s ρ
∂p ∂p ∂ρ ∂p ∂s
= + . (9.459)
∂x t ∂ρ s ∂x t ∂s ρ ∂x t

© 05 December 2025, Joseph M. Powers. All rights reserved.

396 CHAPTER 9. ONE-DIMENSIONAL COMPRESSIBLE FLOW

Now, let us define thermodynamic properties c2 and ζ as follows

∂p ∂p
c2 ≡ , ζ≡ . (9.460)
∂ρ s ∂s ρ

We will see that ζ will be unimportant, and will be able to ascribe to c the physical signifi-
cance of the speed of propagation of small disturbances, the so-called sound speed, that we
have already encountered in acoustics. If we know the equation of state, then we can think
of c2 and ζ as known thermodynamic functions of ρ and s. Our definitions give us
∂p ∂ρ ∂s
= c2 +ζ . (9.461)
∂x ∂x ∂x
Substituting into our governing equations, we see that pressure can be eliminated to give
three equations in three unknowns:
∂ρ ∂ρ ∂u
+u +ρ = 0, (9.462)
∂t ∂x ∂x
∂u ∂u ∂ρ ∂s
ρ + ρu + c2 +ζ = 0, (9.463)
∂t ∂x | ∂x {z ∂x}
∂p
∂x

∂s ∂s
+u = 0. (9.464)
∂t ∂x
Now we can say that if s = s(x, t),

∂s ∂s
ds = dt + dx, (9.465)
∂t ∂x
ds ∂s dx ∂s
= + , (9.466)
dt ∂t dt ∂x
∂s ∂s
= +u . (9.467)
∂t ∂x
Thus, on curves where dx/dt = u (that by definition are particle pathlines), we have from
substituting Eq. (9.467) into the energy equation (9.464)

ds
= 0. (9.468)
dt
Thus we have converted the partial differential equation for energy conservation into an
ordinary differential equation. This can be integrated to give us
dx
s = C, on a particle pathline, = u. (9.469)
dt
This scenario is sketched on the so-called (x, t) diagram of Fig. 9.15.

© 05 December 2025, Joseph M. Powers. All rights reserved.

9.6. METHOD OF CHARACTERISTICS 397

t
pathlines

Figure 9.15: (x, t) diagram showing maintenance of entropy s along particle pathlines
dx/dt = u for isentropic flow.

This result is satisfying, but not complete, as we do not in general know where the
pathlines are. Let us try to apply this technique to the system in general. Consider our
equations in matrix form:
   ∂ρ     ∂ρ   
1 0 0 ∂t
u ρ 0 ∂x
0
 0 ρ 0   ∂u  +  c2 ρu ζ   ∂u  =  0  . (9.470)
∂t ∂x
∂s ∂s
0 0 1 ∂t
0 0 u ∂x
0
These equations are of the form
∂uj ∂uj
Aij + Bij = fi . (9.471)
∂t ∂x
As described by Whitham,13 there is a general technique to analyze such equations. First
pre-multiply both sides of the equation by a yet to be determined vector of variables ℓi :
∂uj ∂uj
ℓi Aij + ℓi Bij = ℓi fi . (9.472)
∂t ∂x
Now, this method will work if we can choose ℓi to render this product to be of the form
similar to ∂/∂t + u(∂/∂x). Let us take
 
∂uj ∂uj ∂uj ∂uj
ℓi Aij + ℓi Bij = mj +λ , (9.473)
∂t ∂x ∂t ∂x
duj dx
= mj on = λ. (9.474)
dt dt
13
Gerald Beresford Whitham, 1927-2014, applied mathematician and developer of theory for non-linear
wave propagation.

© 05 December 2025, Joseph M. Powers. All rights reserved.

398 CHAPTER 9. ONE-DIMENSIONAL COMPRESSIBLE FLOW

So comparing terms, we see that

ℓi Aij = mj , ℓi Bij = λmj , (9.475)

λℓi Aij = λmj , (9.476)

so, we get by eliminating mj that

ℓi (λAij − Bij ) = 0. (9.477)

This is a left eigenvalue problem. We set the determinant of λAij −Bij to zero for a non-trivial
solution and find
λ−u −ρ 0
2
−c ρ(λ − u) −ζ = 0. (9.478)
0 0 λ−u
Evaluating, we get


(λ − u) ρ(λ − u)2 + ρ(λ − u)(−c2 ) = 0, (9.479)


ρ(λ − u) (λ − u)2 − c2 = 0. (9.480)

Solving we get
λ = u, λ = u ± c. (9.481)
Now the left eigenvectors ℓi give us the actual equations. First for λ = u, we get
 
u−u −ρ 0
( ℓ1 ℓ2 ℓ3 )  −c2 ρ(u − u) −ζ  = ( 0 0 0 ) , (9.482)
0 0 u−u
 
0 −ρ 0
( ℓ1 ℓ2 ℓ3 )  −c2 0 −ζ  = ( 0 0 0 ) . (9.483)
0 0 0
Two of the equations require that ℓ1 = 0 and ℓ2 = 0. There is no restriction on ℓ3 . We will
select a normalized solution so that

ℓi = (0, 0, 1). (9.484)

Thus, ℓi Aij (∂uj /∂t) + ℓi Bij (∂uj /∂x) = ℓi fi gives

   ∂ρ     ∂ρ   
1 0 0 ∂t
u ρ 0 ∂x
0
( 0 0 1 )  0 ρ 0   ∂u ∂t
 + ( 0 0 1 )  c2 ρu ζ   ∂u
∂x
 = (0 0 1)0,
∂s ∂s
0 0 1 ∂t
0 0 u 0
 ∂ρ   ∂x
∂ρ

∂t ∂x
( 0 0 1 )  ∂u
∂t
 + ( 0 0 u )  ∂u  = 0,
∂x
∂s ∂s
∂t ∂x
∂s ∂s
+u = 0. (9.485)
∂t ∂x

© 05 December 2025, Joseph M. Powers. All rights reserved.

9.6. METHOD OF CHARACTERISTICS 399

So as before with s = s(x, t), we have ds = (∂s/∂t)dt + (∂s/∂x)dx, and ds/dt = ∂s/∂t +
(dx/dt)(∂s/∂x). Now if we require dx/dt to be a particle pathline, dx/dt = u, then our
energy equation, Eq. (9.485), gives us
ds dx
= 0, on = u. (9.486)
dt dt
The special case in which the pathlines are straight in (x, t) space, corresponding to a uniform
velocity field of u(x, t) = uo , is sketched in the (x, t) diagram of Fig. 9.16.

t
pathlines

Figure 9.16: (x, t) diagram showing maintenance of entropy s along particle pathlines
dx/dt = uo for isentropic flow.

Now let us look at the remaining eigenvalues, λ = u ± c.

 
u±c−u −ρ 0
( ℓ1 ℓ2 ℓ3 )  −c2 ρ(u ± c − u) −ζ  = (0 0 0), (9.487)
0 0 u±c−u
 
±c −ρ 0
( ℓ1 ℓ2 ℓ3 )  −c2 ±ρc −ζ  = ( 0 0 0 ) . (9.488)
0 0 ±c
As one of the components of the left eigenvector should be arbitrary, we will take ℓ1 = 1; we
arrive at the following equations then
1
±c − c2 ℓ2 = 0, =⇒ ℓ2 = ± , (9.489)
c
1
−ρ ± ρcℓ2 = 0, =⇒ ℓ2 = ± , (9.490)
c
ζ
−ζℓ2 ± cℓ3 = 0, =⇒ ℓ3 = 2 . (9.491)
c

© 05 December 2025, Joseph M. Powers. All rights reserved.

400 CHAPTER 9. ONE-DIMENSIONAL COMPRESSIBLE FLOW

Thus, ℓi Aij (∂uj /∂t) + ℓi Bij (∂uj /∂x) = ℓi fi gives

   ∂ρ     ∂ρ   
1 0 0 ∂t
u ρ 0 ∂x
0
1 ζ 
( 1 ± c c2 ) 0 ρ 0   ∂u  1 ζ 
+ ( 1 ± c c 2 ) c2 ρu ζ   ∂u  = ( 1 ±1 ζ
) 0,

∂t ∂x c c2
∂s ∂s
0 0 1 ∂t
0 0 u ∂x
0
(9.492)
 ∂ρ   ∂ρ 
∂t ∂x
( 1 ± ρc ζ
c2
)  ∂u
∂t
 + ( u ± c ρ ± ρu
c
± ζc + ζu
c2
)  ∂u
∂x
 = 0, (9.493)
∂s ∂s
∂t ∂x
  
∂ρ ∂ρ ρ ∂u u  ∂u ζ ∂s ζu ζ ∂s
+ (u ± c) ± +ρ 1± + + ± = 0, (9.494)
∂t ∂x c ∂t c ∂x c2 ∂t c2 c ∂x
     
∂ρ ∂ρ ρ ∂u ∂u ζ ∂s ∂s
+ (u ± c) ± + (u ± c) + 2 + (u ± c) = 0, (9.495)
∂t ∂x c ∂t ∂x c ∂t ∂x
     
2 ∂ρ ∂ρ ∂u ∂u ∂s ∂s
c + (u ± c) ±ρc + (u ± c) +ζ + (u ± c) = 0. (9.496)
∂t ∂x ∂t ∂x ∂t ∂x
| {z } | {z } | {z }
dρ/dt du/dt ds/dt

Now on lines where dx/dt = u±c, we get a transformation of the partial differential equations
to ordinary differential equations:

dρ du ds dx
c2 ± ρc + ζ = 0, on = u ± c. (9.497)
dt dt dt dt
A sketch of the characteristics, the lines on which the differential equations are obtained, is
given in the (x, t) diagram of Fig. 9.17. Lastly, note that if we exchange ρ for p, we get a

pathline
acoustic characteristic acoustic
t characteristic
characteristic

Figure 9.17: (x, t) diagram showing characteristics for pathlines dx/dt = u and acoustic
waves dx/dt = u ± c.

© 05 December 2025, Joseph M. Powers. All rights reserved.

9.6. METHOD OF CHARACTERISTICS 401

more compact form. Note that Eq. (9.458) lets us form

dp dρ ds dx
= c2 + ζ , on = u ± c. (9.498)
dt dt dt dt
Using this, we can recast Eq. (9.497) as
dp du dx
± ρc = 0, on = u ± c. (9.499)
dt dt dt

9.6.2 Homeoentropic flow of a calorically perfect ideal gas

The equations developed so far are valid for a general equation of state. Here let us now
consider the flow of a CPIG, so p = ρRT and e = cv T + ê. Consequently, we have the
standard relation for the square of the sound speed of a CPIG:
p
c2 = γ . (9.500)
ρ
Further let us take the flow to be homeoentropic (as introduced in Ch. 6.5.3), that is to say,
not only does the entropy remain constant on pathlines, that are isentropic, but it has the
same value on each streamline. That is the entropy field is a constant. Consequently, we
have the standard relation for a CPIG:
p
= A, (9.501)
ργ
where A is a constant. Because of homeoentropy, we no longer need consider the energy
equation, and the linear combination of mass and linear momentum equations, Eq. (9.497),
reduces to
dρ du dx
c2 ± ρc = 0, on = u ± c. (9.502)
dt dt dt
Rearranging, we get
du dρ c dx
=∓ , on = u ± c. (9.503)
dt dt ρ dt
√ γ−1
Now c2 = γp/ρ = γAργ−1 ; thus, c = γAρ 2 , so
du p γ−1 dρ
= ∓ γAρ 2 ρ−1 , (9.504)
dt dt
p 2 d  γ−1 
= ∓ γA ρ 2 . (9.505)
γ − 1 dt
Regrouping, we find
 
d p 2 γ−1
u ± γA ρ 2 = 0, (9.506)
dt γ−1
 
d 2
u± c = 0. (9.507)
dt γ−1

© 05 December 2025, Joseph M. Powers. All rights reserved.

402 CHAPTER 9. ONE-DIMENSIONAL COMPRESSIBLE FLOW

Following notation used by Courant14 and Friedrichs,15 (1976) we then integrate each of
these equations, both of which are homogeneous, along characteristics to obtain algebraic
relations
2 dx
u+ c = 2r, on = u + c, C + characteristic, (9.508)
γ−1 dt
2 dx
u− c = −2s, on = u − c, C − characteristic. (9.509)
γ−1 dt
Courant and Friedrichs’ s has no relation to entropy; it is just a new variable introduced
for convenience. A sketch of the characteristics is given in the (x, t) diagram of Fig. 9.18.
Now r and s can take on different values, depending on which characteristic we are on. On
t

arbitrary region
of interest

Figure 9.18: (x, t) diagram showing C + and C − characteristics dx/dt = u ± c.

a given characteristic, they remain constant. Let us define additional parameters α and β
to identify which characteristic we are on. So we have
2 dx
u+ c = 2r(β), on = u + c, C + characteristic, (9.510)
γ−1 dt
2 dx
u− c = −2s(α), on = u − c, C − characteristic. (9.511)
γ−1 dt
14
Richard Courant, 1888-1972, Prussian-born German mathematician, received Ph.D. under David Hilbert
at Göttingen, compiled Hilbert’s course notes into classic two-volume text of applied mathematics, drafted
into German army in World War I, where half of his unit was killed in action, developed telegraph system
that used the earth as a conductor for use in the trenches of the Western front, expelled from Göttingen by
the Nazis in 1933, fled Germany, and founded the Courant Institute of Mathematical Sciences at New York
University, author of classic mathematical text on supersonic fluid mechanics.
15
Kurt Otto Friedrichs, 1901-1982, German-born mathematician who emigrated to the United States in
1937, student of Richard Courant’s at Göttingen, taught at Aachen, Braunschweig, and New York University,
worked on partial differential equations of mathematical physics and fluid mechanics.

© 05 December 2025, Joseph M. Powers. All rights reserved.

9.6. METHOD OF CHARACTERISTICS 403

These quantities are known as Riemann invariants.16

9.6.3 Simple waves

Simple waves are defined to exist when either r(β) or s(α) are constant everywhere in (x, t)
space and not just on characteristics. For example, say s(α) = so . Then the Riemann
invariant
2
u− c = −2so , everywhere, (9.512)
γ −1
is actually invariant over all of (x, t) space. Now the other Riemann invariant,
2
u+ c = 2r(β), on C + , (9.513)
γ−1
takes on many values depending on β. However, it is easily shown that for the simple wave
that the characteristics have a constant slope in the (x, t) plane as sketched in the (x, t)
diagram of Fig. 9.19.
t

arbitrary region
of interest

Figure 9.19: (x, t) diagram showing C + for a simple wave.

Now consider a rarefaction with a prescribed piston motion u = up (t). A sketch is

given in the (x, t) diagram of Fig. 9.20. For this configuration, the Riemann invariant
u − 2c/(γ − 1) = −2so is valid everywhere. Let us evaluate so in terms of more fundamental
variables. For the piston problem we are considering, when t = 0, we have u = 0, c = co , so
2 2
u− c=− co , everywhere. (9.514)
γ−1 γ−1
16
Georg Friedrich Bernhard Riemann, 1826-1866, German mathematician and geometer whose work in
non-Euclidean geometry was critical to Einstein’s theory of general relativity, produced the first major study
of shock waves.

© 05 December 2025, Joseph M. Powers. All rights reserved.

404 CHAPTER 9. ONE-DIMENSIONAL COMPRESSIBLE FLOW

t
+
C

Figure 9.20: (x, t) diagram showing C + characteristics for isentropic rarefaction problem,
along with piston cylinder arrangement.

Thus, so = co /(γ − 1).

Consider now a special characteristic Ĉ + at t = t̂. See Fig. 9.21. At this time the piston
moves with velocity ûp , and the fluid velocity at the piston face is

uf ace (t̂) = ûp . (9.515)

We get cf ace (t̂) from Eq. (9.514), that must be valid everywhere, including the face of the
piston:

2 2
uf ace − cf ace = − co , (9.516)
| {z } γ − 1 γ−1
ûp
γ−1
cf ace (t = t̂) = co + ûp . (9.517)
2
Also from Eq. (9.514), we have

γ−1
c = co + u, everywhere, (9.518)
2
that is valid everywhere.

© 05 December 2025, Joseph M. Powers. All rights reserved.

9.6. METHOD OF CHARACTERISTICS 405

Figure 9.21: (x, t) diagram showing Ĉ + for our rarefaction problem.

Now on Ĉ + , we have from Eq. (9.513) that

 
2 2
u+ c = uf ace + cf ace , (9.519)
γ−1 γ−1 t=t̂
   
2 γ−1 2 γ−1
u+ co + u = ûp + co + ûp , (9.520)
γ−1 2 γ−1 2
| {z } | {z }
c cf ace
2 2
2u + co = 2ûp + co , (9.521)
γ−1 γ−1
u = ûp on Ĉ + . (9.522)

So on Ĉ + , we have from Eq. (9.518) that

γ −1
c = co + ûp . (9.523)
2
So for Ĉ + , we get
dx
= u + c, (9.524)
dt
γ−1
= ûp + co + ûp , (9.525)
| {z2 }
c
γ+1
= ûp + co . (9.526)
2
For a particular characteristic, this slope is a constant, as was earlier suggested.

© 05 December 2025, Joseph M. Powers. All rights reserved.

406 CHAPTER 9. ONE-DIMENSIONAL COMPRESSIBLE FLOW

Now for our prescribed motion, ûp decreases with time and becomes more negative; hence
the slope of our Ĉ + characteristic decreases, and the characteristics diverge in (x, t) space.
The slope of the leading characteristic is co , the ambient sound speed. The characteristic we
consider, Ĉ + is sketched in the (x, t) diagram of Fig. 9.21.
We can use our Riemann invariant along with isentropic relations to obtain other flow
variables. From Eq. (9.514), we get

c γ−1 u
=1+ . (9.527)
co 2 co
γ−1
Because the flow is homeoentropic, we have c/co = (ρ/ρo ) 2 and p/po = (ρ/ρo )γ , so
  2γ
p γ − 1 u γ−1
= 1+ , (9.528)
po 2 co
  2
ρ γ − 1 u γ−1
= 1+ . (9.529)
ρo 2 co

9.6.4 Centered rarefaction

If the piston is suddenly accelerated to a constant velocity, then a family of characteristics
clusters at the origin on the (x, t) diagram and fans out in a centered rarefaction. This can also
be studied using the similarity transformation ξ = x/t that reduces the partial differential
equations to ordinary differential equations. Relevant sketches comparing centered to non-
centered rarefactions are shown in in the (x, t) diagram of Fig. 9.22.

Example 9.13
Analyze a centered rarefaction fan propagating into CPIG air for a piston suddenly accelerated
from rest to up = −100 m/s. Take the ambient air to be at po = 105 Pa, To = 300 K.

The ideal gas law gives ρo = po /RTo = (105 Pa)/((287 J/kg/K)(300 K)) = 1.16 kg/m3 . Now
s  
p 7 J m
co = γRTo = 287 (300 K) = 347 . (9.530)
5 kg K s

On the final characteristic of the fan, Cf+ : u = up = −100 m/s. So

γ−1  m  7/5 − 1  m m
c = co + up = 347 + −100 = 327 . (9.531)
2 s 2 s s
Now the final pressure is

  γ−1
2γ
! 7/5−1
2(7/5)

pf γ − 1 uf 7/5 − 1 −100 ms
= 1+ = 1+ = 0.660. (9.532)
po 2 co 2 347 ms

© 05 December 2025, Joseph M. Powers. All rights reserved.

9.6. METHOD OF CHARACTERISTICS 407

centered rarefaction non-centered rarefaction

t t
up(t)

particle path particle path

x x
u

x
p

Figure 9.22: (x, t) diagram centered and non-centered rarefactions, along with pressure and
velocity profiles for centered fans.

Hence pf = 6.6 × 104 Pa. Because the flow is homeoentropic, we get

  γ1  
pf kg kg
ρf = ρo = 1.16 3 (0.660)5/7 = 0.863 3 . (9.533)
po m m
And the final temperature is
pf 66.0 × 103 Pa
Tf = =    = 266.4 K. (9.534)
ρf R 0.863 mkg
3 287 kgJ K

From linear acoustic theory, Sec. 9.4.6, we deduce that

∆u ∆u
∆ρ ∼ −ρo , ∆p ∼ −ρo co ∆u, ∆T ∼ −(γ − 1)To . (9.535)
co co
We compare the results of this problem with the estimates of linear acoustic theory. and see
kg kg
∆ρexact = −0.298 , ∆ρlinear = −0.335 3 , (9.536)
m3 m
∆pexact = −34.0 × 103 Pa, ∆plinear = −40.3 × 103 Pa, (9.537)
∆Texact = −33.6 K, ∆Tlinear = −34.6 K. (9.538)

© 05 December 2025, Joseph M. Powers. All rights reserved.

408 CHAPTER 9. ONE-DIMENSIONAL COMPRESSIBLE FLOW

9.6.5 Simple compression

We sketch a simple compression generated by a piston gradually accelerating from rest in
the (x, t) diagram of Fig. 9.23. Here the leading wave is at the ambient sonic speed. It

shock
formation

ambient
region

Figure 9.23: (x, t) diagram for simple compression.

slightly raises the temperature downstream. So the local sound speed increases slightly.
This happens for each subsequent compression. So each compression that follows travels
slightly faster. At a downstream point, these compression waves coalesce to form a shock
wave. After the piston reaches a steady velocity, it sends a series of compression waves into
the flow, all propagating at the same speed, and all interacting with the lead shock wave.
This is the mechanism by which energy from the piston is transmitted to the shock front to
support its propagation.

9.6.6 Two interacting expansions

We sketch two interacting expansion waves from two suddenly accelerating pistons in the
(x, t) diagram of Fig. 9.24. Both pistons generate centered rarefactions, one right-propagating,
the other left-propagating. In contrast to the compression wave, each small rarefaction
slightly cools the flow, thus lowering its sound speed. So later rarefactions travel at lower
speeds, and there is no analog to the coalescence of compressions to form a shock. The
rarefaction waves collide and interact in a complicated fashion. The modulated waves then
are transmitted and reflect from the piston walls. This process would continue and generate
a long train of ever-more complicated wave patterns; however, no shocks would be formed.

© 05 December 2025, Joseph M. Powers. All rights reserved.

9.6. METHOD OF CHARACTERISTICS 409

uniform
flow

uniform uniform
flow flow

fluid
at rest

Figure 9.24: (x, t) diagram for two interacting expansion waves.

9.6.7 Wall interactions

We sketch an expansion wall interaction in the (x, t) diagram of Fig. 9.25. Here a suddenly
accelerated left-traveling piston generates a right-propagating centered rarefaction wave.
This wave reflects from a wall at x = L. Though not sketched, the reflected wave will catch
and reflect from the piston. Once again, because the volume is expanding, no shocks would
be formed.

9.6.8 Shock tube

We sketch the behavior of the fluid in a shock tube in the diagrams of Fig. 9.26. In a shock
tube, we have a gas initially segregated into two chambers, separated by a thin diaphragm,
in the initial state. The initial temperature in both chambers is identical, and the gas is at
rest in both chambers. The pressure on one side is different from that on the other, and this
induces a density difference between one side and the other. At t = 0+ , the diaphragm is
removed, and the pressure difference induces a fluid acceleration. As sketched in Fig. 9.26,
where the initial state has higher pressure on the left, a shock wave propagates to the
right, and a rarefaction propagates to the left. These two waves are nonlinear extensions
of acoustic disturbances, and propagate at speeds close to the sound speed. There is also
a contact discontinuity segregating the two regions of different density. This discontinuity
also propagates, but at a speed near the local particle velocity, as it is associated with a

© 05 December 2025, Joseph M. Powers. All rights reserved.

410 CHAPTER 9. ONE-DIMENSIONAL COMPRESSIBLE FLOW

t
rest

uniform
flow

flow
at rest

Figure 9.25: (x, t) diagram for expansion wall interaction.

so-called entropy wave. Across the contact discontinuity, it can be shown that pressure and
velocity must be continuous, while density and temperature can suffer a jump. The shock
and rarefaction both reflect from the walls and interact with each other as well as the contact
discontinuity in a complicated manner.

© 05 December 2025, Joseph M. Powers. All rights reserved.

9.6. METHOD OF CHARACTERISTICS 411

compression
contact
discontinuity
rarefaction
shock
rest rest

uniform
uniform

rest rest
x
p

x
p

x
Τ

Figure 9.26: (x, t) and p, ρ, T versus x behavior for a shock tube.

© 05 December 2025, Joseph M. Powers. All rights reserved.

412 CHAPTER 9. ONE-DIMENSIONAL COMPRESSIBLE FLOW

9.6.9 Inviscid Bateman-Burgers’ equation solution

To this point, we have described a common and traditional approach to the method of
characteristics. Using common notation, we have written what began as partial differential
equations (PDEs) in the form of ordinary differential equations (ODEs), and it is often said
that the method of characteristics is a way to transform PDEs into ODEs. However, the
equations that result are certainly not in a standard form for ODEs; they are burdened with
unusual side conditions.
It is in fact more sound to state that the method of characteristics transforms the PDEs
in (x, t) space to another set of PDEs in a new space (ξ, τ ) in which the integration is much
easier. Consider for example a model equation that is hyperbolic, the inviscid Bateman17 -
Burgers’18 equation:
∂u ∂u
+u = 0. (9.539)
∂t ∂x
This is more commonly known as the inviscid Burgers’ equation, but Bateman’s work has
priority.19 The inviscid Bateman-Burgers’ equation does not apply directly to a fluid; how-
ever, it contains one of the key mathematical features of fluid models: advective nonlinearity
that has profound effect on the dynamics.
Now consider a general transformation (x, t) → (ξ, τ ).

x = x(ξ, τ ), (9.540)
t = t(ξ, τ ). (9.541)

We assume the transformation to be unique and invertible. The chain rule gives
   ∂x ∂x   
dx dξ
= ∂ξ ∂t
∂τ
∂t . (9.542)
dt ∂ξ ∂τ
dτ
| {z }
J

The Jacobian matrix of the transformation is

 ∂x ∂x 
J = ∂ξ∂t
∂τ
∂t . (9.543)
∂ξ ∂τ

And we have for the Jacobian determinant J:

∂x ∂t ∂x ∂t
J = det J = − . (9.544)
∂ξ ∂τ ∂τ ∂ξ
17
Harry Bateman, 1882-1946, well regarded English mathematician who spent many years at Caltech
about whom von Kármán said “He seemed to know everything but did nothing important. I liked him.”
18
Johannes Martinus Burgers, 1895-1981, Dutch physicist.
19
The viscous version of the model equation, ∂u/∂t + u∂u/∂x = ν∂ 2 u/∂x2 , is widely known as Burgers’
equation and is often cited as originating from J. M. Burgers, 1948, A mathematical model illustrating the
theory of turbulence, Advances in Applied Mathematics, 1: 171-199. However, the viscous version was given
earlier by H. Bateman, 1915, Some recent researches in the motion of fluids, Monthly Weather Review, 43(4):
163-170.

© 05 December 2025, Joseph M. Powers. All rights reserved.

9.6. METHOD OF CHARACTERISTICS 413

Now from Eq. (2.286), we can deduce

 ∂   ∂   ∂t ∂t  ∂   ∂t ∂ ∂t ∂ 
T −1 ∂ξ 1 ∂τ
− ∂ξ ∂ξ 1 ∂τ ∂ξ
− ∂ξ ∂τ
∂x = (J ) = = . (9.545)
∂ ∂x ∂x
∂t
∂
∂τ
J − ∂τ ∂ξ
∂
∂τ
J − ∂x ∂
∂τ ∂ξ
+ ∂x ∂
∂ξ ∂τ

With these transformation rules, Eq (9.539) is rewritten as

   
1 ∂x ∂u ∂x ∂u 1 ∂t ∂u ∂t ∂u
− + +u − = 0. (9.546)
J ∂τ ∂ξ ∂ξ ∂τ J ∂τ ∂ξ ∂ξ ∂τ
| {z } | {z }
∂u/∂t ∂u/∂x

Now by assumption, J 6= 0, so we can multiply by J to get

∂x ∂u ∂x ∂u ∂t ∂u ∂t ∂u
− + +u −u = 0. (9.547)
∂τ ∂ξ ∂ξ ∂τ ∂τ ∂ξ ∂ξ ∂τ
Let us now restrict our transformation to satisfy the following requirements:
∂x ∂t
= u , (9.548)
∂τ ∂τ
t(ξ, τ ) = τ. (9.549)

The first says that if we insist that ξ is held fixed, that the ratio of the change in x to
the change in t will be u; this is equivalent to the more standard statement that on a
characteristic line we have dx/dt = u. The second is a convenience simply equating τ to t.
Applying the second restriction to the first, we can also say
∂x
= u. (9.550)
∂τ
With these restrictions, our inviscid Bateman-Burgers’ equation becomes
∂x ∂u ∂x ∂u ∂t ∂u ∂t ∂u
− + +u −u = 0, (9.551)
∂τ ∂ξ
|{z} ∂ξ ∂τ ∂τ ∂ξ
|{z} ∂ξ ∂τ
|{z}
u 1 0

∂u ∂x ∂u ✓
∂u
−u + + u ✓ = 0, (9.552)
∂ξ ∂ξ ∂τ ✓ ∂ξ
∂x ∂u
= 0. (9.553)
∂ξ ∂τ

Let us further require that ∂x/∂ξ 6= 0. Then we have

∂u
= 0, (9.554)
∂τ
u = f (ξ). (9.555)

© 05 December 2025, Joseph M. Powers. All rights reserved.

414 CHAPTER 9. ONE-DIMENSIONAL COMPRESSIBLE FLOW

Here f is an arbitrary function. Substitute this into Eq. (9.548) to get

∂x ∂t
= f (ξ) . (9.556)
∂τ ∂τ
We can integrate Eq. (9.556) to get

x = f (ξ)t + g(ξ). (9.557)

Here g(ξ) is an arbitrary function. Note the coordinate transformation can be chosen for
our convenience. To this end, remove t in favor of τ and set g(ξ) = ξ so that x maps to ξ
when t = τ = 0 giving

x(ξ, τ ) = f (ξ)τ + ξ. (9.558)

We can then state the solution to the inviscid Bateman-Burgers’ equation, Eq. (9.539),
parametrically as

u(ξ, τ ) = f (ξ), (9.559)

x(ξ, τ ) = f (ξ)τ + ξ, (9.560)
t(ξ, τ ) = τ. (9.561)

For this transformation, we have from Eq. (9.543) that

 df

1 + dξ τ f (ξ)
J= . (9.562)
0 1

Thus
df
J = det J = 1 + τ. (9.563)
dξ
We have a singularity in the coordinate transformation whenever J = 0, implying a difficulty
when
1
τ = − df . (9.564)
dξ

Example 9.14
Solve the inviscid Bateman-Burgers’ equation, ∂u/∂t + u∂u/∂x = 0, Eq. (9.539), if

u(x, 0) = 1 + sin πx, x ∈ [0, 1] (9.565)

Let us not be concerned with that portion of u that at t = 0 has x < 0 or x > 1. The analysis is easily
modified to address this.

© 05 December 2025, Joseph M. Powers. All rights reserved.

9.6. METHOD OF CHARACTERISTICS 415

t
u t=0 t=1/8 t=1/4 t=1/2 t=1
2.0

1.8

1.6 u
1.4

1.2

0.5 1.0 1.5 2.0 2.5

x
x

Figure 9.27: Solution to ∂u/∂t + u∂u/∂x = 0 with u(x, 0) = 1 + sin πx.

1.2 1.4 1.7
0.30

0.25 1.4

0.20 1.8
t 1.5 1.2
0.15 1.8
1.3 1.5
0.10 1.9

0.05

0.00
0.0 0.2 0.4 0.6 0.8 1.0 1.2
x

Figure 9.28: Early time solution to ∂u/∂t + u∂u/∂x = 0 with u(x, 0) = 1 + sin πx in the
form of a contour plot in (x, t) space giving contours of constant u.

We know the solution is given in general by Eqs. (9.559-9.561). At t = 0, we have τ = 0, and thus
x = ξ. And we have

f (ξ) = 1 + sin πξ. (9.566)

Thus we can say by inspection that the solution is

u(ξ, τ ) = 1 + sin πξ, (9.567)

x(ξ, τ ) = (1 + sin πξ) τ + ξ, (9.568)
t(ξ, τ ) = τ. (9.569)

Results for u(x, t) are plotted in Fig. 9.27. Another way to view the results is in the (x, t) diagram that
gives contours of u in Fig. 9.28 These contours are generated before the transformation has become
singular. The curves of constant u are the characteristics. Clearly those on the right of the maximum of
u are coalescing, while those to the left are diverging. The coalescence corresponds to shock formation,
and the divergence corresponds to a rarefaction.
One notes the following:

© 05 December 2025, Joseph M. Powers. All rights reserved.

416 CHAPTER 9. ONE-DIMENSIONAL COMPRESSIBLE FLOW

u
2.0

1.8

1.6

1.4

1.2

1.0
x
0.5 1.0 1.5 2.0 2.5 3.0

Figure 9.29: Sketch of response of u that satisfies the inviscid Bateman-Burgers’ equation
∂u/∂t + u∂u/∂x = 0 with u(x, 0) = 1 + sin πx.

• The signal propagates to the right; this is a consequence of u > 0 in the domain we consider.
• Portions of the signal with higher u propagate faster.
• The signal distorts as t increases.
• The wave appears to “break” at t = ts , where 1/4 . ts . 1/2. For t > ts , it is possible to find
multiple values of u at a given x and t. If u were a physical variable, we would not expect to see such
multivaluedness in nature.
It appears to be challenging to write an explicit formula for u(x, t). However, for small ξ, one can
write a useful approximation. Taylor series expansion of Eq. (9.568) for small ξ yields

x(ξ, τ ) ∼ (1 + πτ )ξ + τ + . . . . (9.570)

We invert this and use τ = t to find

x−t
ξ(x, t) ∼ + .... (9.571)
1 + πt
Then, because u = f (ξ), we get

π(x − t)
u(x, t) ∼ 1 + sin + .... (9.572)
1 + πt
This itself has a series expansion for small x and t of

u(x, t) ∼ 1 + π(x − t) + . . . . (9.573)

The sketch of Fig. 9.29 shows how one can envision the portion of the initial sine wave with x > 1/2
steepening, while that portion with x < 1/2 flattens. We place arrows whose magnitude is proportional
to the local value of u on the plot itself.
For our value of f (ξ), we have from from Eq. (9.563) that

J = 1 + πτ cos πξ. (9.574)

Clearly, there exist values of (ξ, τ ) for which J = 0. At such points, we can expect difficulties in our
solution. In Fig. 9.30, we plot a portion of the locus of points for which J = 0 in the (ξ, τ ) plane.

© 05 December 2025, Joseph M. Powers. All rights reserved.

9.6. METHOD OF CHARACTERISTICS 417

2.0 2.8 3.92 2.8 3.92

1.75
1.68

1.5 1.5 1.68

3.36 3.36
1.25

1.0 1

2.24
0.75

0.5 0.5

0.25 2.24

0.0 0.56 1.12

0.0 0.5 1.0 1.5 2.0

Figure 9.30: Curves where J = 0 and of constant x and t in the (ξ, τ ) plane for our coordinate
transformation.

We also see portions of this plane where the transformation is orientation-preserving, for which J > 0,
and orientation-reversing, for which J < 0. Also shown in Fig. 9.30 are contours of constant x and t.
Clearly when J = 0, the contours of constant x are parallel to those of constant t, and there are not
enough linearly independent vectors to form a basis.
From Eq. (9.564), we can expect a singular coordinate transformation when

1 1
τ = − df = − . (9.575)
dξ
π cos πξ

We then substitute this into Eqs. (9.568, 9.569) to get a parametric curve for when the transformation
is singular, xs (ξ), ts (ξ):

1 + sin πξ
xs (ξ) = − + ξ, (9.576)
π cos πξ
1
ts (ξ) = − . (9.577)
π cos πξ

A portion of this curve for where the transformation is singular is shown in Fig. 9.31. Figure 9.31a
plots xs (ξ) from Eq. (9.576). Figure 9.31b plots ts (ξ) from Eq. (9.577). We see a parametric plot of
the same quantities in Fig. 9.31c. At early time the system is free of singularities. It is easily shown
that both xs (ξ) and ts (ξ) have a local minimum at ξ = 1, at which point, we have

1
xs (1) = 1+ , (9.578)
π
1
ts (1) = . (9.579)
π
Examining Fig. 9.27, this appears to be the point at which the solution becomes multivalued. Examining
Fig. 9.30, this is the point on the curve J = 0 that is a local minimum. So while xs and ts are well-
behaved as functions of ξ for the domain considered, when the curves are projected into the (x, t) plane,
there is a cusp at (x, t) = (xs (1), ts (1)) = (1 + 1/π, 1/π).

© 05 December 2025, Joseph M. Powers. All rights reserved.

418 CHAPTER 9. ONE-DIMENSIONAL COMPRESSIBLE FLOW

2.2 1.2

2
2.0 1.0

1.8 0.8

1
1.6 0.6

1.4 0.4

0.6 0.8 1.0 1.2 1.4 0.6 0.8 1.0 1.2 1.4
1 2
a) b) c)

Figure 9.31: Plots indicating where the coordinate transformation of Eqs. (9.568,9.569) is
singular: a) xs (ξ) from Eq. (9.576), b) ts (ξ) from Eq. (9.577), c) representation of the curve
of singularity in (x, t) space.

Let us examine with Taylor series the behavior of ∂u/∂x in the neighborhood of the singularity.
Our expectation is that the slope approaches infinity as the singularity is approached. From Eq. (9.546),
we see that
 
∂u 1 ∂t ∂u ∂t ∂u
= − . (9.580)
∂x J ∂τ ∂ξ ∂ξ ∂τ

For our transformation, we have ∂t/∂τ = 1, ∂t/∂ξ = 0, so

∂u 1 ∂u
= . (9.581)
∂x J ∂ξ

Now use our solution for u, Eq. (9.559), and for J, Eq. (9.563), to say

∂u 1 df
= df dξ
. (9.582)
∂x 1 + dξ τ

With f (ξ) for our example from Eq. (9.566), we get

∂u π cos(πξ)
= . (9.583)
∂x 1 + πτ cos(πξ)

We use computer algebra to perform a Taylor series expansion of ∂u/∂ξ about ξ = 1, τ = 1/π to find
the behavior near the singularity to be
!
∂u 1 π 2
= +
2 (ξ − 1) + . . . (9.584)
∂x τ − π1 2 τ−1 π

By inspection we thus see that

∂u
lim1 = −∞. (9.585)
τ→π ∂x

It is necessary for τ to be increasing towards 1/π for the slope to be negative. This is the physically
relevant approach as τ begins at zero and is increasing.

© 05 December 2025, Joseph M. Powers. All rights reserved.

9.6. METHOD OF CHARACTERISTICS 419

This procedure can be extended to the Euler equations, though it is more complicated.
For the Euler equations, Courant and Friedrichs (1976) give some special solutions for rar-
efactions.

9.6.10 Viscous Bateman-Burgers’ equation solution

Our predictions of u(x, t) change dramatically when diffusion is introduced. Consider the
viscous Bateman-Burgers’ equation:

∂u ∂u ∂2u
+u = ν 2. (9.586)
∂t ∂x ∂x
When we discretized the partial differential equations and simulate via standard numeri-
cal methods the same problem whose diffusion-free solution is plotted in Fig. 9.27 for which
u(x, 0) = 1 + sin πx, we obtain the results plotted in Fig. 9.32 for four different values of
ν = 1/1000, 1/100, 1/10, and 1. While an exact solution to the viscous Bateman-Burgers’
equation is available, in practice, it is complicated. It is often easier to obtain results by
numerical discretization, and that is what we did here. The scheme used was sufficiently
resolved to capture the thin zones present when ν was small. For the case where ν = 1/100,
we plot the (x, t) diagram, where the shading is proportional to the local value of u, in
Fig. 9.33.
We note:

• We restricted our study to positive values of ν, that can be shown to be necessary for
a stable solution as t → ∞.

• If ν = 0, our viscous Bateman-Burgers’ equation reduces to the inviscid Bateman-

Burgers’ equation.

• As ν → 0, solutions to the viscous Bateman-Burgers’ equation seem to relax to a

solution with an infinitely thin discontinuity; they do not relax to those solutions
displayed in Fig. 9.27.

• For all values of ν, the solution u(x, t) at a given time has a single value of u for a
single value of x, in contrast to multi-valued solutions exhibited by the diffusion-free
analog.

• As ν → 0, the peaks retain a larger magnitude. Thus one can conclude that enhancing
ν smears peaks.

• At early time the solutions to the viscous Bateman-Burgers’ equation resemble those
of the inviscid Bateman-Burgers’ equation.

© 05 December 2025, Joseph M. Powers. All rights reserved.

420 CHAPTER 9. ONE-DIMENSIONAL COMPRESSIBLE FLOW

u u
t=0 t=1/8 t=1/4 t=1/2 t=0 t=1/8 t=1/4
2.0 2.0 t=1/2
t=1
t=1
1.8 1.8

1.6 1.6

1.4 1.4

1.2 1.2

1.0 1.0
x x
0.5 1.0 1.5 2.0 2.5 3.0 0.5 1.0 1.5 2.0 2.5 3.0
a) b)
u u
t=0 t=0
2.0 2.0
t=1/8

1.8 t=1/4 1.8

t=1/2
1.6 1.6
t=1
t=1/8
1.4 1.4
t=1/4
1.2 1.2
t=1/2

1.0 1.0 t=1

0.5 1.0 1.5 2.0 2.5 3.0 x 0.5 1.0 1.5 2.0 2.5 3.0 x
c) d)

Figure 9.32: Numerical solution to the viscous Bateman-Burgers’ equation ∂u/∂t+u∂u/∂x =

ν∂ 2 u/∂x2 with u(x, 0) = 1 + sin πx and various values of ν: a) 1/1000, b) 1/100, c) 1/10, d)
1.

1.0 1.1 1.1

1.5
0.8 1.5

0.6 1.4 1.3

t
0.4 1.3 1.4

0.2 1.6

0.0 1.2
0.0 0.5 1.0 1.5 2.0 2.5 3.0
x

Figure 9.33: (x, t) diagram with contours of u for solution to the viscous Bateman-Burgers’
equation ∂u/∂t + u∂u/∂x = ν∂ 2 u/∂x2 with u(x, 0) = 1 + sin πx, ν = 1/100.

© 05 December 2025, Joseph M. Powers. All rights reserved.

9.6. METHOD OF CHARACTERISTICS 421

Let us try to understand this behavior. Fundamentally, it will be seen that in many cases,
nonlinearity, manifested in u∂u/∂x can serve to steepen a waveform. If that steepening is
unchecked by diffusion, either a formal discontinuity is admitted, or multi-valued solutions.
Now diffusion acts most strongly when gradients are steep, that is when ∂u/∂x has large
magnitude. As a wave steepens due to nonlinear effects, diffusion, that may have been
initially unimportant, can reassert its importance and serve to suppress the growth due to
the nonlinearity.

© 05 December 2025, Joseph M. Powers. All rights reserved.

422 CHAPTER 9. ONE-DIMENSIONAL COMPRESSIBLE FLOW

© 05 December 2025, Joseph M. Powers. All rights reserved.

Chapter 10

Viscous incompressible laminar flow

see Powers, Chapters 10 and 11,

see Panton, Chapters 7 and 11,
see Yih, Chapter 7,
see Segel, Chapter 3,
see White and Majdalani.

Here we consider a few standard problems in viscous incompressible laminar flow. For this
entire chapter, we will make the following assumptions:

• the flow is incompressible,

• body forces are negligible, and

• the fluid properties, cp , µ, and k, are constants.

10.1 Fully developed, one-dimensional solutions

The first type of solution we will consider is known as a one-dimensional fully developed
solution. These are commonly considered in first courses in fluid mechanics and heat transfer.
The flows here are essentially one-dimensional, but not absolutely, as they were in the chapter
on one-dimensional compressible flow. In this section, we will further enforce that

• the flow is time-independent, ∂o = 0,

• the velocity and temperature gradients in the x and z directions are zero, ∂v/∂x = 0,
∂v/∂z = 0, ∂T /∂x = 0, ∂T /∂z = 0.

We will see that these assumptions give rise to flows with a non-zero x velocity u that varies
in the y direction, and that other velocities v, and w, will be zero.

423
424 CHAPTER 10. VISCOUS INCOMPRESSIBLE LAMINAR FLOW

b
y
z

Figure 10.1: Pressure gradient-driven flow in a slot.

10.1.1 Pressure gradient-driven flow in a slot

Consider the flow sketched in Fig. 10.1. Here we have a large reservoir of fluid with a long
narrow slot located around y = 0. We take the length of the slot in the z direction, b, to
be long relative to the slot width in the y direction h. Attached to the slot are two parallel
plates, separated by distance in the y direction h. The length of the plates in the x direction
is L. We take L ≫ h. Because of gravity forces, that we neglect in the slot, the pressure
at the entrance of the slot po is higher than atmospheric. At the end of the slot, the fluid
expels to the atmosphere that is at p1 . Hence, there is a pressure gradient in the x direction,
that drives the flow in the slot. We will see that the flow is resisted by viscous stresses. An
analogous flow in a circular duct is defined as a Hagen1 -Poiseuille2 flow.
Near x = 0, the flow accelerates in what is known as the entrance length. If L is sufficiently
long, we observe that sufficiently downstream of x = 0, the fluid particles no longer accelerate.
It is at this point where the viscous shear forces exactly balance the pressure forces to give
rise to the fully developed velocity field.
For this flow, let us make the additional assumptions that

• there is no imposed pressure gradient in the z direction, and

• the walls are held at a constant temperature, To .

• the fluid is liquid with thermal expansion coefficient β ∼ 0.

1
Gotthilf Ludwig Hagen, 1797-1884, German engineer who measured velocity of water in small diameter
tubes.
2
Jean Louis Poiseuille, 1799-1869, French physician who repeated experiments of Hagen for simulated
blood flow.

© 05 December 2025, Joseph M. Powers. All rights reserved.

10.1. FULLY DEVELOPED, ONE-DIMENSIONAL SOLUTIONS 425

Incorporating some of these assumptions, we recast the incompressible constant property

Navier-Stokes equations of Ch. 6.4.6, Eqs. (6.89-6.90) for β = 0, fi = 0, as

∂i vi = 0, (10.1)
ρ∂o vi + ρvj ∂j vi = −∂i p + µ∂j ∂j vi , (10.2)
ρcp ∂o T + ρcp vj ∂j T = k∂i ∂i T + 2µ∂(i vj) ∂(i vj) . (10.3)

Here we have five equations in five unknowns, vi , p, and T .

As for all incompressible flows with constant properties, we can get the velocity field by
only considering the mass and momenta equations; velocity is only coupled one way to the
energy equation. The mass equation, recalling that gradients in x and z are zero, gives us

∂ ∂ ∂
u+ v+ w = 0. (10.4)
∂x
|{z} ∂y ∂z
|{z}
=0 =0

So the mass equation gives us

∂v
= 0. (10.5)
∂y
Now, from our assumptions of steady and fully developed flow, we know that v cannot be a
function of x, z, or t. So the partial becomes a total derivative, and mass conservation holds
that dv/dy = 0. Integrating, we find that v(y) = C. The constant C must be zero, because
we must satisfy a no-penetration boundary condition at either wall that v(y = h/2) = v(y =
−h/2) = 0. Hence, mass conservation, coupled with the no-penetration boundary condition
gives us
v = 0. (10.6)
Now consider the x momentum equation:
 
∂ ∂ ∂ ∂ ∂p  ∂2 ∂2 ∂2 
ρ u + ρu u + ρ |{z}
v u + ρw u = − + µ  2 u + 2 u + 2 u ,
∂t
|{z} ∂x
|{z} ∂y ∂z
|{z} ∂x ∂x
|{z} ∂y ∂z
|{z}
=0
=0 =0 =0 =0 =0
(10.7)
2
∂p ∂ u
0 = − + µ 2. (10.8)
∂x ∂y

For this fully developed flow the acceleration, that is the material derivative of velocity, is
formally zero, and the equation gives rise to a balance of pressure and viscous surface forces.
For the y momentum equation, we get

∂ ∂ ∂ ∂ ∂p
ρ v +ρu
|{z} v +ρ |{z}
|{z} v v +ρw
|{z} v
|{z} = − (10.9)
∂t
|{z} ∂x
|{z} ∂y ∂z
|{z} ∂y
=0 =0 =0 =0 =0
=0 =0 =0

© 05 December 2025, Joseph M. Powers. All rights reserved.

426 CHAPTER 10. VISCOUS INCOMPRESSIBLE LAMINAR FLOW
 
2 2 2
∂ ∂ ∂ 
+µ  2 v + 2 |{z}
v + 2 v ,
∂x
|{z} ∂y ∂z
|{z}
=0
=0 =0
∂p
0 = . (10.10)
∂y
Hence, p = p(x, z), but because we have assumed there is no pressure gradient in the z
direction, we have at most that
p = p(x). (10.11)
For the z momentum equation we get:
∂ ∂ ∂ ∂ ∂p
ρ w + ρu w + ρ |{z}
v w + ρw w = − (10.12)
∂t
|{z} ∂x
|{z} ∂y ∂z
|{z} ∂z
|{z}
=0
=0 =0 =0 =0
 
 ∂2 ∂2 ∂2 
+µ  2 w + 2 w + 2 w  ,
∂x
|{z} ∂y ∂z
|{z}
=0 =0
∂2w
0 = . (10.13)
∂y 2
Solution of this partial differential equation gives us

w = f (x, z)y + g(x, z). (10.14)

Now to satisfy the no-slip condition, we must have w = 0 at y = ±h/2. This leads us to two
linear equations for f and g:
 h    
2
1 f (x, z) 0
= . (10.15)
− h2 1 g(x, z) 0

Because the determinant of the coefficient matrix, h/2 + h/2 = h, is non-zero, the only
solution is the trivial solution f (x, z) = g(x, z) = 0. Hence,

w = 0. (10.16)

Next consider how the energy equation reduces:

   
∂  ∂ ∂ ∂   ∂2 ∂2 ∂2 
ρcp T + ρcp u T + |{z}
v T + |{z}
w T = k  2 T + 2T + 2 T
∂t
|{z} ∂x
|{z} ∂y ∂z ∂x ∂y ∂z
=0 =0 |{z} |{z} |{z}
=0 =0 =0 =0 =0
+2µ∂(i vj) ∂(i vj) , (10.17)
2
∂ T
0 = k 2 + 2µ∂(i vj) ∂(i vj) . (10.18)
∂y

© 05 December 2025, Joseph M. Powers. All rights reserved.

10.1. FULLY DEVELOPED, ONE-DIMENSIONAL SOLUTIONS 427

There is no tendency for a particle’s temperature to increase. There is a balance between

thermal energy generated by viscous dissipation and that conducted away by energy diffusion.
Thus the energy path is 1) viscous work is done to generate thermal energy, 2) thermal energy
diffuses throughout the channel and out the boundary. Now consider the viscous dissipation
term for this flow.
 
∂1 v1 ∂1 v2 ∂1 v3
|{z} |{z} |{z} |{z} |{z}  
 =0 =0 
 =0 =0 =0  0 0 0
 ∂v ∂2 v2 ∂2 v3  
∂i vj =  2 1 |{z} |{z}  = ∂2 v1 0 0  , (10.19)
 =0 =0 
 ∂ v ∂3 v2 ∂3 v3  0 0 0
3 1
|{z} |{z} |{z} |{z} |{z}

=0 =0 =0  =0
=0  
 0 1 ∂2 v1 + ∂1 v2  0 
 2 |{z}   0 1 ∂u
0

   =0  2 ∂y
   1 ∂u
∂(i vj) =   = 2 ∂y 0 0  .(10.20)
1 
 2 ∂2 v1 + ∂1 v2  0 0 0 0 0
 |{z} 
=0
0 0 0
Further,  2  2  2
1 ∂u 1 ∂u 1 ∂u
∂(i vj) ∂(i vj) = + = . (10.21)
2 ∂y 2 ∂y 2 ∂y
So the energy equation becomes finally
 2
∂2T ∂u
0=k 2 +µ . (10.22)
∂y ∂y
At this point we have the x momentum and energy equations as the only two that seem
to have any substance.
∂p ∂2u
0 = − + µ 2, (10.23)
∂x ∂y
2
 2
∂ T ∂u
0 = k 2 +µ . (10.24)
∂y ∂y
This looks like two equations in three unknowns. One peculiarity of incompressible equations
is that there is always some side condition, that ultimately hinges on the mass equation, that
really gives a third equation. Without going into details, it involves for general flows solving
a Poisson equation for pressure that is of the form −∇2 p = f (u, v, w). Sec. 6.4.4 has some
of those details. The Poisson equation involves second derivatives of pressure. Here we can
obtain a simple form of this general equation by taking the partial derivative with respect
to x of the x momentum equation:
∂2p ∂ ∂2u
0 = − + µ , (10.25)
∂x2 ∂x ∂y 2

© 05 December 2025, Joseph M. Powers. All rights reserved.

428 CHAPTER 10. VISCOUS INCOMPRESSIBLE LAMINAR FLOW

∂2p ∂ 2 ∂u
0 = − + µ . (10.26)
∂x2 ∂y 2 |{z}
∂x
=0

The viscous term here is zero because of our assumption of fully developed flow. Moreover,
because p = p(x) only, we then get

d2 p
= 0, p(0) = po , p(L) = p1 , (10.27)
dx2
that has a solution showing the pressure field must be linear in x:
po − p1
p(x) = po − x, (10.28)
L
dp po − p1
= − . (10.29)
dx L
Now, because u is at most a function of y, we can convert partial derivatives to ordinary
derivatives, and write the x momentum equation and energy equation as two ordinary
differential equations in two unknowns with appropriate boundary conditions at the wall
y = ±h/2:
   
d2 u po − p1 h h
= − , u = 0, u − = 0, (10.30)
dy 2 µL 2 2
 2    
d2 T µ du h h
= − , T = To , T − = To . (10.31)
dy 2 k dy 2 2

We could solve these equations directly, but instead let us first cast them in dimensionless
form. This will give our results some universality and efficiency. Moreover, it will reveal more
fundamental groups of terms that govern the fluid behavior. Let us select scales such that
dimensionless variables, denoted by a * subscript, are as follows
y T − To u
y∗ = , T∗ = , u∗ = . (10.32)
h To uc
We have yet to determine the characteristic velocity uc . The dimensionless temperature
has been chosen to render it zero at the boundaries. With these choices, the x momentum
equation becomes

u c d2 u ∗ po − p1
= − , (10.33)
h2 dy∗2 µL
d2 u ∗ (po − p1 )h2
= − , (10.34)
dy∗2 µLuc
uc u∗ (x∗ h = h/2) = uc u∗ (x∗ h = −h/2) = 0, (10.35)
u∗ (x∗ = 1/2) = u∗ (x∗ = −1/2) = 0. (10.36)

© 05 December 2025, Joseph M. Powers. All rights reserved.

10.1. FULLY DEVELOPED, ONE-DIMENSIONAL SOLUTIONS 429

Let us now choose the characteristic velocity to render the x momentum equation to have a
simple form:
(po − p1 )h2
uc ≡ . (10.37)
µL
Now scale the energy equation:
 2
To d2 T∗ µu2c du∗
= − , (10.38)
h2 dy∗2 kh2 dy∗
2 
d2 T∗ µu2c
du∗
= − , (10.39)
dy∗2 kTo
dy∗
 2
µcp u2c du∗
= − , (10.40)
k cp To dy∗
 2
du∗
= −P rEc , (10.41)
dy∗
   
1 1
T∗ − = T∗ = 0. (10.42)
2 2

Here we have grouped terms so that the Prandtl number, Eq. (6.166), P r = µcp /k, explicitly
appears. Further, we have defined the Eckert3 number Ec as
 2
(po −p1 )h2
u2c µL
Ec = = . (10.43)
cp To cp To

In summary our dimensionless differential equations and boundary conditions are

 
d2 u ∗ 1
= −1, u ± = 0, (10.44)
dy∗2 2
 2  
d2 T∗ du∗ 1
2
= −P rEc , T∗ ± = 0. (10.45)
dy∗ dy∗ 2

These boundary conditions are homogeneous; hence, they do not contribute to a non-trivial
solution. The pressure gradient is an inhomogeneous forcing term in the momentum equa-
tion, and the viscous dissipation is a forcing term in the energy equation.
The solution for the velocity field that satisfies the differential equation and boundary
conditions is quadratic in y∗ and is
 2 !
1 1
u∗ = − y∗2 . (10.46)
2 2

3
Ernst R. G. Eckert, 1904-2004, scholar of convective heat transfer.

© 05 December 2025, Joseph M. Powers. All rights reserved.

430 CHAPTER 10. VISCOUS INCOMPRESSIBLE LAMINAR FLOW

Figure 10.2: Velocity profile for pressure gradient-driven flow in a slot.

The maximum velocity occurs at y∗ = 0 and has value

1
u∗max = . (10.47)
8
The mean velocity is found through integrating the velocity field to arrive at
Z 1/2
u∗mean = u∗ (y∗ ) dy∗ , (10.48)
−1/2
Z 1/2  2 !
1 1 2
= − y∗ dy∗ , (10.49)
−1/2 2 2
  1/2
1 1 1 3
= y∗ − y∗ , (10.50)
2 4 3 −1/2
1
= . (10.51)
12
We could have scaled the velocity field in such a fashion that either the maximum or the
mean velocity was unity. The scaling we chose gave rise to a non-unity value of both. In
dimensional terms we could say
 2   !
u 1 1 y 2
(po −p1 )h2
= − . (10.52)
2 2 h
µL

The velocity profile is sketched in Fig. 10.2. This flow is rotational. For the two-dimensional
flow, the only component of vorticity is in the z∗ direction, and we have
∂v∗ ∂u∗
ωz∗ = − , (10.53)
∂x∗ ∂y∗
|{z}
=0
= y∗ . (10.54)

© 05 December 2025, Joseph M. Powers. All rights reserved.

10.1. FULLY DEVELOPED, ONE-DIMENSIONAL SOLUTIONS 431

The vorticity magnitude is maximum at the solid walls at y∗ = ±1/2, and it is zero at the
centerline, y∗ = 0. The deformation tensor is
  
∂u∗ 1 ∂u∗ ∂v∗
+ ∂x∗  y∗

∂x∗ 2 ∂y∗ 0 −
D=   = 2 . (10.55)
1 ∂u∗
+ ∂v∗ ∂v∗ − y2∗ 0
2 ∂y∗ ∂x∗ ∂y∗

It is easy to show the eigenvalues of D are given by λ = ±y∗ /2 and the eigenvectors are
at angles of π/4 and 3π/4 to the horizontal. So on these axes exists the rate of extreme
extensional straining.
Now let us get the temperature field.
 2 !!!2
d2 T∗ d 1 1 2
= −P rEc − y∗ , (10.56)
dy∗2 dy∗ 2 2
= −P rEc (−y∗ )2 , (10.57)
= −P rEc y∗2 , (10.58)
dT∗ 1
= − P rEc y∗3 + C1 , (10.59)
dy∗ 3
1
T∗ = − P rEc y∗4 + C1 y∗ + C2 , (10.60)
12
1 1 1 1
0 = − P rEc + C1 + C2 , y∗ = , (10.61)
12 16 2 2
1 1 1 1
0 = − P rEc − C1 + C2 , y∗ = − , (10.62)
12 16 2 2
P rEc
C1 = 0, C2 = . (10.63)
192
Regrouping, we find that !
 4
P rEc 1 4
T∗ = − y∗ . (10.64)
12 2
In terms of dimensional quantities, we can say
2 4
 4   !
T − To (po − p1 ) h 1 y 4
= − . (10.65)
To 12µL2 kTo 2 h

The temperature profile is sketched in Fig. 10.3.

From knowledge of the velocity and temperature field, we can calculate other quantities
of interest. Let us calculate the field of shear stress and heat flux, and then evaluate both
at the wall. First for the shear stress, recall that in dimensional form we have

τij = 2µ∂(i vj) + λ ∂k vk δij , (10.66)

|{z}
=0
= 2µ∂(i vj) . (10.67)

© 05 December 2025, Joseph M. Powers. All rights reserved.

432 CHAPTER 10. VISCOUS INCOMPRESSIBLE LAMINAR FLOW

Figure 10.3: Temperature profile for pressure gradient-driven flow in a slot.

We have already seen the only non-zero components of the symmetric part of the velocity
gradient tensor are the 12 and 21 components. Thus, the 21 stress component is
∂ v + ∂ v 
2 1 1 2
|{z}
 =0 
τ21 = 2µ∂(2 v1) = 2µ  , (10.68)
2

= µ∂2 v1 . (10.69)

In (x, y) space, we then say here that

du
τyx = µ . (10.70)
dy

This is a stress on the y (tangential) face that points in the x direction; hence, it is certainly
a shearing stress. In dimensionless terms, we can define a characteristic shear stress τc , so
that the scale shear is τ∗ = τyx /τc . Thus, our equation for shear becomes

µuc du∗
τc τ∗ = . (10.71)
h dy∗
Now take  
µuc µ(po − p1 )h2 h
τc ≡ = = (po − p1 ) . (10.72)
h hµL L
With this definition, we get
du∗
τ∗ = . (10.73)
dy∗
Evaluating for the velocity profile of the pressure gradient-driven flow, we find

τ∗ = −y∗ . (10.74)

© 05 December 2025, Joseph M. Powers. All rights reserved.

10.1. FULLY DEVELOPED, ONE-DIMENSIONAL SOLUTIONS 433

Figure 10.4: Shear stress profile for pressure gradient-driven flow in a slot.

The stress is zero at the centerline y∗ = 0 and has maximum magnitude of 1/2 at either
wall, y∗ = ±1/2. In dimensional terms, the wall shear stress τw is
 
1 h
τw = − (po − p1 ) . (10.75)
2 L

The wall shear stress is governed by the pressure difference and not the viscosity. However,
the viscosity plays a determining role in selecting the maximum fluid velocity. The shear
profile is sketched in Fig. 10.4.
Next, let us calculate the heat flux vector. Recall that, for this flow, with no x or z
variation of T , we have the only non-zero component of the heat flux vector as

dT
qy = −k . (10.76)
dy

Now define scale the heat flux by a characteristic heat flux qc , to be determined, to obtain
a dimensionless heat flux:
qy
q∗ = . (10.77)
qc
So,

kTo dT∗
qc q∗ = − , (10.78)
h dy∗
kTo dT∗
q∗ = − . (10.79)
hqc dt∗

Let qc ≡ kTo /h, so

dT∗
q∗ = − , (10.80)
dy∗
1
= P rEc y∗3. (10.81)
3

© 05 December 2025, Joseph M. Powers. All rights reserved.

434 CHAPTER 10. VISCOUS INCOMPRESSIBLE LAMINAR FLOW

Figure 10.5: Heat flux profile for pressure gradient-driven flow in a slot.

For our flow, we have a cubic variation of the heat flux vector magnitude. There is no heat
flux at the centerline, that corresponds to this being a region of no shear. The magnitude
of the heat flux is maximum at the wall, the region of maximum shear. At the upper wall,
we have
1
q∗ |y∗ =1/2 = P rEc. (10.82)
24
The heat flux profile is sketched in Fig. 10.5. In dimensional terms we have
qw 1 (po − p1 )2 h4
kTo
= , (10.83)
h
24 µL2 kTo
1 (po − p1 )2 h3
qw = . (10.84)
24 µL2

10.1.2 Couette flow with pressure gradient

We next consider Couette flow with a pressure gradient. Couette flow implies that there is a
moving plate at one boundary and a fixed plate at the other. It is a common experimental
configuration, and used often to actually determine a fluid’s viscosity. Here we will take
the same assumptions as for pressure gradient-driven flow in a slot, expect for the boundary
condition at the upper surface, that we will require to have a constant velocity U. We will
also shift the coordinates so that y = 0 matches the lower plate surface and y = h matches
the upper plate surface. The configuration for this flow is shown in Fig. 10.6.
Our equations governing this flow are
d2 u po − p1
2
= − , u(0) = 0, u(h) = U, (10.85)
dy µL
 2
d2 T µ du
= − , T (0) = To , T (h) = To . (10.86)
dy 2 k dy
Once again in momentum, there is no acceleration, and viscous stresses balance shear stresses.
In energy, there is no energy increase, and generation of thermal energy due to viscous work
is balanced by diffusion of the thermal energy, ultimately out of the system through the

© 05 December 2025, Joseph M. Powers. All rights reserved.

10.1. FULLY DEVELOPED, ONE-DIMENSIONAL SOLUTIONS 435

L
U

Figure 10.6: Configuration for Couette flow with pressure gradient.

boundaries. Here there are inhomogeneities in both the forcing terms and the boundary
conditions. In terms of work, both the pressure gradient and the pulling of the plate induce
work.
Once again let us scale the equations. This time, we have a natural velocity scale, U, the
upper plate velocity. So take

y T − To u
y∗ = , T∗ = , u∗ = . (10.87)
h To U

The momentum equation becomes

U d2 u ∗ po − p1
2 2
= − , (10.88)
h dy∗ µL
d2 u ∗ (po − p1 )h2
= − . (10.89)
dy∗2 µUL

With dimensionless pressure gradient

(po − p1 )h2
P≡ , (10.90)
µUL

we get

d2 u ∗
= −P, (10.91)
dy∗2
u∗ (0) = 0, u∗ (1) = 1. (10.92)

This has solution

1
u∗ = − Py∗2 + C1 y∗ + C2 . (10.93)
2

© 05 December 2025, Joseph M. Powers. All rights reserved.

436 CHAPTER 10. VISCOUS INCOMPRESSIBLE LAMINAR FLOW

2
=-

=0
=-

=2
Figure 10.7: Velocity profiles for various values of P for Couette flow with pressure gradient.

Applying the boundary conditions, we get

1
0 = − P(0)2 + C1 (0) + C2 , (10.94)
2
= C2 , (10.95)
1
1 = − P(1)2 + C1 (1), (10.96)
2
1
C1 = 1 + P, (10.97)
2  
1 2 1
u∗ = − Py∗ + 1 + P y∗ , (10.98)
2 2
1
= Py∗ (1 − y∗ ) + y∗ . (10.99)
2
| {z } |{z}
Couette effect
pressure effect

We see that the pressure gradient generates a velocity profile that is quadratic in y∗ . This
is distinguished from the Couette effect, that is the effect of the upper plate’s motion, that
gives a linear profile. Because our governing equation here is linear, it is appropriate to
think of these as superposed solutions. Velocity profiles for various values of P are shown in
Fig. 10.7.
Let us now calculate the shear stress profile. With τ = µ(du/dy), and taking τ∗ = τ /τc ,
we get
µU du∗
τc τ∗ = , (10.100)
h dy∗
µU du∗
τ∗ = , (10.101)
hτc dy∗
µU
taking τc ≡ , (10.102)
h
du∗
τ∗ = , so here, (10.103)
dy∗
1
= −Py∗ + P + 1, and (10.104)
2

© 05 December 2025, Joseph M. Powers. All rights reserved.

10.1. FULLY DEVELOPED, ONE-DIMENSIONAL SOLUTIONS 437

1
τ∗ |y∗ =0 = P + 1, (10.105)
2
1
τ∗ |y∗ =1 = − P + 1. (10.106)
2
The wall shear has a pressure gradient effect and a Couette effect as well. In fact we can
select a pressure gradient to balance the Couette effect at one or the other wall, but not
both.
We can also calculate the dimensionless volume flow rate Q∗ , that for incompressible
flow, is directly proportional to the mass flux. Ignoring how the scaling would be done, we
arrive at
Z 1
Q∗ = u∗ dy∗ , (10.107)
0
Z 1   
1 2 1
= − Py∗ + 1 + P y∗ dy∗ , (10.108)
0 2 2
 1   2 1
1 3 1 y∗
= − Py∗ + 1 + P , (10.109)
6 0 2 2 0
 
P 1 1
= − + 1+ P , (10.110)
6 2 2
P 1
= + . (10.111)
12 2
Again there is a pressure gradient contribution and a Couette contribution, and we could
select P to give no net volume flow rate. We summarize some of the special cases as follows

• P → −∞: u∗ = (1/2)Py∗ (1 − y∗ ); τ∗ = P (1/2 − y∗ ), Q∗ = P/12. Here the fluid flows

in the opposite direction as driven by the plate because of the large pressure gradient.

• P = −6. Here we get no net mass flow and u∗ = 3y∗2 − 2y∗ , τ∗ = 2y∗ , Q∗ = 0.

• P = −2. Here we get no shear at the bottom wall and u∗ = y∗2 , τ∗ = 2y∗ , Q∗ = 1/3.

• P = 0. Here we have no pressure gradient and u∗ = y∗ , τ∗ = 1, Q∗ = 1/2.

• P = 2. Here we get no shear at the top wall and u∗ = −y∗2 + 2y∗ , τ ∗ = −2y∗ + 2,
Q∗ = 2/3.

• P → ∞: u∗ = (1/2)Py∗ (1 − y∗ ); τ∗ = P (1/2 − y∗ ), Q∗ = P/12. Here the fluid flows

in the same direction as driven by the plate.

We now consider the heat transfer problem. Scaling, we get

 2
To d2 T∗ µU 2 du∗
= − , T∗ (0) = T∗ (1) = 0, (10.112)
h2 dy∗2 kh2 dy∗

© 05 December 2025, Joseph M. Powers. All rights reserved.

438 CHAPTER 10. VISCOUS INCOMPRESSIBLE LAMINAR FLOW

 2
d2 T∗ µU 2 du∗
= − , (10.113)
dy∗2 kTo dy∗
 2
µc U 2 du∗
= − , (10.114)
k cTo dy∗
 2
du∗
= −P rEc , (10.115)
dy∗
= −P rEc τ∗2 , (10.116)
 2
1
= −P rEc −Py∗ + P + 1 , (10.117)
2
   2 !
1 1
= −P rEc P 2 y∗2 − 2P P + 1 y∗ + 1 + P , (10.118)
2 2
2
   2 !
dT∗ P 3 1 1
= −P rEc y∗ − P P + 1 y∗2 + 1 + P y∗ + C1 , (10.119)
dy∗ 3 2 2
   2 !
P2 4 P 1 1 1
T∗ = −P rEc y∗ − P + 1 y∗3 + 1 + P y∗2
12 3 2 2 2
+C1 y∗ + C2 , (10.120)
T∗ (0) = 0 = C2 , (10.121)
   2 !
P2 P 1 1 1
T∗ (1) = 0 = −P rEc − P +1 + 1+ P + C1 , (10.122)
12 3 2 2 2
 
1 P P2
C1 = P rEc + + , (10.123)
2 6 24
   2 !
P2 4 P 1 1 1
T∗ = −P rEc y − P + 1 y∗3 + 1 + P y∗2
12 ∗ 3 2 2 2
 
1 P P2
+P rEc + + y∗ . (10.124)
2 6 24

Factoring, we can write the temperature profile as

P rEc
T∗ = y∗ (1 − y∗ )(12 + 4P + P 2 − 8Py∗ − 2P 2 y∗ + 2P 2 y∗2). (10.125)
24

For the wall heat transfer, recall qy = −k(dT /dy). Scaling, we get

kTo dT∗
qc q∗ = − , (10.126)
h dy∗
kTo dT∗
q∗ = − . (10.127)
hqc dy∗

© 05 December 2025, Joseph M. Powers. All rights reserved.

10.2. SIMILARITY SOLUTIONS 439

Now choose qc such that

kTo
qc ≡ , (10.128)
h
dT∗
q∗ = − . (10.129)
dy∗
So
   2 !
P2 3 1 1 1 P P 2
q∗ = P rEc y −P P + 1 y∗2 + 1 + P y∗ − − − . (10.130)
3 ∗ 2 2 2 6 24

At the bottom wall y∗ = 0, we get for the heat transfer vector

 
1 P P2
q∗ |y∗ =0 = −P rEc + + . (10.131)
2 6 24

10.2 Similarity solutions

In this section, we will consider problems that can be addressed by what is known as a
similarity transformation. The problems themselves will be fundamental ones that have
variation in either time and one spatial coordinate, or with two spatial coordinates. This is
in contrast with solutions of the previous section that varied only with one spatial coordinate.
Because two coordinates are involved, we must resort to solving partial differential equa-
tions. The similarity transformation actually reveals a hidden symmetry of the partial dif-
ferential equations by defining a new independent variable, that is a grouping of the original
independent variables, under which the partial differential equations transform into ordinary
differential equations. We then solve the resulting ordinary differential equations by standard
techniques.

10.2.1 Stokes’ first problem

The first problem we will consider that uses a similarity transformation is known as Stokes’
first problem, as Stokes addressed it in his original work that developed the Navier-Stokes
equations in the mid-nineteenth century.4 The problem is described as follows, and is
sketched in Fig. 10.8. Consider a flat plate of infinite extent lying at rest for t < 0 on
the y = 0 plane in (x, y, z) space. In the volume described by y > 0 exists a fluid of semi-
infinite extent that is at rest at time t < 0. At t = 0, the flat plate is suddenly accelerated to
a constant velocity of U, entirely in the x direction. Because the no-slip condition is satisfied
for the viscous flow, this induces the fluid at the plate surface to acquire an instantaneous
velocity of u(0) = U. Because of diffusion of linear x momentum via tangential viscous shear
4
Stokes, G. G., 1851, “On the effect of the internal friction of fluids on the motion of pendulums,” Trans-
actions of the Cambridge Philosophical Society, 9(2): 8-106.

© 05 December 2025, Joseph M. Powers. All rights reserved.

440 CHAPTER 10. VISCOUS INCOMPRESSIBLE LAMINAR FLOW

y
x
U

Figure 10.8: Schematic for Stokes’ first problem of a suddenly accelerated plate diffusing
linear momentum into a fluid at rest.

forces, the fluid in the region above the plate begins to acquire a positive velocity in the x
direction as well. We will use the Navier-Stokes equations to quantify this behavior. Let
us make identical assumptions as we did in the previous section, except that 1) we will not
neglect time derivatives, as they are an obviously important feature of the flow, and 2) we
will assume all pressure gradients are zero; hence the fluid has a constant pressure.
Under these assumptions, the x momentum equation,
 
∂ ∂ ∂ ∂ ∂p  ∂2 ∂2 ∂2 
ρ u + ρu u + ρ |{z}
v u + ρw u=− +µ  2 u + 2 u + 2 u ,
∂t ∂x
|{z} ∂y ∂z
|{z} ∂x
|{z} ∂x
|{z} ∂y ∂z
|{z}
=0
=0 =0 =0 =0 =0
(10.132)

is the only relevant component of linear momenta, and reduces to

∂u ∂2u
ρ = µ 2 . (10.133)
∂t
|{z} ∂y
| {z }
(mass)(acceleration) shear force

The energy equation reduces as follows

   
∂  ∂ ∂ ∂   ∂2 ∂2 ∂2 
ρcp T + ρcp u T + |{z}
v T + |{z}
w T = k  2 T + 2T + 2 T
∂t ∂x
|{z} ∂y ∂z ∂x ∂y ∂z
=0 =0 |{z} |{z} |{z}
=0 =0 =0 =0
+2µ∂(i vj) ∂(i vj) , (10.134)
 2
∂T ∂2T ∂u
ρcp = k 2 + µ .
| {z∂t} ∂y
| {z }
∂y
| {z }
energy increase energy diffusion viscous work source
(10.135)

© 05 December 2025, Joseph M. Powers. All rights reserved.

10.2. SIMILARITY SOLUTIONS 441

Let us first consider the x momentum equation. Recalling the momentum diffusivity
definition, Eq. (6.96), ν = µ/ρ, we get the following partial differential equation, initial and
boundary conditions:
∂u ∂2u
= ν 2, (10.136)
∂t ∂y
u(y, 0) = 0, u(0, t) = U, u(∞, t) = 0. (10.137)
Now let us scale the equations. Choose
u t y
u∗ = , t∗ = , y∗ = . (10.138)
U tc yc
We have yet to choose characteristic length, (yc ), and time, (tc ), scales. The equations
become
U ∂u∗ νU ∂ 2 u∗
= , (10.139)
tc ∂t∗ yc2 ∂y∗2
∂u∗ νtc ∂ 2 u∗
= . (10.140)
∂t∗ yc2 ∂y∗2
Wasting no time, we choose
ν µ
yc ≡ = . (10.141)
U ρU
3 kg m s m3 s
Examining the SI units, we see µ/(ρU) has units of length: Nm2s mkg s
m
= s2 m2 kg m
= m.
With this choice, we get
νtc νtc U 2 tc U 2
= = . (10.142)
yc2 ν2 ν
This suggests we choose
ν
tc = 2 . (10.143)
U
With all of these choices the complete system can be written as
∂u∗ ∂ 2 u∗
= , (10.144)
∂t∗ ∂y∗2
u∗ (y∗ , 0) = 0, u∗ (0, t∗ ) = 1, u∗ (∞, t∗ ) = 0. (10.145)
Now for self-similarity, we seek a transformation that reduces this partial differential equa-
tion, as well as its initial and boundary conditions, into an ordinary differential equation with
suitable boundary conditions. If this transformation does not exist, no similarity solution
exists. In this, but not all cases, the transformation does exist.
Let us first consider a general transformation from a y∗ , t∗ coordinate system to a new
η∗ , t̂∗ coordinate system. We assume then a general transformation
η∗ = η∗ (y∗ , t∗ ), (10.146)
t̂∗ = t̂∗ (y∗ , t∗ ). (10.147)

© 05 December 2025, Joseph M. Powers. All rights reserved.

442 CHAPTER 10. VISCOUS INCOMPRESSIBLE LAMINAR FLOW

Following the procedure of Ch. 2.5, this induces

   ∂η∗ ∂η∗   
dη∗ ∂y∗ ∂t∗ dy∗
= ∂ t̂∗ ∂ t̂∗ , (10.148)
dt̂∗ dt∗
| ∂y∗ {z ∂t∗ }
J

with Jacobian matrix

 ∂η∗ ∂η∗ 
∂y∗ ∂t∗
J= ∂ t̂∗ ∂ t̂∗ . (10.149)
∂y∗ ∂t∗

The chain rule for partial differentiation gives

 ∂   ∂η∗ ∂ t̂∗   ∂ 
∂η∗
∂y∗
∂ = ∂y ∗
∂η∗
∂y∗
∂ t̂∗ ∂ . (10.150)
∂t∗ ∂ t̂
| ∂t∗ {z ∂t∗ } ∗

JT

This expands as:

∂ ∂η∗ ∂ ∂ t̂∗ ∂
= + , (10.151)
∂y∗ t∗ ∂y∗ t∗ ∂η∗ t̂∗ ∂y∗ t∗ ∂ t̂∗ η∗

∂ ∂η∗ ∂ ∂ t̂∗ ∂
= + . (10.152)
∂t∗ y∗ ∂t∗ y∗ ∂η∗ t̂∗ ∂t∗ y∗ ∂ t̂∗ η∗

Now we will restrict ourselves to the transformation

t̂∗ = t∗ , (10.153)

so we have ∂ t̂∗ /∂t∗ |y∗ = 1 and ∂ t̂∗ /∂y∗ |t∗ = 0. Thus, our rules for differentiation reduce to
∂ ∂η∗ ∂
= , (10.154)
∂y∗ t∗ ∂y∗ t∗ ∂η∗ t̂∗
∂ ∂η∗ ∂ ∂
= + . (10.155)
∂t∗ y∗ ∂t∗ y∗ ∂η∗ t̂∗ ∂ t̂∗ η∗

The next assumption is key for a similarity solution to exist. We restrict ourselves to
transformations for which all state variables are at most a function of η∗ . That is we allow
no dependence on t̂∗ . Hence we must require that ∂/∂ t̂∗ |η∗ = 0. Moreover, partial derivatives
with respect to η∗ become total derivatives, giving us a final form of transformations for the
derivatives
∂ ∂η∗ d
= , (10.156)
∂y∗ t∗ ∂y∗ t∗ dη∗
∂ ∂η∗ d
= . (10.157)
∂t∗ y∗ ∂t∗ y∗ dη∗

© 05 December 2025, Joseph M. Powers. All rights reserved.

10.2. SIMILARITY SOLUTIONS 443

Now returning to Stokes’ first problem, let us assume that a similarity solution exists
of the form u∗ (y∗ , t∗ ) = u∗ (η∗ ). It is not always possible to find a similarity variable η∗ .
One of the more robust ways to find a similarity variable, if it exists, comes from group
theory,5 and is explained in detail in the monograph of Cantwell (2002). Group theory, that
is too detailed to explicate in full here, relies on a generalized symmetry of equations to find
simpler forms. In the same sense that a snowflake, subjected to rotations of π/3, 2π/3, π,
4π/3, 5π/3, or 2π, is transformed into a form that is indistinguishable from its original form,
we seek transformations of the variables in our partial differential equation that map the
equation into a form that is indistinguishable from the original. When systems are subject
to such transformations, known as group operators, they are said to exhibit symmetry.
Let us subject our governing partial differential equation along with initial and boundary
conditions to a particularly simple type of transformation, a simple stretching of space, time,
and velocity:
t̃ = ea t∗ , ỹ = eb y∗ , ũ = ec u∗. (10.158)
Here the “∼” variables are stretched variables, and a, b, and c are constant parameters. The
exponential will be seen to be a convenience, that is not absolutely necessary. Note that
for a ∈ (−∞, ∞), b ∈ (−∞, ∞), c ∈ (−∞, ∞), that ea ∈ (0, ∞), eb ∈ (0, ∞), ec ∈ (0, ∞).
So the stretching does not change the direction of the variable; that is it is not a reflecting
5
Group theory has a long history in mathematics and physics. Its complicated origins generally include
attribution to Évariste Galois, 1811-1832, a somewhat romantic figure, as well as Niels Henrick Abel, 1802-
1829, the Norwegian mathematician. Critical developments were formalized by Marius Sophus Lie, 1842-
1899, another Norwegian mathematician, in what today is known as Lie group theory. A modern variant,
known as “renormalization group” (RNG) theory is an area for active research. The 1982 Nobel prize in
physics went to Kenneth Geddes Wilson, 1936-2013, of Cornell University and The Ohio State University,
for use of RNG in studying phase transitions, first done in the 1970s. The award citation refers to the
possibilities of using RNG in studying the great unsolved problem of turbulence, a modern area of research
in which Steven Alan Orszag, 1943-2011, made many contributions.
Quoting from the useful Eric Weisstein’s World of Mathematics, available online at
[Link] “A group G is a finite or infinite set of elements to-
gether with a binary operation that together satisfy the four fundamental properties of closure, associativity,
the identity property, and the inverse property. The operation with respect to which a group is defined is
often called the ‘group operation,’ and a set is said to be a group ‘under’ this operation. Elements A, B,
C, . . . with binary operations A and B denoted AB form a group if
1. Closure: If A and B are two elements in G, then the product AB is also in G.
2. Associativity: The defined multiplication is associative, i.e. for all A, B, C ∈ G, (AB)C = A(BC).
3. Identity: There is an identity element I (a.k.a. 1, E, or e) such that IA = AI = A for every element
A ∈ G.
4. Inverse: There must be an inverse or reciprocal of each element. Therefore, the set must contain an
element B = A−1 such that AA−1 = A−1 A = I for each element of G.
. . ., A map between two groups that preserves the identity and the group operation is called a homomorphism.
If a homomorphism has an inverse that is also a homomorphism, then it is called an isomorphism and the two
groups are called isomorphic. Two groups that are isomorphic to each other are considered to be ‘the same’
when viewed as abstract groups.” For example, the group of 90 degree rotations of a square are isomorphic.

© 05 December 2025, Joseph M. Powers. All rights reserved.

444 CHAPTER 10. VISCOUS INCOMPRESSIBLE LAMINAR FLOW

transformation. We note that with this stretching, the domain of the problem remains
unchanged; that is t∗ ∈ [0, ∞) maps into t̃ ∈ [0, ∞); y∗ ∈ [0, ∞) maps into ỹ ∈ [0, ∞).
The range is also unchanged if we allow u∗ ∈ [0, ∞), that maps into ũ ∈ [0, ∞). Direct
substitution of the transformation shows that in the stretched space, the system becomes

∂ ũ ∂ 2 ũ
ea−c = e2b−c 2 , (10.159)
∂ t̃ ∂ ỹ
e−c ũ(ỹ, 0) = 0, −c
e ũ(0, t̃) = 1, e−c ũ(∞, t̃) = 0. (10.160)

In order that the stretching transformation map the system into a form indistinguishable
from the original, that is for the transformation to exhibit symmetry, we must take

c = 0, a = 2b. (10.161)

So our symmetry transformation is

t̃ = e2b t∗ , ỹ = eb y∗ , ũ = u∗ , (10.162)

giving in transformed space

∂ ũ ∂ 2 ũ
= , (10.163)
∂ t̃ ∂ ỹ 2
ũ(ỹ, 0) = 0, ũ(0, t̃) = 1, ũ(∞, t̃) = 0. (10.164)

Now both the original and transformed systems are the same, and the remaining stretching
parameter b does not enter directly into either formulation, so we cannot expect it in the
solution of either form. That is we expect a solution to be independent of the stretching
parameter b. This can be achieved if we take both u∗ and ũ to be functions of special
combinations of the independent variables, combinations that are formed such that b does
not appear. Eliminating b via
ỹ
eb = , (10.165)
y∗
we get  2
t̃ ỹ
= , (10.166)
t∗ y∗
or after rearrangement
y∗ ỹ
√ =√ . (10.167)
t∗ t̃
√
 √ 
We thus expect u∗ = u∗ y∗ / t∗ or equivalently ũ = ũ ỹ/ t̃ . This form also allows
√

u∗ = u∗ αy∗/ t∗ , where α is any constant. Let us then define our similarity variable η∗
as
y∗
η∗ = √ . (10.168)
2 t∗

© 05 December 2025, Joseph M. Powers. All rights reserved.

10.2. SIMILARITY SOLUTIONS 445

Here the factor of 1/2 is simply a convenience adopted so that the solution takes on a
traditional form. We would find that any constant in the similarity transformation would
induce a self-similar result.
Let us rewrite the differential equation, boundary, and initial conditions (∂u∗ /∂t∗ =
∂ u∗ /∂y∗2 , u∗(y∗ , 0) = 0, u∗ (0, t∗ ) = 1, u∗ (∞, t∗) = 0), in terms of the similarity variable η∗ .
2

We first must use the chain rule to get expressions for the derivatives. Applying the general
results just developed, we get
∂u∗ ∂η∗ du∗ 1 y∗ −3/2 du∗ η∗ du∗
= =− t∗ =− , (10.169)
∂t∗ ∂t∗ dη∗ 22 dη∗ 2t∗ dη∗
∂u∗ ∂η∗ du∗ 1 du∗
= = √ , (10.170)
∂y∗ ∂y∗ dη∗ 2 t∗ dη∗
   
∂ 2 u∗ ∂ ∂u∗ ∂ 1 du∗
= = √ , (10.171)
∂y∗2 ∂y∗ ∂y∗ ∂y∗ 2 t∗ dη∗
   
1 ∂ du∗ 1 1 d2 u ∗ 1 d2 u ∗
= √ = √ √ = . (10.172)
2 t∗ ∂y∗ dη∗ 2 t∗ 2 t∗ dη∗2 4t∗ dη∗2
Thus, applying these rules to our governing linear momenta equation, we recover
η∗ du∗ 1 d2 u ∗
− = , (10.173)
2t∗ dη∗ 4t∗ dη∗2
du∗ d2 u ∗
−2η∗ = , (10.174)
dη dη 2
| {z ∗} | {z∗}
acceleration viscous force imbalance
2
d u∗ du∗
2
+ 2η∗ = 0. (10.175)
dη∗ dη∗
Our governing equation has a singularity at t∗ = 0. As it appears on both sides of the
equation, we cancel it on both sides, but we shall see that this point is associated with special
behavior of the similarity solution. The important result is that the reduced equation has
dependency on η∗ only. If this did not occur, we could not have a similarity solution.
Now consider the initial and boundary conditions. They transform as follows:

y∗ = 0, =⇒ η∗ = 0, (10.176)
y∗ → ∞, =⇒ η∗ → ∞, (10.177)
t∗ → 0, =⇒ η∗ → ∞. (10.178)

The three important points for t∗ and y∗ collapse into two corresponding points in η∗ . This
is also necessary for the similarity solution to exist. Consequently, our conditions in η∗ space
reduce to

u∗ (0) = 1, no-slip, (10.179)

u∗ (∞) = 0, initial and far-field. (10.180)

© 05 December 2025, Joseph M. Powers. All rights reserved.

446 CHAPTER 10. VISCOUS INCOMPRESSIBLE LAMINAR FLOW

We solve the second order differential equation by the method of reduction of order, noticing
that it is really two first order equations in disguise:
   
d du∗ du∗
+ 2η∗ = 0. (10.181)
dη∗ dη∗ dη∗
2
multiplying by the integrating factor eη∗ , (10.182)
   
η∗2 d du∗ η∗2 du∗
e + 2η∗ e = 0. (10.183)
dη∗ dη∗ dη∗
 
d η∗2 du∗
e = 0, (10.184)
dη∗ dη∗
2 du∗
eη∗ = A, (10.185)
dη∗
du∗ 2
= Ae−η∗ , (10.186)
dη∗
Z η∗
2
u∗ = B + A e−s ds. (10.187)
0

Now applying the condition u∗ = 1 at η∗ = 0 gives

Z 0
2
1 = B+A e−s ds, (10.188)
| 0 {z }
=0
B = 1. (10.189)

So we have
Z η∗
2
u∗ = 1 + A e−s ds. (10.190)
0

Now applying the condition u∗ = 0 at η∗ → ∞, we get

Z ∞
2
0 = 1+A e−s ds, (10.191)
| 0 {z }
√
= π/2
√
π
= 1+A , (10.192)
2
2
A = −√ . (10.193)
π
Though not immediately obvious, it can be shown by a simple variable transformation
√ to
a polar coordinate system that the integral from 0 to ∞ has the value π/2. It is not
surprising that this integral has finite value over the semi-infinite domain as the integrand
is bounded between zero and one, and decays rapidly to zero as s → ∞.

© 05 December 2025, Joseph M. Powers. All rights reserved.

10.2. SIMILARITY SOLUTIONS 447

Let us divert to evaluate this integral. To do so, consider the related integral I2 defined
over the first quadrant in s − t space, where
Z ∞Z ∞
2 2
I2 ≡ e−s −t ds dt, (10.194)
0 0
Z ∞ Z ∞
−t2 2
= e e−s ds dt, (10.195)
0 0
Z ∞  Z ∞ 
−s2 −t2
= e ds e dt , (10.196)
0 0
Z ∞ 2
−s2
= e ds , (10.197)
0
p Z ∞
2
I2 = e−s ds. (10.198)
0

Now transform to polar coordinates with s = r cos θ, t = r sin θ. With this, we can easily
show ds dt = r dr dθ and s2 + t2 = r 2 . Substituting this into Eq. (10.194) and changing the
limits of integration appropriately, we get
Z π/2 Z ∞
2
I2 = e−r r dr dθ, (10.199)
0 0
Z π/2  ∞
1 −r2
= − e dθ, (10.200)
0 2 0
Z π/2  
1
= dθ, (10.201)
0 2
π
= . (10.202)
4
Comparing with Eq. (10.198), we deduce
p Z ∞ √
−s2 π
I2 = e ds = . (10.203)
0 2
With this verified, we can return to our original analysis and say that the velocity profile
can be written as
Z η∗
2 2
u∗ (η∗ ) = 1 − √ e−s ds, (10.204)
π 0
Z √y∗
2 2 t∗ 2
u∗ (y∗ , t∗ ) = 1 − √ e−s ds, (10.205)
π 0
 
y∗
= erfc √ . (10.206)
2 t∗
In the last form here, we have introduced the so-called error function complement, “erfc.”
Plots for the velocity profile in terms of both η∗ and y∗ , t∗ are given in Fig. 10.9. We see that

© 05 December 2025, Joseph M. Powers. All rights reserved.

448 CHAPTER 10. VISCOUS INCOMPRESSIBLE LAMINAR FLOW

t*
=3
1
t
*=
2
t =
* 1

1 1

Figure 10.9: Sketch of velocity field solution for Stokes’ first problem in both similarity
coordinate η∗ and primitive coordinates y∗ , t∗ .

in similarity space, the curve is a single curve that in which u∗ has a value of unity at η∗ = 0
and has nearly relaxed to zero when η∗ = 1. In dimensionless physical space, we see that at
early time, there is a thin momentum layer near the surface. At later time more momentum
is present in the fluid. We can say in fact that momentum is diffusing into the fluid.
We define the momentum diffusion length as the length for which significant momentum
has diffused into the fluid. This is well estimated by taking η∗ = 1. In terms of physical
variables, we have
y
√∗ = 1, (10.207)
2 t∗
√
y∗ = 2 t∗ , (10.208)
s
y t
ν = 2 ν , (10.209)
U U2
r
2ν U 2 t
y = , (10.210)
U√ ν
= 2 νt. (10.211)

We can in fact define this as a boundary layer thickness. That is to say the momentum
boundary layer thickness in Stokes’ first problem grows at a rate proportional to the square
root of momentum diffusivity and time. This class of result is a hallmark of all diffusion
processes, be it mass, momentum, or energy. Taking standard properties of air at 300 K
and
p one atmosphere pressure, we find after one minute that its boundary layer thickness is
2 (15.89 × 10p−6 (m2 /s)(60 s) = 0.062 m. For motor oil at the same conditions, we get a
thickness of 2 (550 × 10−6 (m2 /s)(60 s) = 0.363 m.
We next consider the shear stress field. For this problem, the shear stress reduces to

© 05 December 2025, Joseph M. Powers. All rights reserved.

10.2. SIMILARITY SOLUTIONS 449

simply
∂u
τ =µ . (10.212)
∂y
Scaling as before by a characteristic stress τc , we get
µU ∂u∗
τ∗ τc = ν , (10.213)
U
∂y ∗
µU 2 1 ∂u∗
τ∗ = . (10.214)
ν τc ∂y∗
Taking τc = µU 2 /ν = µU 2 /(µ/ρ) = ρU 2 , we get
∂u∗ 1 du∗
τ∗ = = √ , (10.215)
∂y∗ 2 t∗ dη∗
 
1 2 −η∗2
= √ −√ e , (10.216)
2 t∗ π
1 2
= − √ e−η∗ , (10.217)
πt∗
 2
1 y∗
= −√ exp − √ . (10.218)
πt∗ 2 t∗
Now at the wall, y∗ = 0, and we get
1
τ∗ |y∗ =0 = − √ . (10.219)
πt∗
So the shear stress does not have a similarity solution, but is directly related to time variation.
The equation holds that the stress is infinite at t∗ = 0, and decreases as time increases. This
is because the velocity gradient flattens as time progresses. It can also be shown that while
the stress is unbounded at a single point in time, that the impulse over a finite time span
is finite, even when the time span includes t∗ = 0. It can also be shown that the flow
corresponds to a pulse of vorticity being introduced at the wall, that subsequently diffuses
into the fluid.
In dimensional terms, we can say
τ 1
2
= −q , (10.220)
ρU 2
πU t ν
2
ρU
τ = −√ √ , (10.221)
π √Uν t
q
ρU µρ
= − √ , (10.222)
πt
√
U ρµ
= − √ . (10.223)
πt

© 05 December 2025, Joseph M. Powers. All rights reserved.

450 CHAPTER 10. VISCOUS INCOMPRESSIBLE LAMINAR FLOW

For this two-dimensional flow, the dimensionless vorticity vector is confined to the z∗ direction
and has magnitude
 2
∂v∗ ∂u∗ 1 y∗
ωz∗ = − =√ exp − √ . (10.224)
∂x∗ ∂y∗ πt∗ 2 t∗
|{z}
=0

The dimensionless acceleration vector is confined to the x∗ direction and has magnitude
 2
∂u∗ 1 y∗
ax∗ = = √ 3/2 exp − √ . (10.225)
∂t∗ 2 πt∗ 2 t∗
The dimensionless deformation tensor reduces to
  2 
 1 ∂u∗  0 y∗
− 2√1πt∗ exp − 2√
0 t∗
D∗ = 1 ∂u∗ 2 ∂y∗ =   2  . (10.226)
2 ∂y∗
0 − √1 exp − √ y∗
0
2 πt∗ 2 t∗

The deformation tensor has eigenvalues that represent the extreme values of extensional
strain, and they are
 2
1 y∗
D∗ = ± √ exp − √ . (10.227)
2 πt∗ 2 t∗
It is straightforward to show that the eigenvectors of D∗ are aligned with coordinates axes
that have been rotated through an angle of π/4.
Now let us consider the heat transfer problem. Recall the governing equation, initial and
boundary conditions are
 2
∂T ∂2T ∂u
ρcp = k 2 +µ , (10.228)
∂t ∂y ∂y
T (y, 0) = To , T (0, t) = To , T (∞, t) = To . (10.229)
We will adopt the same time tc and length yc scales as before. Take the dimensionless
temperature to be
T − To
T∗ = . (10.230)
To
So we get
 2
ρcp To ∂T∗ kTo ∂ 2 T∗ µU 2 ∂u∗
= + 2 , (10.231)
tc ∂t∗ yc2 ∂y∗2 yc ∂y∗
 2
∂T∗ kTo tc ∂ 2 T∗ µU 2 tc ∂u∗
= + 2 , (10.232)
∂t∗ yc2 ρcp To ∂y∗2 yc ρcp To ∂y∗
k tc kU 2 ν 1 k k 1
now 2 = 2 2
= = = , (10.233)
yc ρcp ν U ρcp ρcp ν µcp Pr
µT 2 tc µU 2 U 2 ν 1 µU 2 U2
= = µ = = Ec. (10.234)
yc2 ρcp To ν 2 U 2 ρcp To ρ
ρc p To c p T o

© 05 December 2025, Joseph M. Powers. All rights reserved.

10.2. SIMILARITY SOLUTIONS 451

So we have in dimensionless form

 2
∂T∗ 1 ∂ 2 T∗ ∂u∗
= + Ec , (10.235)
∂t∗ P r ∂y∗2 ∂y∗
T∗ (y∗ , 0) = 0, T∗ (0, t∗ ) = 0, T∗ (∞, t∗ ) = 0. (10.236)

Notice that the only driving inhomogeneity is the viscous work. Now we know from our
solution of the linear momentum equation that
 
∂u∗ 1 y∗2
= −√ exp − . (10.237)
∂y∗ πt∗ 4t∗
So we can rewrite the equation for temperature variation as
 
∂T∗ 1 ∂ 2 T∗ Ec y∗2
= + exp − , (10.238)
∂t∗ P r ∂y∗2 πt∗ 2t∗
T∗ (y∗ , 0) = 0, T∗ (0, t∗ ) = 0, T∗ (∞, t∗ ) = 0. (10.239)

Before considering the general solution, let us consider some limiting cases.
• Ec → 0
In the limit as Ec → 0, we get a trivial solution, T∗ (y∗ , t∗ ) = 0.

• Pr → ∞
Recalling that the Prandtl number is the ratio of momentum diffusivity to thermal
diffusivity, Eq. (6.166), this limit corresponds to materials for which momentum diffu-
sivity is much greater than thermal diffusivity. For example for SAE 30 oil, the Prandtl
number is around 3500. Naı̈vely assuming that we can simply neglect conduction, we
write the energy equation in this limit as
 
∂T∗ Ec y∗2
= exp − . (10.240)
∂t∗ πt∗ 2t∗
√
and with T∗ = T∗ (η∗ ) and η∗ = y∗ /(2 t∗ ), we get the transformed partial time deriva-
tive to be
∂T∗ η∗ dT∗
=− . (10.241)
∂t∗ 2t∗ dη∗
So the governing equation reduces to
η∗ dT∗ Ec −2η∗2
− = e , (10.242)
2t∗ dη∗ πt∗
dT∗ 2Ec 1 −2η∗2
= − e , (10.243)
dη∗ π η∗
Z
2Ec ∞ 1 −2s2
T∗ = e ds. (10.244)
π η∗ s

© 05 December 2025, Joseph M. Powers. All rights reserved.

452 CHAPTER 10. VISCOUS INCOMPRESSIBLE LAMINAR FLOW

We cannot satisfy both boundary conditions; the equation has been solved so as to
satisfy the boundary condition in the far field of T∗ (∞) = 0.
Unfortunately, we notice that we cannot satisfy the boundary condition at η∗ = 0.
We simply do not have enough degrees of freedom. In actuality, what we have found
is an outer solution, and to match the boundary condition at 0, we would have to
reintroduce conduction, that has a higher derivative.
First let us see how the outer solution behaves near η∗ = 0. Expanding the differential
equation in a Taylor series about η∗ = 0 and solving gives
 
dT∗ 2Ec 1 3
=− − 2η∗ + 2η∗ + . . . , (10.245)
dη∗ π η∗
 
2Ec 2 1 4
T∗ = − ln η∗ − η∗ + η∗ + . . . . (10.246)
π 2
It turns out that solving the inner layer problem and the matching is of about the
same difficulty as solving the full general problem, so we will defer this until later in
this section.
• Pr → 0
In this limit, we get
∂ 2 T∗
= 0. (10.247)
∂y∗2
The solution that satisfies the boundary conditions is
T∗ = 0. (10.248)
In this limit, momentum diffuses slowly relative to energy. So we can interpret the
results as follows. In the boundary layer, momentum is generated in a thin layer.
Viscous dissipation in this layer gives rise to a local change in temperature in the layer
that rapidly diffuses throughout the entire flow. The effect of smearing a localized
finite thermal energy input over a semi-infinite domain has a negligible influence on
the temperature of the global domain.
So let us bring back diffusion and study solutions for finite Prandtl number. Our gov-
erning equation in similarity variables then becomes
η∗ dT∗ 1 1 d2 T∗ Ec −2η∗2
− = 2
+ e , (10.249)
2t∗ dη∗ P r 4t∗ dη∗ πt∗
dT∗ 1 d2 T∗ 4Ec −2η∗2
−2η∗ = 2
+ e , (10.250)
dη∗ P r dη∗ | π {z }
| {z } | {z }
temperature evolution energy diffusion dissipation
2
d T∗ dT∗ 4 −2η∗2
+ 2P r η∗ = − EcP r e , (10.251)
dη∗2 dη∗ π
T∗ (0) = 0, T∗ (∞) = 0. (10.252)

© 05 December 2025, Joseph M. Powers. All rights reserved.

10.2. SIMILARITY SOLUTIONS 453

The second order differential equation is really two first order differential equations in dis-
2
guise. There is an integrating factor of eP r η∗ . Multiplying by the integrating factor and
operating on the system, we find
2
η∗2 d
T∗ 2 dT∗ 4 2
eP r 2
+ 2P r η∗ eP r η∗ = − EcP r e(P r−2)η∗ , (10.253)
dη∗ dη∗ π
 
d 2 dT∗ 4 2
eP r η∗ = − EcP r e(P r−2)η∗ , (10.254)
dη∗ dη∗ π
Z η∗
P r η∗2 dT∗ 4 2
e = − EcP r e(P r−2)s ds + C1 . (10.255)
dη∗ π 0

Continuing, we get
Z η∗
dT∗ 4 −P r η∗2 2 2
= − EcP r e e(P r−2)s ds + C1 e−P rη∗ , (10.256)
dη∗ π 0
Z η∗ Z p Z η∗
4 −P r p2 (P r−2)s2 s2
T∗ = − EcP r e e ds dp + C1 e−P r ds + C2 .(10.257)
π 0 0 0

The boundary condition T∗ (0) = 0 gives us C2 = 0. The boundary condition at ∞ gives us

Z ∞ Z p Z ∞
4 −P r p2 (P r−2)s2 2
0 = − EcP r e e ds dp + C1 e−P r s ds . (10.258)
π 0 0 0
| {z
p }
1 π
2 Pr
(10.259)
Therefore, we get
Z ∞ Z p
r
4 p2 (P r−2)s2 C1 π
EcP r e−P r e ds dp = , (10.260)
π 0 0 2 Pr
Z ∞ Z p
8 3/2 p2 2
C1 = EcP r e−P r e(P r−2)s ds dp. (10.261)
π 3/2 0 0

So finally, we have for the temperature profile

Z η∗ Z p
4 −P r p2 2
T∗ (η∗ ) = − EcP r e e(P r−2)s ds dp
π 0
 Z ∞ 0 Z p  Z η∗
8 3/2 −P r p2 (P r−2)s2 s2
+ EcP r e e ds dp e−P r ds.(10.262)
π 3/2 0 0 0

This simplifies somewhat to

Z η √ 
2EcP r p2
T∗ (η∗ ) = − p e−P r erf 2 − Pr p
dp
π(2 − P r) 0
√ Z ∞ √  
p2
− erf Pr η e−P r erf 2 − P r p dp . (10.263)
0

© 05 December 2025, Joseph M. Powers. All rights reserved.

454 CHAPTER 10. VISCOUS INCOMPRESSIBLE LAMINAR FLOW

This analysis simplifies considerably in the limit of P r = 1, that is when momentum and
energy diffuse at the same rate. This is a close to reality for many gases. In this case, the
temperature profile becomes
Z η∗ Z p Z η∗
4 −p2 −s2 2
T∗ (η∗ ) = − Ec e e ds dp + C1 e−s ds. (10.264)
π 0 0 0

Rp 2 2
Now if h(p) = 0
e−s ds, we get dh/dp = e−p . Using this, we can rewrite the temperature
profile as
Z η∗ Z η∗
4 dh 2
T∗ (η∗ ) = − Ec h(p) dp + C1 e−s ds, (10.265)
π dp
Z 0η∗  2  Z η∗ 0
4Ec h 2
= − d + C1 e−s ds, (10.266)
π 0 2 0
  Z η∗ 2 Z η∗
4Ec 1 −s2 2
= − e ds + C1 e−s ds, (10.267)
π 2
 Z η∗ 0  Z η∗ 0
2Ec 2 2
= − e−s ds + C1 e−s ds. (10.268)
π 0 0

Now for T (∞) = 0, we get

 Z Z ∞
2Ec ∞ −s2 2
0 = − e ds + C1 e−s ds, (10.269)
π 0 0
 √ √
2Ec π π
= − + C1 , (10.270)
π 2 2
Ec
C1 = √ . (10.271)
π

So the temperature profile can be expressed as

Z η∗  Z η∗ 
Ec −s2 2 −s2
T∗ (η∗ ) = √ e ds 1− √ e ds . (10.272)
π 0 π 0

We notice that we can write this directly in terms of the velocity as

Ec
T∗ (η∗ ) = u∗ (η∗ ) (1 − u∗ (η∗ )) . (10.273)
2

This is a consequence of what is known as Reynolds’ analogy that holds for P r = 1 that
the temperature field can be directly related to the velocity field. The temperature field for
Stokes’ first problem for P r = 1, Ec = 1 is plotted in Fig. 10.10.

© 05 December 2025, Joseph M. Powers. All rights reserved.

10.2. SIMILARITY SOLUTIONS 455

Pr = 1
Ec = 1

0 0.12

Figure 10.10: Plot of temperature field for Stokes’ first problem for P r = 1, Ec = 1.

10.2.2 Blasius boundary layer

We next consider the well known problem of the flow of a viscous fluid over a flat plate. This
problem forms the foundation for a variety of viscous flows over more complicated geometries.
It also illustrates some important features of viscous flow physics, as well as giving the original
motivating problem for the mathematical technique of matched asymptotic expansions. Here
we will consider, as sketched in Fig. 10.11, the incompressible flow of viscous fluid of constant
viscosity and thermal conductivity over a flat plate. In the far field, the fluid will be a uniform
stream with constant velocity. At the plate surface, the no-slip condition must be enforced,
that will give rise to a zone of adjustment where the fluid’s velocity changes from zero at the
plate surface to its freestream value. This zone is called the boundary layer. The sketch has
a small inaccuracy as it only gives the u velocity component variation. Actually there is also
a v component of the velocity vector, and that induces streamline curvature, not apparent
in Fig. 10.11.
Considering first the velocity field, we find, assuming the flow is steady as well, that the
dimensionless two-dimensional Navier-Stokes equations are as follows

∂u ∂v
+ = 0, (10.274)
∂x ∂y
 
∂u ∂u ∂p 1 ∂2u ∂2u
u +v = − + + , (10.275)
∂x ∂y ∂x Re ∂x2 ∂y 2
 
∂v ∂v ∂p 1 ∂2v ∂2v
u +v = − + + . (10.276)
∂x ∂y ∂y Re ∂x2 ∂y 2

© 05 December 2025, Joseph M. Powers. All rights reserved.

456 CHAPTER 10. VISCOUS INCOMPRESSIBLE LAMINAR FLOW

Figure 10.11: Schematic for flat plate boundary layer problem.

The dimensionless boundary conditions are

u(x, y → ∞) = 1, (10.277)
p(x, y → ∞) = 0, (10.278)
u(x, 0) = 0, (10.279)
v(x, 0) = 0. (10.280)

In this section, we are dispensing with the ∗’s and assuming all variables are dimensionless.
In fact we have assumed a scaling of the following form, where dim is a subscript denoting
a dimensional variable.
udim vdim xdim ydim pdim − po
u= , v= , x= , y= , p= . (10.281)
U U L L ρU 2

For our flat plate of semi-infinite extent, we do not have a natural length scale. This suggests
that we may find a similarity solution that removes the effect of L. It also may be appropriate
to think of making the requirement that
ν
L≫ . (10.282)
U
This will insure that Re = UL/ν ≫ 1. This is not entirely satisfying as we really want our
domain to be semi-infinite with L → ∞.
Now let us consider that for Re → ∞, we have an outer solution of u = 1 to be valid
for most of the flow field sufficiently far away from the plate surface. In fact the solution
u = 1, v = 0, p = 0, satisfies all of the governing equations and boundary conditions except
for the no-slip condition at y = 0. Because in the limit as Re → ∞, we effectively ignore the

© 05 December 2025, Joseph M. Powers. All rights reserved.

10.2. SIMILARITY SOLUTIONS 457

high order derivatives found in the viscous terms, we cannot expect to satisfy all boundary
conditions for the full problem. We call this the outer solution, that is also an inviscid
solution to the equations, allowing for a slip condition at the boundary.
Let us rescale our equations near the plate surface y = 0 to
• bring back the effect of the viscous terms,
• bring back the no-slip condition, and
• match our inviscid outer solution to a viscous inner solution.
This is the first example of the use of the method of matched asymptotic expansions as
introduced by Prandtl and his student Blasius in the early twentieth century.
With some difficulty, we could show how to choose the scaling, let us simply adopt a
scaling and show that it indeed achieves our desired end. So let us take a scaled y distance
and velocity, denoted by a˜superscript, to be
√ √
ṽ = Re v, ỹ = Re y. (10.283)
With this scaling, assuming the Reynolds number is large, when we examine small y or
v, we are examining an order unity ỹ or ṽ. Our equations rescale as
√
∂u 1/ Re ∂ṽ
+ √ = 0, (10.284)
∂x 1/ Re ∂ ỹ
√  
∂u 1/ Re ∂u ∂p 1 ∂2u ∂2u
u + √ ṽ = − + + Re 2 , (10.285)
∂x 1/ Re ∂ ỹ ∂x Re ∂x2 ∂ ỹ
√ √
1 ∂ṽ (1/ Re)(1/ Re) ∂ṽ 1 ∂p
√ u + √ ṽ = − √
Re ∂x 1/ Re ∂ ỹ 1/ Re ∂ ỹ
√ !
1 1 ∂ 2 ṽ 1/ Re ∂ 2 ṽ
+ √ + . (10.286)
Re Re ∂x2 1/Re ∂ ỹ 2

Simplifying, this reduces to

∂u ∂ṽ
+ = 0, (10.287)
∂x ∂ ỹ
∂u ∂u ∂p 1 ∂2u ∂2u
u + ṽ = − + + , (10.288)
∂x ∂ ỹ ∂x Re ∂x2 ∂ ỹ 2
∂ṽ ∂ṽ ∂p 1 ∂ 2 ṽ ∂ 2 ṽ
u + ṽ = −Re + + . (10.289)
∂x ∂ ỹ ∂ ỹ Re ∂x2 ∂ ỹ 2
Now in the limit as Re → ∞, the rescaled equations reduce to the well known boundary
layer equations:
∂u ∂ṽ
+ = 0, (10.290)
∂x ∂ ỹ

© 05 December 2025, Joseph M. Powers. All rights reserved.

458 CHAPTER 10. VISCOUS INCOMPRESSIBLE LAMINAR FLOW

∂u ∂u ∂p ∂ 2 u
u + ṽ = − + , (10.291)
∂x ∂ ỹ ∂x ∂ ỹ 2
∂p
0 = . (10.292)
∂ ỹ

To match the outer solution, we need the boundary conditions that are

u(x, ỹ → ∞) = 1, (10.293)
p(x, ỹ → ∞) = 0, (10.294)
u(x, 0) = 0, (10.295)
ṽ(x, 0) = 0. (10.296)

The ỹ momentum equation gives us

p = p(x). (10.297)
In general, we can consider this to be an imposed pressure gradient that is supplied by the
outer inviscid solution. For general flows, that pressure gradient dp/dx will be non-zero.
For the Blasius problem, we will choose to study problems for which there is no pressure
gradient. That is we take

p(x) = 0, for Blasius flat plate boundary layer. (10.298)

So called Falkner-Skan solutions consider flows over curved plates, for which the outer inviscid
solution does not have a constant pressure. This ultimately affects the behavior of the fluid in
the boundary layer, giving results that differ in important features from our Blasius problem.
With our assumptions, the Blasius problem reduces to

∂u ∂ṽ
+ = 0, (10.299)
∂x ∂ ỹ
∂u ∂u ∂2u
u + ṽ = . (10.300)
∂x ∂ ỹ ∂ ỹ 2

The boundary conditions are now only on velocity and are

u(x, ỹ → ∞) = 1, (10.301)
u(x, 0) = 0, (10.302)
ṽ(x, 0) = 0. (10.303)

Now to simplify, we invoke the stream function ψ, that allows us to satisfy continuity au-
tomatically and eliminate u and ṽ at the expense of raising the order of the differential
equation. So taking
∂ψ ∂ψ
u= , ṽ = − , (10.304)
∂ ỹ ∂x

© 05 December 2025, Joseph M. Powers. All rights reserved.

10.2. SIMILARITY SOLUTIONS 459

we find that mass conservation reduces to ∂ 2 ψ/∂x∂ ỹ − ∂ 2 ψ/∂ ỹ∂x = 0. The x momentum
equation and associated boundary conditions become
∂ψ ∂ 2 ψ ∂ψ ∂ 2 ψ ∂3ψ
− = , (10.305)
∂ ỹ ∂x∂ ỹ ∂x ∂ ỹ 2 ∂ ỹ 3
∂ψ
(x, ỹ → ∞) = 1, (10.306)
∂ ỹ
∂ψ
(x, 0) = 0, (10.307)
∂ ỹ
∂ψ
(x, 0) = 0. (10.308)
∂x
Let us try stretching all the variables of this system to see if there are stretching transforma-
tions under which the system exhibits symmetry; that is we seek a stretching transformation
under which the system is invariant. Take

x̂ = ea x, ŷ = eb ỹ, ψ̂ = ec ψ. (10.309)
Under this transformation, the x momentum equation and boundary conditions transform
to
∂ ψ̂ ∂ 2 ψ̂ ∂ ψ̂ ∂ 2 ψ̂ ∂ 3 ψ̂
ea+2b−2c − ea+2b−2c = e3b−c , (10.310)
∂ ŷ ∂ x̂∂ ŷ ∂ x̂ ∂ ŷ 2 ∂ ŷ 3
∂ ψ̂
eb−c (x̂, ŷ → ∞) = 1, (10.311)
∂ ŷ
∂ ψ̂
eb−c (x̂, 0) = 0, (10.312)
∂ ŷ
∂ ψ̂
ea−c (x̂, 0) = 0. (10.313)
∂ x̂
If we demand b = c and a = 2c, then the transformation is invariant, yielding

∂ ψ̂ ∂ 2 ψ̂ ∂ ψ̂ ∂ 2 ψ̂ ∂ 3 ψ̂
− = , (10.314)
∂ ŷ ∂ x̂∂ ŷ ∂ x̂ ∂ ŷ 2 ∂ ŷ 3
∂ ψ̂
(x̂, ŷ → ∞) = 1, (10.315)
∂ ŷ
∂ ψ̂
(x̂, 0) = 0, (10.316)
∂ ŷ
∂ ψ̂
(x̂, 0) = 0. (10.317)
∂ x̂
Now our transformation is reduced to

x̂ = e2c x, ŷ = ec ỹ, ψ̂ = ec ψ. (10.318)

© 05 December 2025, Joseph M. Powers. All rights reserved.

460 CHAPTER 10. VISCOUS INCOMPRESSIBLE LAMINAR FLOW

Because c does not appear explicitly in either the original equation set nor the transformed
equation set, the solution
p must not depend on this stretching. Eliminating c from the
c
transformation by e = x̂/x we find that
r r
ŷ x̂ ψ̂ x̂
= , = , (10.319)
ỹ x ψ x
or
ŷ ỹ ψ̂ ψ
√ =√ , √ =√ . (10.320)
x̂ x x̂ x
Thus motivated, let us seek solutions of the form
 
ψ ỹ
√ =f √ . (10.321)
x x
That is taking
ỹ
η=√ , (10.322)
x
we seek √
ψ= xf (η). (10.323)
Let us check that our similarity variable is independent of L our unknown length scale.
√ √ r √ r
ỹ Re y Re ydim /L UL ydim L U ydim
η=√ = √ = p = √ = √ . (10.324)
x x xdim /L ν L xdim ν xdim
So indeed, our similarity variable is independent of any arbitrary length scale we happen to
have chosen.
With our similarity transformation, we have
∂η 1 1η
= − ỹx−3/2 = − , (10.325)
∂x 2 2x
∂η 1
= √ . (10.326)
∂ ỹ x
Now we need expressions for ∂ψ/∂x, ∂ψ/∂ ỹ, ∂ 2 ψ/∂x∂ ỹ, ∂ 2 ψ/∂ ỹ 2 , and ∂ 3 ψ/∂ ỹ 3 . First, con-
sider the partial derivatives of the stream function ψ. Operating on each partial derivative,
we find
∂ψ ∂ √

= xf (η) , (10.327)
∂x ∂x
√ df ∂η 1 1
= x + √ f, (10.328)
dη ∂x 2 x
 
√ 1 η df 1 1
= x − + √ f, (10.329)
2 x dη 2 x
 
1 df
= √ f −η . (10.330)
2 x dη

© 05 December 2025, Joseph M. Powers. All rights reserved.

10.2. SIMILARITY SOLUTIONS 461

So we get for ṽ that

 
∂ψ 1 df
ṽ = − = √ η −f . (10.331)
∂x 2 x dη
And then we find
∂ψ ∂ √

= xf (η) , (10.332)
∂ ỹ ∂ ỹ
√ ∂
= x (f (η)), (10.333)
∂ ỹ
√ df ∂η
= x , (10.334)
dη ∂ ỹ
√ df 1
= x √ , (10.335)
dη x
df
= . (10.336)
dη
Thus for u, we get
∂ψ df
u= = . (10.337)
∂ ỹ dη
So
   
∂2ψ ∂ ∂ψ ∂ df d2 f ∂η 1 d2 f
= = = 2 = − η 2. (10.338)
∂x∂ ỹ ∂x ∂ ỹ ∂x dη dη ∂x 2x dη
2
    2
∂ ψ ∂ ∂ψ ∂ df d f ∂η 1 d2 f
= = = = √ . (10.339)
∂ ỹ 2 ∂ ỹ ∂ ỹ ∂ ỹ dη dη 2 ∂ ỹ x dη 2
     
∂3ψ ∂ ∂2ψ ∂ 1 d2 f 1 ∂ d2 f 1 d3 f ∂η
= = √ = √ = √ , (10.340)
∂ ỹ 3 ∂ ỹ ∂ ỹ 2 ∂ ỹ x dη 2 x ∂ ỹ dη 2 x dη 3 ∂ ỹ
1 d3 f
= . (10.341)
x dη 3
Now we substitute each of these expressions into the x momentum equation and get
   
df 1 d2 f 1 df 1 d2 f 1 d3 f
− η 2 + √ η −f √ = , (10.342)
dη 2x dη 2 x dη x dη 2 x dη 3
|{z} | {z } | {z } | {z } | {z }
u ∂u ṽ ∂u ∂2 u
∂x ∂ ỹ ∂ ỹ 2
  2
df d2 f df df d3 f
−η + η − f = 2 , (10.343)
dη dη 2 dη dη 2 dη 3
d2 f d3 f
−f 2 = 2 3 , (10.344)
dη dη
| {z } | {z }
advection momentum diffusion
d3 f 1 d2 f
+ f = 0. (10.345)
dη 3 2 dη 2

© 05 December 2025, Joseph M. Powers. All rights reserved.

462 CHAPTER 10. VISCOUS INCOMPRESSIBLE LAMINAR FLOW

This is a third order non-linear ordinary differential equation for f (η). We need three
boundary conditions. Now at the surface ỹ = 0, we have η = 0. And as ỹ → ∞, we have
η → ∞. To satisfy the no-slip condition on u at the plate surface, we require
df
= 0. (10.346)
dη η=0

For no-slip on ṽ, we require

 
1df
ṽ(0) = 0 = √
η −f , (10.347)
2 x dη
df
0 = 0 −f (0), (10.348)
dη η=0
| {z }
=0
f (0) = 0. (10.349)
And to satisfy the freestream condition on u as η → ∞, we need
df
= 1. (10.350)
dη η→∞

The most standard way to solve non-linear ordinary differential equations of this type is to
reduce them to systems of first order ordinary differential equations and use some numerical
technique, such as a Runge6 -Kutta integration. We recall that Runge-Kutta techniques, as
well as most other common techniques, require a well-defined set of initial conditions to
predict the final state. To achieve the desired form, we define
df d2 f
g≡ , h≡ . (10.351)
dη dη 2
Thus the x momentum equation becomes
dh 1
+ f h = 0. (10.352)
dη 2
But this is one equation in three unknowns. We need to write our equations as a system of
three first order equations, along with associated initial conditions. They are
df
= g, f (0) = 0, (10.353)
dη
dg
= h, g(0) = 0, (10.354)
dη
dh 1
= − f h, h(0) =?. (10.355)
dη 2
6
Carl David Tolmè Runge, 1856-1927, German mathematician and physicist, close friend of Max Planck,
studied spectral line elements of non-Hydrogen molecules, held chairs at Hanover and Göttingen, entertained
grandchildren at age 70 by doing handstands.

© 05 December 2025, Joseph M. Powers. All rights reserved.

10.2. SIMILARITY SOLUTIONS 463

Everything is well-defined except we do not have an initial condition on h. We do however

have a far-field condition on g that is g(∞) = 1. One viable option we have for getting a
final solution is to use a numerical trial and error procedure, guessing h(0) until we find that
g(∞) → 1. We will use a slightly more efficient method here, that only requires one guess.
To do this, let us first demonstrate the following lemma: If f = F (η) is a solution
to the Blasius equation d3 f /dη 3 + 21 f (d2 f /dη 2) = 0, then f = aF (aη) is also a solution.
The proof is as follows. Consider the transformation f = aF and ξ = aη. Then we have
d/dη = (dξ/dη)d/dξ = a d/dξ. Then the Blasius equation transforms to

d3 F 1 4 d2 F d3 F 1 d2 F
a4 + a F 2 = 0, + F = 0. (10.356)
dξ 3 2 dξ dξ 3 2 dξ 2
This is again the Blasius equation, and F (ξ) is its solution. Hence f = aF (aη) is a solution.
This is true for any boundary conditions.
So to solve our non-linear system, let us first solve the following related system:
dF
= G, F (0) = 0, (10.357)
dη
dG
= H, G(0) = 0, (10.358)
dη
dH 1
= − F H, H(0) = 1. (10.359)
dη 2
After one numerical integration, we find that with this guess for H(0) that

G(∞) = 2.08540918... (10.360)

Now our numerical solution also gives us F , and so we know that f = aF (aη) is also a
solution. Moreover
df dF (aη)
= a2 , that is (10.361)
dη dη
g(η) = a2 G(aη). (10.362)

Now we want g(∞) = 1, so take 1 = a2 G(∞), so a2 = 1/G(∞). So

1
a= p . (10.363)
G(∞)

Now
d2 f 3d
2
F (aη)
= a , (10.364)
dη 2 dη 2
d2 f 2
3 d F (aη)
= a , (10.365)
dη 2 η=0 dη 2 η=0

© 05 December 2025, Joseph M. Powers. All rights reserved.

464 CHAPTER 10. VISCOUS INCOMPRESSIBLE LAMINAR FLOW

0 1 u

Figure 10.12: Velocity component profile for u for a Blasius boundary layer.

h(0) = a3 H(0), (10.366)

= a3 (1), (10.367)
= a3 = G−3/2 (∞), (10.368)
= (2.08540918...)−3/2, (10.369)
= 0.332057335... (10.370)

This is the proper choice for the initial condition on h. Numerically integrating once
more, we get the behavior of f , g, and h as functions of η that indeed satisfies the condition
at ∞. A plot of u = df /dη as a function of η is shown in Fig. 10.12. From Fig. 10.12, we see
that when η = 5, the velocity has nearly acquired the freestream value of u = 1. We can plot
streamlines and the velocity vector field as well using the transformations to acquire ψ, u
and ṽ as functions of x and ỹ. They are plotted in in Fig. 10.13. Notice that the streamlines
have curvature, consistent with the velocity vector having non-zero components in the x and
ỹ directions.
The associated kinematic topics of acceleration and vorticity vector fields, deformation
tensors, are not straightforward and need to be carefully interpreted in light of the fact that
we have scaled our equations in a particular fashion. Examination of the streamlines and
velocity vector field of Fig. 10.13 suggests that the flow is decelerating in the streamwise
direction, and that streamline curvature induces a stream-normal component of centripetal
acceleration. The only forces available to induce such an acceleration are imbalanced viscous
shear forces. Certainly one can visualize that a fluid element is both rotating as well as
deforming in a volume-preserving fashion.
Panton, Ch. 20, shows that to leading order, the dimensionless vorticity field is given by

d2 f
ω=− = −h(η). (10.371)
dη 2

© 05 December 2025, Joseph M. Powers. All rights reserved.

10.2. SIMILARITY SOLUTIONS 465

1.0

0.8

0.6

0.4

0.2

0.0
0.0 0.5 1.0 1.5 2.0
x

Figure 10.13: Streamlines and velocity vector field for Blasius boundary layer.

A plot is given in in Fig. 10.14. We note the vorticity is maximum at the no-slip boundary
at η = 0. As η → ∞, the flow becomes irrotational, consistent with the uniform freestream
in the far field.
Examination of the numerical results shows that when η = 4.9, that the u component of
velocity has 0.99 of its freestream value. As the velocity only reaches its freestream value at
∞, we define the boundary layer thickness, δ0.99 , as that value of ydim for which the velocity
has 0.99 of its freestream value. Recalling that
r
U ydim
η= √ , (10.372)
ν xdim

we say that r
U δ0.99
4.9 = √ . (10.373)
ν xdim
Rearranging, we get
r
δ0.99 ν
= 4.9 , (10.374)
xdim Uxdim
= 4.9Re−1/2
xdim . (10.375)

Here we have taken a Reynolds number based on local distance to be

Uxdim
Rexdim = . (10.376)
ν
This formula is valid for laminar flows, and has been seen to be valid for Rexdim < 3 × 106 .
For greater lengths, there can be a transition to turbulent flow. For water flowing a 1 m/s
and a downstream distance of 1 m, we find δ0.99 = 0.5 cm. For air under the same conditions,

© 05 December 2025, Joseph M. Powers. All rights reserved.

466 CHAPTER 10. VISCOUS INCOMPRESSIBLE LAMINAR FLOW

10

- 0.35 - 0.30 - 0.25 - 0.20 - 0.15 - 0.10 - 0.05

Figure 10.14: Dimensionless vorticity field for Blasius boundary layer.

we find δ0.99 = 1.9 cm. We also note that the boundary layer grows with the square root
of distance along the plate. We further note that higher kinematic viscosity leads to thicker
boundary layers, while lower kinematic viscosity lead to thinner boundary layers.
The velocity ṽ has some non-intuitive√behavior. It is plotted in Fig. 10.15. As seen from
its definition in Eq. (10.331), scaling by 2 x is required to capture the exclusive dependency
on η. And the calculation reveals that
√
lim 2 x ṽ = 1.72. (10.377)
η→∞

We can unravel the various scalings then to get the following as η → ∞:

√
x ṽ = 0.86, (10.378)
r
xdim √
Re v = 0.86, (10.379)
L
r r
xdim UL vdim
= 0.86, (10.380)
L ν U
r
Uxdim vdim
= 0.86, (10.381)
ν U
vdim
= 0.86Re−1/2
xdim . (10.382)
U
We might expect our theory to force vdim → 0 as ydim → ∞. This is obviously not the case
for our approximation, especially for small xdim . The remedy is a complicated problem in
asymptotic analysis as outlined by Van Dyke (1982), Ch. 7.
Now let us determine the shear stress at the wall, and the viscous force acting on the
wall. So let us find
∂udim
τw = µ . (10.383)
∂ydim ydim =0

© 05 December 2025, Joseph M. Powers. All rights reserved.

10.2. SIMILARITY SOLUTIONS 467

0.5 1.0 1.5

√
Figure 10.15: Velocity component profile for 2 x ṽ in Blasius boundary layer.

Consider
∂u ∂2ψ 1 d2 f
= = √ , (10.384)
∂ ỹ ∂ ỹ 2 x dη 2


∂ udim
U 1 d2 f
q  = p xdim 2 , (10.385)
∂ U L ydim
L
dη
ν L
s
∂udim ρU 1 d2 f
= U √ , (10.386)
∂ydim µ xdim dη 2
r
∂udim ρUµ d2 f
τ = µ =U , (10.387)
∂ydim xdim dη 2
r
τ (0) µ d2 f
1 = Cf = 2 (0), (10.388)
2
ρU 2 ρUxdim dη 2
d2 f
Cf = 2Re−1/2
xdim (0), (10.389)
dη 2
0.664...
= p . (10.390)
Rexdim
We notice that at xdim = 0 that the stress is infinite. This seeming problem is seen not to be
one when we consider the actual viscous force on a finite length of plate. Consider a plate
of length L and width b. Then the viscous force acting on the plate is
Z L
F = τ dA, (10.391)
0

© 05 December 2025, Joseph M. Powers. All rights reserved.

468 CHAPTER 10. VISCOUS INCOMPRESSIBLE LAMINAR FLOW

Z L
= τ (xdim , 0)b dxdim , (10.392)
0
Z L p 1
= b f ′′ (0)U ρUµ √ dxdim , (10.393)
0 xdim
p Z L
′′ dxdim
= bf (0)U ρUµ √ , (10.394)
0 xdim
√ p L
= bf ′′ (0)U ρUµ (2 xdim )0 , (10.395)
p √
= 2bf ′′ (0)U ρUµ L, (10.396)
r
F ′′ µ −1/2 −1/2
1 2
= CD = 4f (0) = 4f ′′ (0)ReL = 1.328ReL . (10.397)
2
ρU Lb ρUL
Now let us consider the thermal boundary layer. Here we will take the boundary condi-
tions so that the wall and far field are held at a constant fixed temperature Tdim = To . We
need to do the scaling on the energy equation, so let us start with the steady incompressible
two-dimensional dimensional energy equation:
   2 
∂Tdim ∂Tdim ∂ Tdim ∂ 2 Tdim
ρcp udim + vdim = k + (10.398)
∂xdim ∂ydim ∂x2dim 2
∂ydim
 2  2  2 !
∂udim ∂vdim ∂udim ∂vdim
+µ 2 +2 + + .
∂xdim ∂ydim ∂ydim ∂xdim
Taking as before,
xdim ydim Tdim − To udim vdim
x= , y= , T = , u= , v= . (10.399)
L L To U U
Making these substitutions, we get
   
ρcp UTo ∂T ∂T kTo ∂ 2 T ∂2T
u +v = + (10.400)
L ∂x ∂y L2 ∂x2 ∂y 2
 2  2  2 !
µU 2 ∂u ∂v ∂u ∂v
+ 2 2 +2 + + ,
L ∂x ∂y ∂y ∂x
 2 
∂T ∂T k ∂ T ∂2T
u +v = + (10.401)
∂x ∂y ρcp UL ∂x2 ∂y 2
 2  2  2 !
µU ∂u ∂v ∂u ∂v
+ 2 +2 + + .
ρcp LTo ∂x ∂y ∂y ∂x
(10.402)
Now we have
k k µ 1 1
= = , (10.403)
ρcp UL cp µ ρUL P r Re
µU µ U2 Ec
= = . (10.404)
ρcp LTo ρUL cp To Re

© 05 December 2025, Joseph M. Powers. All rights reserved.

10.2. SIMILARITY SOLUTIONS 469

So the dimensionless energy equation with boundary conditions can be written as

 2 
∂T ∂T 1 ∂ T ∂2T
u +v = + (10.405)
∂x ∂y P rRe ∂x2 ∂y 2
 2  2  2 !
Ec ∂u ∂v ∂u ∂v
+ 2 +2 + + ,
Re ∂x ∂y ∂y ∂x
T (x, 0) = 0, T (x, ∞) = 0. (10.406)
Now as Re → ∞, we see that T = 0 is a solution that satisfies the energy equation and
all boundary conditions. For finite Reynolds number, non-zero velocity gradients √ generate
a temperature
√ field. Once again, we rescale in the boundary layer using ṽ = Re v, and
ỹ = Re y. This gives
 2 
∂T 1 1 ∂T 1 ∂ T ∂2T
u +√ √ ṽ = + Re 2 (10.407)
∂x Re 1/ Re ∂ ỹ P rRe ∂x2 ∂ ỹ
 2  2  2 !
Ec ∂u ∂ṽ √ ∂u 1 ∂ṽ
+ 2 +2 + Re +√ .
Re ∂x ∂ ỹ ∂ ỹ Re ∂x
 
∂T ∂T 1 1 ∂2T ∂2T
u + ṽ = + (10.408)
∂x ∂ ỹ P r Re ∂x2 ∂ ỹ 2
 2  2  2 !
2 ∂u 2 ∂ṽ ∂u 1 ∂ṽ
+Ec + + + .
Re ∂x Re ∂ ỹ ∂ ỹ Re ∂x
(10.409)
Now as Re → ∞,
 2
∂T ∂T 1 ∂2T ∂u
u + ṽ = + Ec . (10.410)
∂x ∂ ỹ P r ∂ ỹ 2 ∂ ỹ
√ √
Now take T = √ T (η) with η = ỹ/ x as well as u = df /dη, ṽ = (1/(2 x)) (η(df /dη) − f ) and
∂u/∂ ỹ = (1/ x)(d2 f /dη 2 ). We also have for derivatives, that
 
∂T dT ∂η dT 1η
= = − , (10.411)
∂x dη ∂x dη 2x
∂T dT ∂η dT 1
= = √ , (10.412)
∂ ỹ dη ∂ ỹ dη x
   
∂2T ∂ ∂T ∂ 1 dT 1 ∂ dT 1 d2 T
= = √ = √ = . (10.413)
∂ ỹ 2 ∂ ỹ ∂ ỹ ∂ ỹ x dη x ∂ ỹ dη x dη 2
The energy equation is then rendered as
     2
df 1 η dT 1 df 1 dT 1 1 d2 T Ec d2 f
− + √ η −f √ = + , (10.414)
dη 2 x dη 2 x dη x dη P r x dη 2 x dη 2
|{z} | {z } | {z } | {z } | {z } | {z }
u ∂T ṽ ∂T 1 ∂2 T ∂u 2
∂x ∂ ỹ P r ∂ ỹ 2 Ec( ∂ ỹ )

© 05 December 2025, Joseph M. Powers. All rights reserved.

470 CHAPTER 10. VISCOUS INCOMPRESSIBLE LAMINAR FLOW

   2
1 df dT 1 df dT 1 d2 T d2 f
− η + η −f = + Ec , (10.415)
2 dη dη 2 dη dη P r dη 2 dη 2
 2 2
1 dT 1 d2 T df
− f = + Ec ,(10.416)
2 dη P r dη 2 dη 2
| {z } | {z } | {z }
advection energy diffusion dissipation
2
 2
2
dT 1 dT df
2
+ Pr f = −P rEc , (10.417)
dη 2 dη dη 2
T (0) = 0, T (∞) = 0. (10.418)

Now for Ec → 0, we get T = 0 as a solution that satisfies the governing differential equation
and boundary conditions. Let us consider a solution for non-trivial Ec, but for P r = 1.
We could extend this for general values of P r as well. Here, following Reynolds analogy,
when thermal diffusivity equals momentum diffusivity, we expect the temperature field to
be directly related to the velocity field. For P r = 1, the energy equation reduces to
 2 2
d2 T 1 dT df
2
+ f = −Ec , (10.419)
dη 2 dη dη 2
|{z} | {z } | {z }
energy diffusion advection dissipation
T (0) = 0, T (∞) = 0. (10.420)

Here the integrating factor is Rη 1

e 0 2 f (t) dt
. (10.421)
Multiplying the energy equation by the integrating factor gives
2
 2
Rη 1
f (t) dt d T 1 R η 1 f (t) dt dT Rη 1
f (t) dt d2 f
e 0 2 + f e 0 2 = −Ec e 0 2 , (10.422)
dη 2 2 dη dη 2
 R   2
d η 1
f (t) dt dT Rη 1 d2 f
e0 2 = −Ec e 0 2 f (t) dt
. (10.423)
dη dη dη 2

Now from the x momentum equation, f ′′′ + 12 f f ′′ = 0, we have

f ′′′
f = −2 . (10.424)
f ′′
So we can rewrite the integrating factor as
f ′′ (0)
′′′  ′′ 
Rη 1
Rη 1 (−2)f f (η)
dt − ln f ′′ (0)
e 0 2 f (t) dt
=e 0 2 f ′′ =e = . (10.425)
f ′′ (η)
So the energy equation can be written as
   ′′   2 2
d f ′′ (0) dT f (0) d f
= −Ec , (10.426)
′′
dη f (η) dη ′′
f (η) dη 2

© 05 December 2025, Joseph M. Powers. All rights reserved.

10.2. SIMILARITY SOLUTIONS 471

d2 f
= −Ecf ′′ (0) , (10.427)
dη 2
Z η 2
f ′′ (0) dT ′′ df
= −Ecf (0) 2
ds + C1 , (10.428)
f ′′ (η) dη 0 ds
Z
dT d2 f η d2 f d2 f
= −Ec 2 ds + C 1 , (10.429)
dη dη 0 ds2 dη 2
 
2
d f df d2 f
= = −Ec 2  − f ′ (0) + C1 2 , (10.430)
dη dη | {z } dη
=0
2 2
d f df df
= −Ec 2
+ C1 2 , (10.431)
dη dη dη
 2 !
d 1 df d2 f
= −Ec + C1 2 , (10.432)
dη 2 dη dη
 2
Ec df df
T = − + C1 + C2 , (10.433)
2 dη dη
Ec
T (0) = 0 = − (f ′(0))2 + C1 f ′ (0) +C2 , (10.434)
2 | {z } | {z }
=0 =0
C2 = 0, (10.435)
Ec ′
T (∞) = 0 = − (f (∞))2 + C1 f ′ (∞), (10.436)
2 | {z } | {z }
=1 =1
Ec
C1 = , (10.437)
2  
Ec df df
T (η) = 1− , (10.438)
2 dη dη
Ec
= u(η)(1 − u(η)). (10.439)
2
A plot of the temperature profile for P r = 1 and Ec = 1 is given in Fig. 10.16.

© 05 December 2025, Joseph M. Powers. All rights reserved.

472 CHAPTER 10. VISCOUS INCOMPRESSIBLE LAMINAR FLOW

6
Pr = 1
Ec = 1
5

1 T

Figure 10.16: Temperature profile for Blasius boundary layer, Ec = 1, P r = 1.

© 05 December 2025, Joseph M. Powers. All rights reserved.

Bibliography

This bibliography focuses on books that are closely related to the material presented in
this course in classical fluid mechanics, especially with regard to graduate level treatment
of continuum mechanical principles applied to fluids, compressible flow, viscous flow, and
vortex dynamics. It also has some general works of historic importance. It is by no means
a comprehensive survey of works on fluid mechanics. Only a few works are given here
that focus on such important topics as low Reynolds number flows, turbulence, bio-fluids,
computational fluid dynamics, microfluids, molecular dynamics, magneto-hydrodynamics,
geo-physical flows, rheology, astrophysical flows, as well as elementary undergraduate texts.
That said, those that are listed are among the best that exist and would be useful to examine.

I. H. Abbott, Theory of Wing Sections, Dover, New York, 1980.

D. J. Acheson, Elementary Fluid Dynamics, Oxford University Press, Oxford, 1990.

J. D. Anderson, Modern Compressible Flow with Historical Perspective, 4th edition, McGraw-
Hill, New York, 2021.

J. D. Anderson, Hypersonic and High Temperature Gas Dynamics, 3rd edition, AIAA,
Reston, Virginia, 2019.

H. Aref and S. Balachandar, A First Course in Computational Fluid Dynamics, Cambridge

University Press, Cambridge, UK, 2018.

R. Aris, Vectors, Tensors, and the Basic Equations of Fluid Mechanics, Dover, New York,
1962.

H. Ashley and M. Landahl, Aerodynamics of Wings and Bodies, Dover, New York, 1985.

G. I. Barenblatt, Scaling, Self-Similarity, and Intermediate Asymptotics, Cambridge Uni-

versity Press, New York, 1996.

A. B. Basset, 1888, A Treatise on Hydrodynamics with Numerous Examples, Vols. 1,2,

Deighton, Bell, and Co., Cambridge, UK.

G. Batchelor, An Introduction to Fluid Dynamics, Cambridge University Press, Cambridge,

UK, 2000.

473
J. Bear, Dynamics of Fluids in Porous Media, Dover, New York, 1988.

P. S. Bernard, Fluid Dynamics, Cambridge University Press, New York, 2015.

D. Bernoulli, Hydrodynamica, Sive de Viribus et Motibus Fluidorum Commentarii, Johann

Reinhold Dulsseker, Strasbourg, 1738.

R. B. Bird, W. E. Stewart, and E. N. Lightfoot, Transport Phenomena, Revised 2nd edition,

John Wiley, New York, 2007.

R. B. Bird, R. C. Armstrong, and O. Hassager, Dynamics of Polymeric Liquids, Vol. 1:

Fluid Mechanics, Vol. 2: Kinetic Theory, John Wiley, New York, 1987.

A. I. Borisenko and I. E. Tarapov, Vector and Tensor Analysis with Applications, Dover,
New York, 1968.

R. S. Brodkey, The Phenomena of Fluid Motions, Dover, New York, 1995.

B. J. Cantwell, Introduction to Symmetry Analysis, Cambridge University Press, New York,

2002.

S. Chandrasekhar, Hydrodynamic and Hydromagnetic Stability, Dover, New York, 1961.

S. Chapman and T. G. Cowling, The Mathematical Theory of Non-Uniform Gases, 3rd

edition, Cambridge University Press, Cambridge, UK, 1991.

R.-H. Chen, Foundations of Gas Dynamics, Cambridge University Press, New York, 2017.

R. Chevray and J. Mathieu, Topics in Fluid Mechanics, Cambridge University Press, Cam-
bridge, UK, 1993.

A. J. Chorin and J. E. Marsden, A Mathematical Introduction to Fluid Mechanics, 3rd

edition, Springer, New York, 2000.

T. J. Chung, Computational Fluid Dynamics, 2nd edition, Cambridge University Press,

New York, 2010.

T. C. Corke, Flow Control: a Fluid Instability Approach, Cambridge University Press,

Cambridge, UK, 2024.

R. Courant and K. O. Friedrichs, Supersonic Flow and Shock Waves, Springer, New York,
1976.

W. O. Criminale, T. L. Jackson, and R. D. Joslin, Theory and Computation in Hydrody-

namic Stability, 2nd edition, Cambridge University Press, Cambridge, UK, 2019.

474
R. M. Cummings, W. H. Mason, S. A. Morton, and D. R. McDaniel, Applied Computational
Aerodynamics: A Modern Engineering Approach, Cambridge University Press, New
York, 2015.

I. G. Currie, Fundamental Mechanics of Fluids, 4th edition, CRC Press, Boca Raton,
Florida, 2013.

A. W. Date, Introduction to Computational Fluid Dynamics, Cambridge University Press,

New York, 2005.

S. R. de Groot and P. Mazur, Non-Equilibrium Thermodynamics, Dover, New York, 1984.

P. G. Drazin and W. H. Reid, Hydrodynamic Stability, Cambridge University Press, Cam-
bridge, UK, 1981.
P. G. Drazin, Introduction to Hydrodynamic Stability, Cambridge University Press, Cam-
bridge, UK, 2002.

D. A. Drew and S. L. Passman, Theory of Multicomponent Fluids, Springer, New York,

1999.

P. A. Durbin and G. Medic, Fluid Dynamics with a Computational Perspective, Cambridge

University Press, New York, 2007.

F. Durst, Fluid Mechanics: An Introduction to the Theory of Fluid Flows, Springer, Berlin,
2008.

G. Emanuel, Analytical Fluid Dynamics, 3rd edition, CRC Press, Boca Raton, Florida,
2016.

G. Emanuel, Gasdynamics: Theory and Applications, AIAA, New York, 1986.

A. C. Eringen, Mechanics of Continua, 2nd edition, Krieger, Malabar, Florida, 1989.

J. H. Ferziger and M. Peric, Computational Methods for Fluid Dynamics, 3rd revised edition,
Springer, New York, 2002.

R. P. Feynman, R. B. Leighton, and M. Sands, The Feynman Lectures on Physics, 3 vols.,

Addison-Wesley Reading, MA, 1963.

Y. C. Fung, Foundations of Solid Mechanics, Prentice-Hall, Englewood Cliffs, New Jersey,

1965.
M. Gad-el-Hak, Flow Control: Passive, Active and Reactive Flow Management, Cambridge
University Press, Cambridge, UK, 2000.
H. Glauert, The Elements of Aerofoil and Airscrew Theory, Cambridge University Press,
Cambridge, UK, 1983.

475
H. Goldstein, Classical Mechanics, Addison-Wesley, Reading, Massachusetts, 1950.

S. Goldstein, ed., Modern Developments in Fluid Dynamics, Vols. I and II, Dover, New
York, 1965.

W. T. Grandy, Entropy and the Time Evolution of Macroscopic Systems, Oxford University
Press, Oxford, 2008.

R. A. Granger, Fluid Mechanics, Dover, New York, 1995.

W. D. Hayes and R. F. Probstein, Hypersonic Flow Theory, Academic Press, New York,
1959.

H. Helmholtz, On the Sensations of Tone, Dover, New York, 1954.

I. H. Herron and M. R. Foster, Partial Differential Equations in Fluid Dynamics, Cambridge

University Press, New York, 2008.

C. Hirsch, Numerical Computation of Internal and External Flows, Vols. 1 and 2, John
Wiley, New York, 1989.

J. O. Hirschfelder, C. F. Curtiss, and R. B. Bird, Molecular Theory of Gases and Liquids,

Wiley, New York, 1954.

W. F. Hughes and E. W. Gaylord, Basic Equations of Engineering Science, Schaum’s

Outline Series, McGraw-Hill, New York, 1964.

J. D. Jackson, Classical Electrodynamics, 3rd edition, John Wiley, New York, 1999.

C. S. Jog, Fluid Mechanics: Foundations and Applications of Mechanics, Vol. 2, Cambridge

University Press, Delhi, 2015.

R. T. Jones, Wing Theory, Princeton University Press, Princeton, 1990.

D. D. Joseph, Fluid Dynamics of Viscoelastic Liquids, Springer, New York, 1990.

G. E. Karniadakis and A. Beskok, Microflows: Fundamentals and Simulation, Springer,

Berlin, 2001.

G. E. Karniadakis and S. J. Sherwin, Spectral/HP Element Methods for CFD, 2nd edition,
Oxford University Press, Oxford, 2005.

J. Katz and A. Plotkin, Low-Speed Aerodynamics, 2nd edition, Cambridge University Press,
New York, 2001.

D. C. Kay, Tensor Calculus, Schaum’s Outline Series, McGraw-Hill, New York, 2011.

C. Kleinstreuer, Modern Fluid Dynamics, 2nd edition, CRC Press, Boca Raton, 2018.

476
E. L. Koschmieder, Bènard Cells and Taylor Vortices, Cambridge University Press, Cam-
bridge, UK, 1993.

A. M. Kuethe and C.-Y. Chow, Foundations of Aerodynamics: Bases of Aerodynamic De-

sign, 5th edition, John Wiley, New York, 1998.

P. K. Kundu, I. M. Cohen, D. R. Dowling, and J. Capecelatro, Fluid Mechanics, 7th edition,

Academic Press, Amsterdam, 2025.

H. Lamb, Hydrodynamics, 6th edition, Dover, New York, 1993.

F. W. Lanchester, Aerodynamics, Constable, London, 1905.

C. B. Laney, Computational Gasdynamics, Cambridge University Press, Cambridge, UK,

1998.

M. T. Landahl, Unsteady Transonic Flow, Cambridge University Press, Cambridge, UK,

1989.

L. D. Landau and E. M. Lifshitz, Fluid Mechanics, Pergamon Press, Oxford, 1959.

L. G. Leal, Advanced Transport Phenomena: Fluid Mechanics and Convective Transport

Processes, Cambridge University Press, New York, 2007.

L. G. Leal, Laminar Flow and Convective Transport Processes: Scaling Principles and
Asymptotic Analysis, Butterworth-Heinemann, Stoneham, Massachusetts, 1992.

R. J. LeVeque, Numerical Methods for Conservation Laws, Birkhäuser, Basel, 1992.

R. J. LeVeque, Finite Volume Methods for Hyperbolic Problems, Cambridge University

Press, Cambridge, UK, 2002.

H. W. Liepmann and A. Roshko, Elements of Gasdynamics, Dover, New York, 2002.

V. L. Liseikin, Grid Generation Methods, 2nd edition, Springer, Dordrecht, 2010.

J. Lighthill, Waves in Fluids, Cambridge University Press, Cambridge, UK, 1978.

A. J. McConnell, Applications of Tensor Analysis, Dover, New York, 1957.

R. E. Meyer, Introduction to Mathematical Fluid Dynamics, Dover, New York, 1982.

L. M. Milne-Thompson, Theoretical Aerodynamics, 4th edition, Dover, New York, 1958.

L. M. Milne-Thompson, Theoretical Hydrodynamics, 5th edition, Dover, New York, 1996.

P. M. Morse and H. Feshbach, Methods of Theoretical Physics, Vols. 1 and 2, McGraw-Hill,

New York, 1953.

477
I. Newton, Principia, Volume I, The Motion of Bodies, University of California Press,
Berkeley, 1934.

J. M. Ottino, The Kinematics of Mixing: Stretching, Chaos, and Transport, Cambridge

University Press, Cambridge, UK, 1989.

S. Paolucci, Continuum Mechanics and Thermodynamics of Matter, Cambridge University

Press, New York, 2016.

R. L. Panton, Incompressible Flow, 5th edition, John Wiley, New York, 2024.

D. Pnueli and C. Gutfinger, Fluid Mechanics, Cambridge University Press, Cambridge, UK,
1992.

J. M. Powers and M. Sen, Mathematical Methods in Engineering, Cambridge University

Press, New York, 2015.

J. M. Powers, Combustion Thermodynamics and Dynamics, Cambridge University Press,

New York, 2016.

J. M. Powers, Mechanics of Fluids, Cambridge University Press, Cambridge, UK, 2024.

C. Pozrikidis, Fluid Dynamics: Theory, Computation, and Numerical Simulation, 3rd edi-
tion, Springer, New York, 2017.

L. Prandtl and O. G. Tietjens, Fundamentals of Hydro and Aeromechanics, Dover, New

York, 1957.

M. Rieutord, Fluid Dynamics: An Introduction, Springer, Heidelberg, 2015.

L. Rosenhead, ed., Laminar Boundary Layers, Oxford University Press, Oxford, 1963.

R. H. Sabersky, A. J. Acosta, E. G. Hauptmann, and E. M. Gates, Fluid Flow: A First

Course in Fluid Mechanics, 4th edition, Prentice-Hall, Upper Saddle River, New Jersey,
1999.

P. G. Saffman, Vortex Dynamics, Cambridge University Press, Cambridge, UK, 1992.

M. Samimy, K. S. Breuer, L. G. Leal, and P. H. Steen, A Gallery of Fluid Motion, Cambridge

University Press, Cambridge, UK, 2004.

J. A. Schetz, Boundary Layer Analysis, Prentice-Hall, Englewood Cliffs, New Jersey, 1993.

H. M. Schey, Div, Grad, Curl, and All That, 4th edition, W.W. Norton, London, 2004.

H. Schlichting and K. Gersten, Boundary Layer Theory, 9th edition, McGraw-Hill, New
York, 2017.

478
L. I. Sedov, Similarity and Dimensional Methods in Mechanics, Academic Press, New York,
1959.

L. A. Segel, Mathematics Applied to Continuum Mechanics, Dover, New York, 1987.

A. H. Shapiro, The Dynamics and Thermodynamics of Compressible Fluid Flow, Volume I,

John Wiley, New York, 1953.

A. H. Shapiro, The Dynamics and Thermodynamics of Compressible Fluid Flow, Volume II,
Krieger, Malabar, Florida, 1954.

F. S. Sherman, Viscous Flow, McGraw-Hill, New York, 1990.

J. Smoller, Shock Waves and Reaction-Diffusion Equations, 2nd edition, Springer, New
York, 1994.

J. H. Spurk, Fluid Mechanics: Problems and Solutions, Springer, New York, 1997.

G. G. Stokes, Mathematical and Physical Papers, Vols. 1-5, University of Michigan Library,
Ann Arbor, 2005.

J. W. Strutt (Lord Rayleigh), The Theory of Sound, Vols. 1 and 2, Dover, New York, 1945.

G. I. Taylor, Scientific Papers, Cambridge University Press, Cambridge, UK, 1958.

J. C. Tannehill, D. A. Anderson, and R. H. Pletcher, Computational Fluid Mechanics and

Heat Transfer, 3rd edition, CRC, Boca Raton, Florida, 2013.

P. A. Thompson, Compressible-Fluid Dynamics, McGraw-Hill, New York, 1971.

W. Thomson (Lord Kelvin), Mathematical and Physical Papers, Cambridge University

Press, Cambridge, UK, 1911.

K. S. Thorne and R. D. Blandford, Modern Classical Physics: Optics, Fluids, Plasmas,

Elasticity, and Statistical Physics, Princeton University Press, Princeton, New Jersey,
2017.

T. T. C. Ting, Anisotropic Elasticity: Theory and Applications, Oxford University Press,

New York, 1996.

D. Tong, Fluid Mechanics: Lectures on Theoretical Physics, Vol. 4, Cambridge University

Press, Cambridge, UK, 2025.

E. F. Toro, Riemann Solvers and Numerical Methods for Fluid Dynamics, Springer, New
York, 1999.

D. J. Tritton, Physical Fluid Dynamics, 2nd edition, Oxford University Press, Oxford, 1988.

479
C. A. Truesdell, A First Course in Rational Continuum Mechanics, Vol. 1, 2nd edition,
Academic Press, Boston, 1991.

P. G. Tucker, Advanced Computational Fluid and Aerodynamics, Cambridge University

Press, New York, 2016.

M. Van Dyke, Perturbation Methods in Fluid Mechanics, Parabolic Press, Stanford, Cali-
fornia, 1975.

M. Van Dyke, An Album of Fluid Motion, Parabolic Press, Stanford, California, 1982.

W. G. Vincenti and C. H. Kruger, Introduction to Physical Gas Dynamics, Wiley, New

York, 1965.

R. von Mises, Theory of Flight, McGraw-Hill, New York, 1945.

R. von Mises, Mathematical Theory of Compressible Flow, Dover, Mineola, New York, 2004.

J. von Neumann, Collected Works, Pergamon, New York, 1961.

S. Whitaker, Introduction to Fluid Mechanics, Krieger, Malabar, Florida, 1968.

F. M. White and J. Majdalani, Viscous Fluid Flow, 4th edition, McGraw-Hill, New York,
2022.

F. M. White, Fluid Mechanics, McGraw-Hill, New York, 1986.

G. B. Whitham, Linear and Nonlinear Waves, John Wiley, New York, 1974.

L. C. Woods, The Thermodynamics of Fluid Systems, Clarendon Press, Oxford, 1975.

L. C. Woods, The Theory of Subsonic Plane Flow, Cambridge University Press, Cambridge,
UK, 1961.

C.-S. Yih, Fluid Mechanics, West River Press, East Hampton, Connecticut, 1977.

Y. B. Zeldovich and Y. P. Raizer, Physics of Shock Waves and High-Temperature Hydrody-

namic Phenomena, Dover, New York, 2002.

R. Kh. Zeytounian, Theory and Applications of Nonviscous Fluid Flows, Springer, Berlin,
2002.

R. Kh. Zeytounian, Theory and Applications of Viscous Fluid Flows, Springer, Berlin, 2004.

M. J. Zucrow and J. D. Hoffman, Gas Dynamics, Vol. I, John Wiley, New York, 1976.

480