DIMENSIONAL ANALYSIS

• Mathematical technique which makes use of the study of dimensions for the
solution of several engineering problems.

❖ Main applications of dimensional analysis (DA) are:

(i) To establish a relationship among the variables in a given phenomenon.

(ii) To conduct tests on models for getting the desired information i.e. models study.

• Each physical phenomenon can be expressed by an equation consisting of variables.

➢ These variables may be dimensional or non-dimensional.

• By using the methods of DA, a systematic relationship among the variables can be
established.

• To know the performance of actual structures in advance, models are made and
the tests are conducted on the models.

❖ If performance of model is satisfactory, only then actual structure is made.

PRIMARY QUANTITIES AND SECONDARY QUANTITIES

• To describe a given phenomenon, various physical quantities influencing the phenomenon

can be described by a set of quantities, which are independent of each other.

❖ These independent quantities are known as primary quantities, also called

fundamental quantities or fundamental dimensions (FD).

❖ Four fundamental dimensions are mass, length, time and temperature denoted by
M, L, T and  respectively

❖ Temperature is used as FD only in compressible fluids.

➢ All other quantities are known as secondary quantities, also called derived quantities, can
be expressed in terms of primary quantities.

Examples: Area, volume, velocity - secondary quantities.

Definition of dimensions:

• Expression of a physical quantity in terms of fundamental dimensions.

❖ Dimensions of any physical quantity can be derived from the relationship of

quantity with other variables or units of quantity.

Classification of physical quantities and their dimensions

1
 • Various physical quantities involved in the problems of fluid flow may be classified as
geometric, kinematic and dynamic quantities.

Geometric Quantities

• Most commonly used geometric quantities and their dimensions are Length [L],
diameter [L], area [L2], volume [L3], curvature (1/R) [L-1], slope [M0L0T0], angle
[M0L0T0] etc.

Kinematic Quantities

• Time [T], velocity [LT-1], acceleration [LT-2], angular velocity [T-1], angular
acceleration [T-2], frequency [T-1], discharge [L3T-1], kinematic viscosity [L2 T-1 ]etc.

Dynamic Quantities

• Mass [M], force [MLT-2], weight [MLT-2], mass density (mass/volume) [ML-3],
specific weight (weight/volume) [ML-2T-2], specific gravity[M0L0T0], , intensity of
pressure and shear stress [ML-1T-2], dynamic viscosity {τ = μ(du/dy)} [ML-1T-1],
surface tension (F/L) [MT-2], modulus of elasticity, momentum (mass  velocity) ,
work, energy and torque, Power (work/time), mass rate of flow (Q), weight rate
of flow (Q), Moment of a force (force  distance), moment of momentum
(momentum  distance) etc.

DIMENSIONAL HOMOGENEITY
• An equation is said to be dimensionally homogeneous if dimensions of terms on
left-hand side = dimensions of terms on the right hand.

❖ In other words, powers of fundamental dimensions are equal for a dimensionally

homogeneous equation.
Example: Consider the eq. for time period of oscillation of a simple pendulum i.e. t = 2  L
g
• Dimension of the term on the left-hand side = [T]
1/ 2
• Dimensions of the terms on the right-hand side =  L 
 LT − 2 

❖ 2 is a dimensionless constant.
Equation is dimensionally homogeneous.
➢ Some other dimensionally homogeneous equations are:

2
 2
• Flow over a rectangular notch, Q = CdL 2g H3 / 2
3
• Flow over a triangular notch, Q = 8 Cd 2g tan  H5 / 2
15 2
2
f LV
• Head loss due to friction in pipes, h f = 1
D(2g)
❖ Cd and f1 are dimensionless.

Characteristics of dimensionally homogeneous Equations

(i) A dimensionally homogeneous eq. is independent of the units chosen for its measurement i.e.
can be used in any one of the system of units (CGS, MKS or SI).

(ii) It is always possible to reduce a dimensionally homogeneous eq. to a non-dimensional form.

➢ Following dimensionally homogeneous eqs. can be converted into non-dimensional form:

 t   Q  8 
  = 2 ,  5/2 
= Cd 2 tan = cons tan t
 L / g   gH  15 2

Dimensionally non-homogeneous equations

❖ Ideally an equation should be dimensionally homogenous

❖ There are several equations which are non-homogeneous but are still used within
their limited ranges.

Examples: Followings two equations are commonly used for finding mean velocity of
flow, V in open channels.

V = C RhS (Chezy’s equation)

V=
1 2 / 3 1/ 2
Rh S (Manning’s equation)
N
➢ C = Chezy’s constant, Rh = Hydraulic radius or Hydraulic mean depth (Area of
flow/wetted perimeter) and S = Bed slope, N = Manning’s constant.

❖ Both equations are dimensionally non-homogeneous.

• Consider Chezy’s equation, V = C RhS
−1
• Dimension of term on the left-hand side = [LT ]
1/ 2
• Dimensions of the terms on the right-hand side = [L ]
❖ C is a constant and is dimensionless.
Chezy’s eq. is dimensionally non-homogeneous.

3
Observations:
(i) Dimensionally non-homogeneous equations can be used only in that system of unit in
which the value of constant involved in the equation is given.
(ii) Dimensionally non-homogeneous eqs. can be converted into another system of unit
by modifying the value of constant.

Example:

• Eq. V = 10.8 Rh 2 / 3 S1/ 2 is applicable in FPS system of units. The various terms
have their usual meaning. Modify this equation for SI system of units.

Solution:
• The given eq. is dimensionally non-homogenous.
• Can be written as:
 V 
10 .8 =  2 / 3 1/ 2 
 Rh S 
• Writing dimensions of terms on the RHS, to get

  LT 
−1
 V

1 / 3 −1
 2 / 3 1/ 2  =  2 / 3  = L T 
 Rh S   L 

❖ Constant 10.8 has units of (ft1/3/s)

• Since 1 m = 3.28 ft, the value of constant 10.8 in SI units will be 10.8(1/3.28)1/3 = 7.27

 Given equation in SI system of units is:

V = 7.27Rh 2 / 3 S1/ 2

METHODS OF DIMENSIONAL ANALYSIS

• If numbers of variables involved in a phenomenon are known, then the relationship
among the variables can be determined by the following two methods:

➢ Rayleigh’s method and Buckingham’s method.

Rayleigh’s method

❖ In this method, functional relationship among the variables is expressed in the

form of an exponential equation:

• If X is some function of variables say (X 1 , X 2 , X 3 − − − − − − , X n ) i.e.

X = f (X 1 , X 2 , X 3 ,− − − − −−, X n ) (1)

4
 ➢ X is called a dependent variable and (X 1 , X 2 , X 3 ,− − − − −−, X n ) are called
independent variables.

• According to Rayleigh’s method, Eq. (1) can be expressed as:

X = K(X1a  Xb2  Xc3  − − − − − − Xnn )

➢ exponents (a, b, c, - - -,n) are constants and are evaluated using the principle of
dimensional homogeneity.

❖ K is a dimensionless constant - determined either from the physical

characteristics of the problem or from experiments.

❑ To explain this method, consider the following problems.

Problem 1: Find an expression for drag force/resistance FD on a smooth sphere of

diameter D moving with constant velocity Uo in a stationary fluid of viscosity .

Solution: FD = f (D, Uo, ) (1)

• According to Rayleigh’s method, Eq. (1) may be expressed as:

FD = K (Da  Uob  c )
➢ Exponents (a, b, c, - - -,n) are constants and are to be evaluated using the principle
of dimensional homogeneity
• Writing dimensions of each variable, to write

 [MLT -2 ] = [L] a [LT -1 ]b [ML-1T -1 ]c

❖ K is a dimensionless constant.

• Equating powers of M, L and T, to get

➢ Powers of M, 1 = c

➢ Powers of L, 1 = (a + b - c)

➢ Powers of T, - 2 = - b – c

❖ On solving, a = 1, b = 1 and c =1

 FD = K (D1Uo11)

• In this problem, K = 3

FD = 3 UoD

❖ Stoke’s Law

5
Problem 2: Find an expression for drag force FD on a smooth sphere of diameter D
moving with a constant velocity Uo in a fluid of density  and viscosity .
Solution: The additional variable is 

 FD = f (D, Uo, , ), can be written as

FD = K (Da  Uob  c   d )
 [MLT -2 ] = [L]a [LT -1 ]b [ML-3 ]c [ML-1T -1 ]d

❖ Equating powers of M, L and T, to get

• Powers of M, 1 = c + d
• Powers of L, 1 = (a + b - 3c - d)
• Powers of T, -2 = - b - d
❖ There are four unknowns a, b, c and d but equations are three. Hence, it is not
possible to find the values of a, b, c and d.

❖ Any three variables can be expressed in terms of fourth variable, which is the most
significant in the problem.

❖ In this problem, viscosity is the most significant variable affecting the given
phenomenon.

Expressinga,b and c in terms of d (power to viscosity), to get

❖ c = (1 – d), b = (2 – d), a = (2 – d)

 FD = K (D2- d  U2- d  1− d  d )

 2 2   d 
= K D U   
  UD  
• can be further expressed as:
  
FD = Uo2D2 

 o 
U D
• Required expression.

6
Buckingham’s method
• Rayleigh’s method becomes increasingly tedious when more than four
independent variables are involved in the phenomenon.
• Rayleigh’s method is best suited when the numbers of independent variables are
three.
• When four or more than four independent variables are involved in the
phenomenon, then it is more convenient to use Buckingham’s method,
• According to this method, “if there are n numbers of variables (independent and
dependent) and if these variables are described by m fundamental dimensions, then the
variables can be expressed into (n-m) independent dimensionless terms, each term is
called a -term”.
• Each -term contains (m+1) variables, m (number of fundamental dimensions)
are also known as repeating variables.
❖ Each -term is solved using the principle of dimensional homogeneity.
❖ Important step in this method is the selection of repeating variables as discussed below:

Method of selecting repeating variables

• Following points should be kept in mind while selecting repeating variables:

(i) Generally dependent variable is not selected as a repeating variable.

(ii) No two repeating variables should have the same dimensions.

(iii) Repeating variables should be chosen in such a manner that;

(a) First variable should represent the geometric property like length, diameter, height etc.

(b) Second variable should represent a flow property like velocity, acceleration, discharge etc.

(c) Third variable should represent a characteristic fluidproperty like mass density or viscosity.

❖ Viscosity is important only in case of viscous or laminar flow problems otherwise

density is chosen as the repeating variable.

To explain this method, let us consider the previous problem of finding drag force on a
sphere:

 FD = f (D, Uo, , ) (1)

Or f1(FD, D, Uo, , ) = 0 (2)

• Writing dimensions of each variable as

7
 FD = [MLT -2 ], D = [L], Uo = [LT -1 ],  = [ML-3 ],  = [ML-1T -1 ]
No. of fundamental dimensions involved, m = 3
• Total number of variables = 5
 Number of  -terms = (5 - 3) = 2
• Equation (2) can be written as: f1 (1,  2 ) = 0 (3)
❖ Each -term contains (m+1) variables, m is the number of fundamental
dimensions and also called repeating variables.
❖ Taking D, Uo and  as repeating variables.
• Each  -term can be written as:
a b c
1 = D 1  Uo1   1  FD (4)
b2
a2 c2
2 = D  Uo    (5)
• exponents ai, bi and ci are constants, determined using the principle of
dimensional homogeneity i.e. each - term is solved using the principle of
dimensional homogeneity.
First  - term
• Writing dimensions of each variable on both sides, to get
[M0L0 T 0 ] = [L] a1 [LT -1 ]b1 [ML-3 ]c1 [MLT -2 ]
• Equating powers of M, L and T on both sides, to get
➢ Powers of M, 0 = C1 + 1
➢ Powers of L, 0 = a1 + b1 -3c1 + 1
➢ Powers of T, 0 = - b1 - 2
❖ On solving, a1 = -2, b1 = -2 and c1 = -1
• Substituting a1, b1 and c1in Eq. (4), to write
FD
1 = D2  Uo− 2  −1  FD  1 =
U2oD2
• Similarly, solving second  - term, to get

• a2 = -1, b2 = -1, c2 = -1

 2 = D −1  Uo−1   −1    2 =
UoD
• Substituting the values of 1 and 2in Eq. (3), to get

 F  
f1  2D 2 , 8  = 0
 UoD UoD 
 • can be expressed as:

FD      
=     FD = U2oD2  
UoD
2 2
 UoD   UoD 

• Required expression (same).

❖ To reduce the calculations and to obtain an expression in the desired form, the
following characteristics of -terms are used:

CHARACTERISTICS OF -TERMS

(i) If a variable is dimensionless it can be straightway assigned as one of the -term.

(ii) If any two variables have the same dimensions, then their ratio will be one of the -terms.
Example: If in a problem L and B are involved, then one of the -terms will be L/B or B/L.
(iii) Reciprocal of a -term, its square and its square root is non-dimensional i.e. its
character will not change (it would remain dimensionless). Thus, any -term may be

replaced by 1/ ,  2 , 

(iii) Product and division of any two -terms is non-dimensional. Thus, any two -terms
say 1 and 2 may be replaced by  1   2 ,  1 /  2 or  2 /  1

(iv) Multiplication, division, addition or subtraction of any -term by any numerical

constant does not change the character of -term. Thus, for example any -term may be
replaced by 3,  / 3, ( + 3) or ( − 3).

Illustration: To obtain the expression for drag force on a sphere in the standard from:

  
The equation is: FD = U2oD 2   
 UoD 

 1  UoD
 FD = Uo D 2 
2
; Re =
 Re  
 1 
• We know that drag coefficient is a function of Reynolds number i.e.   can be
 Re 
replaced by CD.

2FD
• Also, first  term i.e. 1 can be written as: (Using the characteristics of -terms)
 2 2
U D
4 o

9
 Uo 
2
 FD = CD A , A = D2
2 4

• Required expression.

MODELS STUDY
• A model is a small-scale replica of the actual structure or machine.

• Small-scale replica of the actual structure is known as model while the actual
structure is called prototype.

• Models are generally smaller than the prototypes, but in some cases the models
may be larger than the prototype.

• For example, performance of a carburetor and a wrist watch may require to be

tested on a large scale model.

❖ Before any new project is undertaken for the first time, it is often desirable and
advantageous to conduct model studies.

Examples:

• Testing of models of building, rivers, spillways, objects moving through air

(designing streamline shapes of objects) etc.

• A model study involves:

➢ Model design and fabrication

➢ Tests on model and analysis of test results

❖ For model design, similarities are to be established between the model and the
prototype and for analysis of test results, model laws are applied.

SIMILITUDE (TYPES OF SIMILARITIES)

• Results obtained from model tests are applicable to prototype if there exist
similitude between the model and the prototype.
• For this purpose, three types of similarities are to be established between the
model and prototype [Link], Kinematic and Dynamic similarity
Geometric Similarity:
• Geometric similarity between the model and prototype implies similarity in shapes
[Link] of all the corresponding dimensions in model and prototype should be equal.
❖ Such a ratio is known as scale ratio.

10
 • On geometric similarity basis, different scale ratios are linear scale ratio, area scale
ratio and volume scale ratio.
Lm Bm Hm
• Linear scale ratio, Lr = = =
Lp Bp Hp
➢ Subscripts m and p corresponds to model and prototype, respectively.

Area scale ratio, Ar = L2r

Volume scale ratio, r = Lr

3

Kinematic Similarity
➢ is the similarity of motion i.e. both the systems should undergo same rates of
change of motion.
• On kinematic similarity basis, we can have velocity scale ratio, acceleration scale
ratio and discharge scale ratio etc.
• Thus, velocity scale ratio Vr can be expressed as:
( V1 )m ( V2 )m
Vr = =
( V1 )p ( V2 )p

➢ (V1)m is the velocity at point 1 in the model , (V1)p is the velocity at the corresponding
point in the prototype, (V2)m and (V2)p are the corresponding values at point 2.

❖ The corresponding points in the two systems are called homologous points.

➢ Velocity scale ratio can also be expressed as:

Vm (L m / Tm ) L r T
Vr = = = , Tr is the time scale ratio = m
Vp (L p / Tp ) Tr Tp

• Similarly, acceleration scale ratio, ar and discharge scale ratio Qr can be written as:

Dynamic Similarity

• is the similarity of forces i.e. ratio of all the forces acting at homologous points in the
two systems should be equal.
❖ Force scale ratio Fr on the basis of dynamic similarity is given by
(Fi )m (Fv )m (Fg )m
Fr = = =
(Fi )p (Fv )p (Fg )p

11
 • Fi, Fv and Fg are the inertia, viscous and gravity forces, respectively.
❖ When the two systems are geometrically, kinematically and dynamically similar, they are
said to be completely similar or complete similitude exists between the two systems.
❖ If the two systems are only dynamically similar, they are also said to be completely
similar.

DIMENSIONLESS NUMBERS
• Forces acting on a moving mass of fluid are: Inertia force (Fi), viscous force (Fv),
Gravity force (Fg), pressure force (Fp), surface tension force (Fs) and elastic force (Fe).

❖ Dimensionless numbersare obtained by dividing the inertia force with any one of
the remaining five forces listed above. Accordingly, there are five dimensionless
numbers i.e. Reynolds no. (Re), Froude no. (Fr), Euler no. (Eu), Weber no. (We),
Mach no. (Ma)

❖ Before determining expression for dimensionless numbers, let us first find

expressions for these six forces:

• Inertia force = Mass  Acceleration of the flowing fluid

2
3
• can be expressed as: Fi =   a = L  2  = L2   = L2 V 2
L L
T  T

➢ L is the characteristic dimension, V is the velocity,  is the dynamic viscosity

• Viscous force, Fv = 
du V
 Surface area =  L2 = VL
dy L

Gravity force, Fg = Mass  Acc. due to gravity = =   g = L g

3
•

• Pressure force, Fp = Intensity of pressure  Area = pL2

• Surface tension force, Fs = L,  is the surface tension

• Elastic force, Fe = Bulk modulus of elasticity  Area = KL2

Reynolds number (Re)

• ratio of inertia force of a flowing fluid to the viscous force of fluid i.e. Fi/Fv
V 2L2 VL VL
Re = = =
VL  
Froude number (Fr)
• ratio of square root of inertia force to the gravity force (Fg) of fluid

12
 Fi V 2L2 V
 Fr = = =
Fg L3 g gL

Euler’s number (Eu)

• ratio of square root of inertia force to pressure force (Fp)
Fi V 2L2 V
 Eu = = =
p/
2
Fp pL

Weber’s number (We)

• ratio of square root of inertia force to surface tension force (Fs)

Fi V 2L2 V
 We = = =
Fs L ( / L )

Mach number (Ma)

• ratio of square root of inertia force to the elastic force Fe

Fi V 2L2 V V K
Ma = = = = ; C=
Fe KL 2
K/ C 

❖ V is the fluid velocity; C is the velocity of sound through that fluid medium whoseK
and  are considered.

• The ratio of fluid velocity to the velocity of sound in the fluid medium is also
known as Mach number.
➢ When V > C i.e. Ma > 1 flow is designated as supersonic.
➢ When V = C i.e. Ma = 1, flow is designated as sonic and
➢ When V >>C i.e. Ma >> 1, flow is hypersonic.

❖ If Ma < 1, the flow can be considered as incompressible (V < C) and If Ma ≥ 1, flow

is compressible (V  C).

MODEL LAWS
• For most of the hydraulic problems of model studies, it is quite rare that all the
forces are simultaneously predominant/ significant in the given problem.

• Generally, only one force in addition to inertia force is relatively more significant than the
rest of forces, which may not either exist or may be of negligible magnitude.

❖ Various model laws have been developed depending upon the significant forces
influencing the different flow situations.
13
 ❖ There are five model laws viz. Reynolds model law, Froude, Euler, Weber, Mach-
model law

Reynolds Model Law: This law is applicable to flows which are influenced by viscous
effects or used when there is dominant action of viscous forces. Examples: Boundary
layer flows (Flow- through pipes, meters, over surfaces; completely submerged flows)
etc.
Froude Model Law: applicable to flows that are influenced by gravity. Examples: Free
surface flows (flow in channels, rivers, flow over spillways, notches, weirs etc.)
Euler Model Law: applicable to flows in which pressure drops are significant Examples:
Flow through pipes, cavitation phenomenon etc.
Weber Model Law: applicable to those flow situations in which surface tension
influences the flow Examples: Flow with interface (capillary rise), flow over notches and weirs with
small heads etc.

Mach Model Law: applicable to flows where compressibility of fluid is significant.

Examples: Aerodynamic testing, water hammer problems etc.

❖ In some problems, only one model law may be applicable where as in some other
cases, more than one model law may be applicable.

Reynolds Model Law

❖ According to Reynolds model law, similarity of flow in model and prototype can
be established if Reynolds number of model is equal to Reynolds number of
prototype i.e. (R e )m = (R e )p

 VL   VL  r Vr Lr Vr Lr
   =   OR = 1 or =1
  m   p r r

➢ quantities with subscript r represent the corresponding scale ratios.

Scale ratios for some physical quantities based on RML(in terms of r, r and Lr)

r
• Velocity scale ratio, Vr =
Lr
Discharge ratio, Qr = Vr Lr =  r Lr
2
•
 L2 V 2  r  r2
Intensity of pressure ratio, pr = 
Force 
•  =   =  =
2
V
 Area r  L r
2 r r
L2r
• Work/energy ratio, Er = (Force  dis tan ce)r = L V ( 3 2
) = (  L )
r r
2
r r

14
 • Power ratio, Pr = (Force  velocity)r = L V ( 2 3
)
r =
r  r3
Lr
• Or Power = Work/Time

Froude Model Law

❖ According to Froude model law, (Fr )m = (Fr )p

 V     Vr 
   =  V  or   = 1 or Vr = L r gr
    
 Lg  m  Lg  p  L r gr 

❖ If gr = 1 (g is same for both model and prototype), then Vr = Lr

Scale ratios for some physical quantities based on FML

Lr L
• Time scale ratio, Tr = = r = Lr
Vr Lr
Lr
• Acceleration scale ratio, ar = Vr = =1
Tr Lr
L3r L3
• Discharge scale ratio, Qr = = r = L2r.5
Tr Lr
or Qr = Ar Vr = Lr 2 Lr = Lr 2.5
Force scale ratio, Fr = r Lr Vr = Lr Lr = Lr
2 2 2 3
• (when r = 1 i.e. m =p)

• Intensity of pressure scale ratio, pr = r Vr2 = Lr when r = 1.

• Work, energy, torque or moment scale ratio, Wr = L3rLr = L4r
L4r L4
• Power scale ratio, Pr = = r = L3r.5
Tr Lr
Euler Model Law
❖ According to Euler model law, (Eu )m = (Eu )p

 V    Vr
  = V  or =1 or Vr = pr (when r = 1)
 p/   
 m  p /  p pr / r

Weber Model Law

❖ According to Weber model law, (We )m = (We )p

15
  V   V  Vr
   =
 

 or =1
 ( / L )  m  ( / L ) p ( r /  r L r )

Mach Model Law

❖ According to Mach model law, (Ma )m = (Ma )p

 V    Vr
  =  V  or =1
 K/   
 m  K /  p K r / r
Problems:

Q1: A pipe of diameter 1.5 m is required to transport oil of relative density 0.90 and
viscosity 0.03 Stoke at the rate of 3 m 3/s. If a 150 mm diameter pipe with water at 20 oC
(viscosity = 0.01 Stoke) is used to model the above flow, find velocity and discharge in the
model.

Solution:

• Here, Reynolds model law is applicable.

 VL   VL 
  = 
   m   p

 VD 
• In Reynolds's no., L = D   r r  = 1
 r 

16
Q2: A 1:50 spillway model has a discharge of 1.25 m 3/s. What is the corresponding
prototype discharge? If a flood takes 12 h to occur in the prototype, how long will it take
to occur in the model?

Solution:

CLASSIFICATION OF MODELS
• Two categories - undistorted and distorted models.

Undistorted Models: are those which are geometrically similar to their prototypes i.e.
scale ratio for the corresponding linear dimensions of model and prototype is same.
Distorted Models: are not geometrically similar to their prototypes. For a distorted
model, different scale ratios for linear dimensions are adopted.
Example: For river models, two different scale ratios are taken, one for horizontal and
the other for vertical dimensions (depth).
❖ If same scale ratio is taken, then depth of water in the model will be very-very
small which may not be measured accurately.

Scale ratios for distorted models

17
 L m Bm
• Scale ratio for horizontal dimensions, (L r )H = =
Lp Bp
Hm
• Scale ratio for vertical dimensions, (L r ) V =
Hp
❖ Velocity scale ratio Vr, area ratio Ar and discharge ratio Qr for distorted models:

Vm 2gmHm
Vr = = = (Lr ) V
Vp 2gpHp
A m Bm  Hm
Ar = = = (Lr )H  (Lr ) V
Ap Bp  Hp
Q m Am  Vm
Qr = = = (Lr )H  (Lr ) V  (Lr ) V = (Lr )H  (Lr )1V.5
Qp Ap  Vp

Scale effect in models

• If similitude does not exist between a model and its prototype, then there will be
some discrepancy between the results obtained from the model tests and those
which will be indicated by the prototype after its construction. This discrepancy
is known as scale effect.
Reasons for scale effect
(i) In some problems, if several forces have predominance, then similitude will be
ensured only if all the pertinent forces are simultaneously satisfied. However,
it is quite difficult to satisfy all the model laws involved in the phenomenon.
For example, if in particular problem, both viscous and gravity forces are predominant, then both
Reynolds and Froude model laws should be satisfied simultaneously.
• The necessary condition by combining both the laws is:

Vr L r
= 1; V r = Lr  r = L3r / 2 (1)
r
• Equation (1) indicates that once model scale is selected, a liquid of appropriate viscosity
as required Eq. (1) must be found for the model testing or if a particular liquid is to be
used for the model testing, then the model scale Lr has to be found in accordance with Eq.
(1).

• Further, for model testing if same liquid as for the prototype is used then Lr =1 i.e. model
must be as large as the prototype.

❖ Difficult to find liquids of any desired viscosity and also, models with full scale are
impractical.

18
 (ii) Scale effect may also be developed in cases where forces which have practically no
effect on the prototype significantly affect the behaviour of its model.

Example: For flow over weir models, if head is less than about 15 mm, surface tension plays
a significant part than in the full size weir having large head than in the model.

Application of model laws to some specific model studies:

• Submerged objects
• Partially submerged objects
• Models of rivers

(1) Submerged objects

• Consider the case of an aeroplane model testing in a wind tunnel, to find the drag
(resistance to its motion).
❖ In this case, only viscous and inertia forces are involved.
• Gravity forces may be neglected (fluid being air has very small density)
• Surface tension is not present as there is no interface between a liquid and another liquid.
• If it is assumed that it is a low speed aeroplane (velocity of air in the wind tunnel is small),
then the effects of compressibility can also be neglected.

•  Drag encountered by the aeroplane model is affected by the viscous force only.

Drag FD will thus be a function velocity Uo, density  and viscosity  and some characteristic
length L (to specify the size of model) i.e. FD = f (Uo, ,, L)

• Using the method of dimensional analysis, the following dimensionless functional

relationship is obtained.
FD  U L 
=   o  = (Re)
UoL   
2 2

• Eq. is true for both model and prototype.

• For similitude, the ratio of forces at homologous points must be same for both.

 Reynolds no. must be same for both.

19
 • Consequently, function (Re) is same for both model and prototype.
FD
Dimensionless no. is same for both model and prototype i.e.
U2oL2    
F F
  2D 2  =  2D 2 
 UoL  m  UoL p

❖ This eq. is valid only if tests on the model are carried out under the conditions that only
Reynolds no. is same for the prototype and model.

Observation:

• If compressibility effects are also important for the prototype, then Mach no. must also be
same in addition to Reynolds numbers for similitude between the model and prototype.
Thus,

• To satisfy both the conditions simultaneously, we have

❖ As the range of available fluids being limited, it is not possible to satisfy the above
condition.

• In such cases, one of the forces is considered as having secondary effect and may be
neglected. This would however result in introducing the scale effect which needs to be
accounted for before making any predictions about prototype behaviour.

(2) Partially Submerged objects

• Consider the case of a ship moving through water, to find the drag

❖ Drag in this case is due to three causes/reasons: Viscosity, surface waves and wake
formation.

• Due to velocity gradient and viscosity of fluid, viscous forces are set up between the fluid
layers close to the surface of ship, resulting in causing frictional resistance to the motion
of ship, known as skin frictional drag.

• Ship being only partly immersed in water, its motion gives rise to waves on the surface,
resulting in imparting resistance to its motion.

20
 • As ship moves, a part of flow towards the rear of ship separates from its surface so as to
form a ‘wake’. Eddies so formed considerably modify the pressure distribution around the
moving ship, resulting in developing resistance to its motion, known as pressure or form
drag.

❖ Total drag on a ship moving through water is the sum of friction drag, drag due to
surface waves and pressure drag.

• Since ships are generally streamlined, pressure drag is relatively small and the formation
of surface waves is associated with gravitational action.

• (For this case, compressibility of water is insignificant and surface tension forces may be
neglected if the model is not too small).

❖ Thus, resistance to the motion of ship is affected by viscosity as well as gravity.

Total drag FD encountered by the ship will thus be a function of velocity Uo of the ship, viscosity
of water, density of water , characteristic length L to specify the size of ship and acceleration
due to gravity g i.e. FD = f (Uo, ,, L, g)

• Using the method of dimensional analysis, the following dimensionless functional

relationship is obtained.

 UoL U2o 
FD
=  
 ,  (
 =  Re,Fr 2 )
UoL   gL 
2 2

• Thus, for complete similarity between the prototype and its model, Re and Fr numbers
must be same i.e.
 V   V 
 VL   VL    = 
  =  and  gL   gL 
   m   p  m  p

• In practice, gm = gp
For both the conditions to be satisfied simultaneously, we have

(3) Models of rivers

• River models are frequently used to predict the behaviour of rivers when some works are
undertaken across the river.

❖ Froude’s model law is the basic similitude criteria. Surface roughness should also be given
due considerations.

• River models may have either vertical or slope enhancement/enlargement.

21
 • In vertically enlarged models, depth scale is enhanced (distorted) and in slope
enhanced model, model is geometrically similar (undistorted) but is made
relatively steeper by titling it.

❖ Surface roughness for river models is calculated using Manning’s Eq. i.e.
1
Vr = (R h )r2 / 3 S 1r/ 2
Nr
Vertically enlarged river models (distorted):
• For these models, as depth scale is enlarged, let (Lr)Hand(Lr)V represent the horizontal
and vertical length (depth) scale ratios, respectively.
❖ Hydraulic radius is dependent on both horizontal and vertical dimensions, so it cannot be
uniquely represented by a single scale ratio.
❖ For rivers, it can be assumed that , (R h ) r  (L r ) V

• Also, slope scale ratio, S = (L r ) V

r
(L r )H

1 (L r )v
7/6
 Velocity scale ratio, Vr =
Nr (L r )1H/ 2

• According to Froude law,

As Vr = (Lr )1v/ 2
(Lr )2v / 3
Scale ratio for rugosity coefficient,  Nr =
(Lr )1H/ 2
• By suitably fixing the scale ratios (Lr)H and (Lr)V , the value of Nr can be obtained.

Observation: For undistorted models, since (L r )H = (L r )V  Nr = L1r/ 6

Slope enlarged river models:
Models are geometrically similar (undistorted) but slope of models is increased by titling them
[Link] models are therefore also known as tilted models.
• For such models, 1 1 2 / 3 1/ 2
Vr = (R h )r2 / 3 S1r / 2  Lr = Lr Sr
Nr Nr

 Nr = L1r/ 6S1r/ 2
• By suitably selecting the slope and length scale ratios, the corresponding value of Nr can
be obtained.

Problem: To estimate the frictional head loss in a pipeline of diameter 1 m through which oil of
specific gravity 0.80 and viscosity 0.02 P is required to be transported at the rate of 2000 litres

22
per sec, a model test was conducted. The test was conducted on a 100 mm diameter pipe using
water at 20oC. Determine (i) Discharge required for the model pipe. (ii) Head loss in the
prototype, if head loss in 30 m length of model pipe is 440 cm of water. Given, viscosity of water
at 20oC = 1 cP.

Solution: The data given in the problem can be written in the following manner:

R
• Here, resistance to flow, R is governed by = (Re )
V 2L2
• For similitude, function (Re) must be same for both model and prototype i.e. Reynolds
no. must be same for both model and prototype.
R
• Also, dimensionless no. will be same for both model and prototype.
V 2L2
 First applying Reynolds model law, to write

23
24