3 Balance Equations

In this chapter we provide in short form the classical balance principles. For detailed
studies and consequences of these relations in continuum mechanics we refer to the
fundamental works of Coleman, Truessdell and Noll [CN64, TN65, TT60], and to com-
prehensive textbooks as, e.g., [Ber05, Cha76, Hol00, Mül01].
In classical solid mechanics the wanted fields determined by the balance equations are
defined on time t ∈ IR+ and position x ∈ IR3 of a set of particles P occupying a region
B ⊂ IR3 . The hypothesis of a continuum states that a volume element dV is associated
with each event in space-time (x, t); this allows to define macroscopic fields with a certain
continuity. Primarily we focus on bodies or systems of sub-bodies, respectively, which
are modelled as a closed system. In a closed system the mass of the bodies remains
constant under all deformation. This excludes processes of mass growth and loss for
which the system is commonly modelled as open, presuming a constant volume instead,
see Section 3.7.
The solids under consideration here deform and move over a period of time under the
action of a system of externally applied forces and prescribed displacements. The forces
may consist of body forces which are the result of interaction at a distance, as well as
of traction acting on the surface of the body. By B̄ : B → IR3 we denote the material
body force field per unit mass and by b̄ : ϕ(B) → IR3 the spatial body force field
per unit mass. A relation between the two fields may be established considering an
infinitesimal neighborhood of mass dm and letting df be the resultant body force acting
on the neighborhood. With the deformation mapping x = ϕ(X, t) we may write
df = B̄(X, t) dm = b̄(x, t) dm. (3.1)
Thus we get b̄ ◦ ϕ = B̄.
Analogously, we denote by T̄ : S → IR3 the material surface traction field per
unit reference area, and by t̄ : ϕ(S) → IR3 the spatial surface traction field per
unit deformed area. Denote by dS and ds the infinitesimal areas containing X and x,
respectively, and let N and n be the corresponding outward unit normals, Figure 3.1.
Then, the total force acting on an element of area is
df = T̄ (X, t) dS = t̄(x, t) ds. (3.2)

16
 Figure 3.1: Piola transformation.

We employ Cauchy’s tetrahedron theorem from which it follows that the traction
depend linearly on the corresponding unit normal, i.e.,

t̄(x, t) = σ(x, t)n(x, t) (3.3)

where σ : ϕ(B) → IR3×3 is the Cauchy stress tensor comprising the ”true” stress at
the deformed material point. Correspondingly, we may formulate these relations for the
reference placement,

T̄ (X, t) = P (X, t)N (X, t), (3.4)

where P : B → IR3×3 is the first Piola-Kirchhoff stress tensor. Inserting (3.3) and
(3.4) in (3.2) and applying Nanson’s formula (2.20) yields the following relation between
the stress tensors,

σ = J −1 P F ⊤ . (3.5)

Relation (3.5) is known as Piola transformation. Written in components it reads

σij = J −1 (X, t) PiJ (X, t) FJj (X, t) ⇔ −1

PiJ = J(X, t) σij (X, t) FjJ (X, t),

where the indices emphasize that σ is defined on the current placement whereas P is a
two-point tensor, i.e., it works between current and reference placement.
Additionally, we introduce here a symmetric tensor measuring the stress with respect to
the reference placement, namely, the second Piola-Kirchhoff stress tensor, S : B →

17
3.1 Conservation of mass

IR3×3 . The following relations hold

σ = J −1 F SF ⊤ (3.6)
P = F S, (3.7)
and, vice versa,
S = J F −1 σF −⊤ (3.8)
= F −1 P . (3.9)

3.1 Conservation of mass

Let ̺0 (X, t) be the mass density per unit undeformed volume and ̺(x, t) the corre-
sponding mass density per unit deformed volume. Now we consider an infinitesimal
neighborhood of X ∈ B at time t of volume dV0 and let dV be the volume of the corre-
sponding spatial neighborhood. Since both neighborhoods are in correspondence by the
deformation mapping, the total mass contained in them must be identical, i.e.
dm = ̺0 (X, t)dV0 = ̺(x, t)dV. (3.10)
Combining this relation with (2.16) gives the simple algebraic equation1
̺0 = J̺, (3.11)
which is the Lagrangian local form of mass balance. Within the Eulerian frame we obtain
a partial differential equation known as continuity equation, see Table 3.1.

3.2 Conservation of linear momentum

The total linear momentum L of the body expressed in material form is

Z
L= ̺0 V dV, (3.12)
B

where V denotes the material velocity. Likewise the resultant of the forces acting on
the body F̄ may be written in the form
Z Z
F̄ = ̺0 B̄ dV + P N dS, (3.13)
B S
1
For brevity we omit in the following the arguments but we silently understand all relevant quantities
to depend on position and time.

18
3.3 Conservation of angular momentum

with the first Piola-Kirchhoff tensor P and the outward unit normal N as above. Then,
the principle of linear momentum (or Newtons second law) simply states
dL
= F̄ . (3.14)
dt
Inserting the corresponding integral expressions equation (3.14) reads
Z Z Z
̺0 AdV = ̺0 B̄dV + P N dS, (3.15)
B B S

where A is the material acceleration defined by (2.8). Here and further on we assume
the material mass density ̺0 to be constant in time. An application of the divergence
theorem gives2
Z


̺0 (A − B̄) − Div P dV = 0. (3.16)
B

The balance equation applies to the body as well as to all ’sub-bodies’ of B, i.e., an
identical identity must hold for every open subset Ω ⊂ B. This in turn requires the
integrand in (3.15) to equal zero.

̺0 A = ̺0 B̄ + Div P a.e. in B, (3.17)

Equation (3.17) states the Lagrangian local form of linear momentum balance. In qua-
sistatic deformations, i.e, deformations with negligible influence of acceleration, the left
hand side of equation (3.17) vanishes and the remaining terms state the local equilibrium
condition.

3.3 Conservation of angular momentum

The total angular momentum G of the body reads in material form

Z
G= x × (̺0 V ) dV. (3.18)
B

Likewise, the resultant moment M acting on the body may be expressed in the form
Z Z
M= x × (̺0 B̄) dV + x × (P N ) dS. (3.19)
B S
2
By Div(·) and Grad(·) we denote the divergence and the gradient with respect to the material frame,
respectively, see Appendix 2.

19
3.4 The deformation power identity

Under these conditions conservation of angular momentum means

dG
= M. (3.20)
dt
Equation (3.19) applied to a deformed (non-polar) material element simply requires the
Cauchy stress tensor to be symmetric,

σ = σ⊤ in ϕ(B), (3.21)

which is the spatial local form of conservation of angular momentum. Pulled back into
the reference placement by virtue of (3.5) it follows the Lagrangian local form of angular
momentum,

PF⊤ = FP⊤ in B. (3.22)

3.4 The deformation power identity

The kinetic energy of the body written in Lagrangian form is

1
Z
K= ̺0 |V |2 dV. (3.23)
2 B

The external power P E is the power of all externally applied force fields, i.e., the sum
of resulting body forces and traction.
Z Z
E
P = ̺0 B̄ · V dV + P N · V dS. (3.24)
B S

The deformation power P D is by definition the amount of the external power supply
which is not converted into kinetic energy of the body,

P D = P E − K̇. (3.25)

Inserting (3.23) and (3.24) in (3.25) gives

Z Z Z
D
P = ̺0 B̄ · V dV + P N · V dS − ̺0 A · V dV. (3.26)
B S B

By application of the divergence theorem we get

Z Z
D
P = ̺0 (B̄ − A) + Div P ) dV + P · Grad V dV, (3.27)
B B

20
3.5 Conservation of mechanical energy

and simple algebra gives Grad(V ) = Ḟ . If the motion satisfies the linear momentum
balance the first integral in (3.27) vanishes. Then, the deformation power follows as,
Z
D
P = P · Ḟ dV. (3.28)
B

Equation (3.28) shows, that the Piola-Kirchhoff stress tensor P and the deformation gra-
dient F are work-conjugate variables, i.e., their scalar product defines a rate of internal
mechanical work. With the definitions of Section 2.4 other work-conjugate expressions
are given by

P · Ḟ = S · Ė = J σ · d = τ · d. (3.29)

Tensor τ is the frequently used Kirchhoff-stress tensor, τ : B → IR3×3 with τ = Jσ.

3.5 Conservation of mechanical energy

Let U : B 7→ IR denote the internal energy density per unit undeformed volume and
Z Z
E= U dV = J −1 U dV (3.30)
B ϕ(B)

the total internal energy of the body. For elastic bodies the conservation of energy than
demands

P E = Ė + K̇, (3.31)

i.e., the external power will be converted, without gains or losses, into internal or kinetic
energy. By the definition of deformation power (3.25) and the deformation-power identity
(3.28) this reduces to the tautology Ė = P D .
Z Z Z Z
̺0 B̄ · V dV + T̄ · V dS = P · Ḟ dS + ̺0 A · V dV. (3.32)
B S B B

Writing the conservation of mechanical energy in Lagrangian local form results in

U̇ = P · Ḟ . (3.33)

Equation (3.33) does not furnish a new field equation.

Table 3.1 summarizes the local balance equations in material and spacial form.

21
3.6 Thermal energy

mass:
d̺ dvi
̺0 = J̺ dt
= ̺ dx i
linear momentum:
̺0 Ai = ̺0 B̄i + PiJ,J ̺ai = ̺b̄i + σiJ,J
angular momentum:
σij = σij PiJ FjJ = PjJ FiJ
mechanical energy:
U̇ = PiJ ḞiJ U̇ = JσiJ diJ
thermal energy:
∂QL ∂qL
U̇ = PiJ ḞiJ − ∂XL
+R J −1 U̇ = σiJ diJ − ∂xL
+r

Table 3.1.: Local balance equation in Lagrangian and Eulerian form.3

3.6 Thermal energy

In this section we deviate from the restriction to reversible processes and consider both
mechanical and thermal energy. For later reference we state here the fundamental laws
of thermodynamics.

First law of thermodynamics

Let Q : B 7→ IR be the thermal power or the rate of thermal work. With the vector of
heat flux per unit time and per unit reference and current surface Q and q, respectively,
and with the heat source R and r per unit reference and current surface, the thermal
work is determined to be
Z Z Z Z
Q = R dV − Q · N dS = r dV − q · n dS. (3.34)
B S ϕ(B) ϕ(S)

In a thermomechanical system the conservation of energy demands

P E + Q = Ė + K̇ . (3.35)

The thermal power will, additionally to the external (mechanical) power, completely
be invested to raise the internal or kinetic energy of the body. Thus, the first law of
3
Written in components, here and subsequently, a comma denotes partial differentiation.

22
3.6 Thermal energy

thermodynamics in material form reads

Z Z Z Z



̺0 B̄ · V + R dV + T̄ V − Q̄ · N dS = U̇ dV + ̺0 A · V dV. (3.36)
B S B B

Following arguments as above we find the Lagrangian local form of thermal energy
balance to have the form

U̇ = P · Ḟ − Div Q + R. (3.37)

Second law of thermodynamics

Let S : B → IR denote the entropy per unit reference volume. Let Σ : B → IR+ be
the total production of entropy within the body, i.e., the difference between the rate
of change of entropy and the rate of entropy input into the body. The second law of
thermodynamics states that in a thermodynamic process the total entropy production
can never be negative,

Σ ≥ 0. (3.38)

A reversible thermodynamic process is not accompanied by any entropy production, i.e.,

Σ = 0, whereas a process is called irreversible if Σ > 0.
We present the second law of thermodynamics in form of the Clausius-Duhem in-
equality. This approach links the thermal quantities heat flux Q and heat source R via
the absolute temperature T to the entropy flux Q/T and entropy source R/T , respec-
tively. Consequently, the material form the Clausius-Duhem inequality reads
R
Z Z Z
Q
Σ= Ṡ dV + · N dS − dV ≥ 0. (3.39)
B S T B T

In an adiabatic closed system, i.e., in a system without heat source and without heat
flux, R = 0, Q = 0, equation (3.39) simplifies to

Ṡ ≥ 0. (3.40)

Roughly speaking, the entropy measures the probability of a state, i.e., in equilibrium
the state with the highest entropy is the most likely one. The fact that the entropy can
only rise defines a direction for irreversible processes.
Let A : B → IR denote the Helmholtz free energy density

A = U − ST. (3.41)

23
3.7 General balance equations in a Schottky system

The Helmholz free energy density defines the internal energy density which may be con-
verted into mechanical work, sometimes named as ”work content“. With that definition
we can formulate equation (3.37) as
Ȧ = P · Ḟ − Div Q + R − ṠT − S Ṫ , (3.42)
and we derive a local form of (3.39)
Q
Ȧ + S Ṫ − P · Ḟ + Grad T ≤ 0. (3.43)
T
The inequality (3.43) is often referred to as Clausius-Planck inequality, the terms on
the left hand side describe the internal dissipation Dint . At constant temperature (3.43)
reduces to
Dint = P · Ḟ − Ȧ ≥ 0. (3.44)

For completeness we define here also the Gibbs free energy density, G : B → IR
G = U − ST + pV, (3.45)
which is a thermodynamic potential and, correspondingly, a state function of a thermo-
dynamic system.
In the variational approaches favored in this work we will repeatedly exploit the simple
idea, that every system seeks to archive a minimum of its free energy function.

3.7 General balance equations in a Schottky system

A domain Ω(t) exchanging heat, power and material with its environment is commonly
defined to be a Schottky system [Mus90, Mül85], see Figure 3.2 left. A Schottky
system is, therefore, an extension of the above defined body with constant mass B. Let
now Ψ(x, t) be an extensive quantity in a Schottky system Ω(t). A change of quantity
Ψ can be invoked by a production or source within the volume, here both denoted by
Πψ , by a supply Σψ and by a flux of quantity Ψ over the boundaries of the volume.
The latter is written as the vector of outward flux per unit time and unit surface J ψ
in direction of the normal n of surface S ≡ ∂Ω. Then the generic form of a global
balance equation reads in spacial form
dΨ d
Z
= ρψ(x, t)d V
dt dt Ω(t)
Z Z
ψ
= − J (x, t) · n d S + (Πψ (x, t) + Σψ (x, t))d V , (3.46)
∂Ω(t) Ω(t)

24
3.7 General balance equations in a Schottky system

where ρψ is the quantity Ψ(x, t) per unit volume.

ty
t icle veloci
r
heat ndS ≡ dS v : pa
∂Ω w
control
volume
Ω Ω
w
power
∂Ω
material w:
velocity of the
environment control volume

Figure 3.2: Definition of a Schottky system Ω with boundary ∂Ω, velocity w and particle
velocity v.

If the domain Ω(t) of a Schottky system coincides with a body (i.e., the system is a closed
system), there is no particle transfer over the boundary ∂Ω. The total mass contained
in it remains constant and equation (3.10) holds, in other words

d
Z
ρ(x, t)dV = 0. (3.47)
dt Ω(t)

Equation (3.47) represents a special case of equation (3.46) with ψ ≡ 1, J ψ = Πψ =

Σψ ≡ 0. On the other hand, a Schottky system may be modelled as open presuming
a constant volume moving with velocity w (right in Figure 3.2). The velocity w is
superposed to the material velocity v(x, t) of the particles P. Then holds J ψ ≡ J 1 = J
and we notify
d
Z Z
ρ(x, t)dV = − J (x, t)dS 6= 0 , (3.48)
dt Ω(t) ∂Ω(t)

where the flux of mass is given by

J = ρ(v − w). (3.49)

Note that in a control volume the total outward flux J ψ of quantity ψ is the sum of
a convective (mass related) part ρ(v − w)ψ and a conductive part J ψcond . Clearly, if
domain Ω is a body the convective part of the flux vanishes, i.e., v = w.

25
3.8 Jump conditions

With the same arguments as before and by applying the divergence theorem and Reynolds’
transport theorem (see Appendix 2) we obtain the local form of the generic balance
equation (cf. [Mül01])

∂ρψ
+ ∇ · w ρ ψ + J ψ = Πψ + Σψ .


(3.50)
∂t
The specific balances of mass, momentum, total and internal energy follow by inserting
the known expressions (here stated in spatial form):
• mass:
ψ ≡ 1 , J 1 ≡ ρ(v − w) , Π1 ≡ 0 , Σ1 ≡ 0. (3.51)
• momentum:

ψ ≡ v , J v ≡ ρ(v − w)v − σ T , Πv ≡ 0 , Σv ≡ ρb̄. (3.52)

• total energy:

ψ ≡ e , J e ≡ ρ(v − w)e − σ T · v + q , Πe ≡ 0 , Σe ≡ ρb̄ · v + r. (3.53)

• internal energy:

ψ ≡ u , J u ≡ ρ(v − w)u + q , Πu ≡ ∇v : σ , Σu ≡ ρr. (3.54)

Note that the internal energy balance is not a conservation law, because of the nonzero
production Πu .

3.8 Jump conditions

In the foregoing we presented the global balance equations for the extensive quantities
mass, momentum and energy, and we assumed continuity to derive the local, differen-
tial form. In some problems the considered fields are not smooth but only piecewise
continuous. Let Λ be a discontinuous surface within a body B invoking (finite) jumps
in some fields quantities. Assume this singular surface travels through the body with
velocity v Λ 6= v. Then, the body is divided into two sub-bodies B+ (t) and B− (t) such
that B+ ∪ B− = B and B+ ∩ B− = Λ. Indicate the values of a field quantity Ψ on the
positive side of Λ by + and on the negative side by -. On Λ we define the jump in Θ as
the difference of the limits Θ+ and Θ− and we write

[[Ψ]] = Ψ− − Ψ+ . (3.55)

26
3.8 Jump conditions

To compute the time rate of a field quantity Ψ Reynolds’ transport theorem (??) now
needs to be applied to the domains of both sub-bodies, Ω(B+ , t) and Ω(B− , t), and, more-
over, to the surface Λ where the quantity Ψ jumps. The latter results in an additional
term in equation (A ??), which is of the form
Z
[[Ψ · v Λ − v ]] · n dS.


(3.56)
Λ

Consequently, the generic form of the balance equations for a body with a discontinuous
surface, i.e., the general global jump balance equation, reads in spacial form
d
Z Z Z
ψ
ρψ(x, t) dΩ = − J (x, t) · n dS + (Πψ + Σψ ) dΩ
dt
Ω+ (t)∪Ω− (t) (∂Ω+ ∪∂Ω− )\Λ Ω+ (t)∪Ω− (t)
Z
+ [[ρψ(x, t) · v Λ − v ]] · n dS,


(3.57)
Λ

with the physical field quantities, flux and source terms of equations (3.52–3.54).
The local form of equation (3.57) yields to local forms for both sub-bodies B+ and B− ,
respectively, and to an Eulerian local jump balance equation on surface Λ,

[[ρψ · v Λ − v − J ψ ]] · n = 0.


(3.58)

Note that there might be a source term on the right-hand side of equation (3.58) re-
flecting a possible area-distributed production onto the singular surface. The analogous
Lagrangian local jump balance equation on surface Λ reads

[[ρψ VNΛ ]] − [[J ψ · N ]] = 0, (3.59)

where VNΛ is the normal component of the material velocity of the travelling front Λ, i.e.,
V Λ = V · N |Λ .

27
Bibliography

[AA95] J. Altenbach and H. Altenbach. Einfüuhrung in die Kontinuumsmechanik. Teubner

Studienbücher, Stuttgart, 1995.
[BB75] E. Becker and W. Bürger. Kontinuumsmechanik. B. G. Teubner Verlag, 1975.
[Ber05] A. Bertram. Elasticity and Plasticity of Large Deformations. Springer, 2005.
[Bet93] J. Betten. Kontinuumsmechanik. Elasto-, Plasto- und Kriechmechanik. Springer-
Verlag, Berlin, 1993.
[BHW97] K Burg, H. Haf, and I. Wille. Höhere Mathematik für Ingenieure, Band I - V. B.
G. Teubner Verlag, 1997.
[BS88] I. N. Bronstein and K. A. Semendjajew. Ergänzende Kapitel zum Taschenbuch der
Mathematik. Teubner Verlagsgesellschaft, Leipzig, 1988.
[BS94] S. Brenner and L. Scott. The Mathematical Theory of Finite Element Methods.
Springer, New York, 1994.
[BW76] Ray M. Bowen and C.-C. Wang. Introduction to Vectors and Tensors, Part A:
Linear and Multilinear Algebra. Plenum Press, 1976.
[Cha76] P. Chadwick. Continuum Mechanics - Concise Theory and Problems. G. Allen &
Unwin, 1976.
[CL91] P. G. Ciarlet and J. L. Lions, editors. Handbook of Numerical Analysis. Finite
Element Methods. Elsevier Science Publishers B.V. (North–Holland), 1991.
[CN64] B. D. Coleman and W. Noll. Material symmetry and thermostatic inequalities in
finite elastic deformations. Arch. Rational Mech. Anal., 15:87–111, 1964.
[Gek06] E. Gekeler. Mathematische Methoden zur Mechanik. Springer, Berlin (to appear),
2006.
[GR80] A. Göpfert and T. Riedrich. Funktionalanalysis. In: Mathematik für Inge-
nieure, Naturwissenschaftler, Ökonomen und Landwirte. Bd.22. Teubner Verlagsge-
sellschaft, Leipzig, 1980.

28
Bibliography

[GRT88] H. Goering, H. G. Roos, and L. Tobiska. Finite Element Methode. Akademie–Verlag,

Berlin, 1988.

[Gur81] M.E. Gurtin. Introduction to Continuums Mechanics. Springer, 1981.

[Hac92] W. Hackbusch. Theorie und Numerik elliptischer Differentialgleichungen. Teubner–

Verlag, 1992.

[Hau02] P. Haupt. Continuum Mechanics and Theory of Materials. Springer, 2002.

[Hol00] G. Holzapfel. Nonlinear Solid Mechanics. Wiley & Sons, 2000.

[Lee69] E.H. Lee. Elastic-plastic deformation at finite strain. J. Appl. Mech., 36:1–6, 1969.

[Mal69] L. Malvern. Introduction to the Mechanics of a Continous Medium. Prentice-Hall,

1969.

[MH83] J. Marsden and T. Hughes. Mathematical Foundations of Elasticity. Prentice-Hall,

Englewood Cliffs, NJ, 1983.

[Mül85] Ingo Müller. Thermodynamics. Pitman, 1985.

[Mül01] Ingo Müller. Grundzüge der Thermodynamik. Springer Verlag, 2001.

[Mus90] W. Muschik. Aspects of Non-Equilibrium Thermodynamics. World Scientific, Sin-

gapore, 1990.

[R.83] Hughes T. J. R. Analysis of transient algorithms with particular reference to stability

behaviour, 1983.

[RO99] R. Radovitzky and M. Ortiz. Error estimation and adaptive meshing in strongly
nonlinear dynamic problems. Computer Methods in Applied Mechanics and Engi-
neering, 172(1-4):203–240, 1999.

[Sch97] H. Schade. Tensoranalysis. Walter de Gruyter, Berlin, 1997.

[Set64] B. R. Seth. Generalized strain measure with applications to physical problems. In

D. Abir M. Reiner, editor, Second-Order Effects in Elasticity, Plasticity and Fluid
Dynamics, pages 162–172. Pergamon Press, Oxford, 1964.

[SH00] J. J. Skrzypek and R. B. Hetnarski. Plasticity and Creep. CRC Press, 2000.

[SK91] H. Schwetlick and H. Kretzschmar. Numerische Verfahren für Naturwissenschaftler

und Ingenieure. Fachbuchverlag Leipzig, 1991.

[Tho57] T.Y. Thomas. Extended compatibility conditions for the study of surfaces of dis-
continuity in continuum mechanics. J. Math. Mech., 6:311–322, 1957.

29
Bibliography

[TN65] C. Truesdell and W Noll. The Non-Linear Field Theories of Mechanics. In Flügge,
S.: Handbuch der Physik. Volume III/3. Springer-Verlag, Berlin, 1965.

[TT60] C. Truesdell and R. A. Toupin. The Classical Field Theories. In Flügge, S.: Hand-
buch der Physik. Volume III/1. Springer-Verlag, Berlin, 1960.

[Wri01] Peter Wriggers. Nichtlineare Finite-Elemente-Methode. Springer, Berlin, 2001.

[ZT03] O. C. Zienkiewicz and R. L. Taylor. The Finite Element Method. Vol. 1-3. McGraw–
Hill Book Company, 2003.

30