CEE 4361

Fluid
Mechanics

Dimensional Analysis and

Similitude
INTRODUCTION
• Most of the problems in fluid mechanics can only be investigated
experimentally.
• Only some problems are available than can be solved analytically.
• Students of hydraulic engineering must know how to plan these
experiments.
• After conducting the experiment, we must correlate the experiment in a
real life specific problem. Example: simulation of flume flow with real
life river flow.
• Most important thing is we must make these experiments widely
applicable. Widely applicable means not only the experiments will be
applicable for the conditions assumed, but also for the other scenarios.
SIMILITUDE
• One way to make the experiments applicable for various types of
scenarios is called similitude or Similarity.
Similarity laws enables us to make experiments with a convenient fluid
such as water or air, for example, and then apply the results to a fluid which
is less convenient to work with such as hydrogen, steam, or oil. Also in both
hydraulics and aeronautics, valuable results can be obtained at a minimum
cost by test made with small scale models of the full size apparatus. The
laws of similitude make it possible to predict the performance of the
prototype, which means the full size device from test made with the model
• Why simulation is needed?
• Normally, lab experiments are performed carefully in controlled
conditions. However, in natural water flow like rivers, controlled condition
can not be provided.
• Therefore, similitude is a way to make lab experiment results simulate real
life outcomes.
EXAMPLE PROBLEM
Pipe flow example: Pressure drop per unit length

The pressure drop per unit length that develops along a pipe as the result of
friction cannot be determined analytically without the use of experimental
data.
First we determine the variables in the flow that are related to pressure drop:

ΔP/l = f (D, ρ, µ, V)

ΔP/l = Pressure drop per unit length

D = Diameter of the pipe
ρ = density of fluid
V = Velocity of
fluid
µ = viscosity of fluid
EXAMPLE PROBLEM
ΔP/l = f (D, ρ, µ, V)
• How do we approach this problem?
• If we do experiments to prove this equation, how can we do these
experiments?
• We will vary only one variable, while keeping the other three variables
constant.
Now we have done five experiments for each plot with
other variables held constant (total 20 experiments)
What we have gained?
Our analysis is very narrow and specific, not widely
applicable.

Now, what if we do 10 experiments for each variable,

and let other three variables vary for 10 values.
Total combination : 10 x 10 x 10 x 10
10,000 experiments!!
DIMENSIONLESS ANALYSIS
Fortunately, there is a simpler approach: dimensionless groups
The original list of variables can be collected into two dimensionless groups.

ρV2
( )
ΔP/l = φ ρVD
µ

Now, instead of working with 5 variables, there are only two.

The experiment would consist of varying the independent variable and
determining the dependent variable which is related to pressure drop.
DIMENSIONLESS ANALYSIS
Dimensions are Mass (M), Length (L), Time (T) and Force (F or MLT-2) Then we
check our dimensionless groups
V = LT-1
µ = FL-2T Substituting, we see no dimensions of our two variables.
ΔP/ꙇ = FL-3 -
3
ΔP/ꙇ
D=L ρV2 =
-
4 2
-
1 2
= F0L0T0
ρ = FL-4T2 - -
ρVD = (FL 4 T2 )(LT 1 )(L) = F0L0T0
-
µ (FL 2 T)

Not only have we reduced the number of variables from five to two, but
the dimensionless plot is independent of the system of units based.
DIMENSIONLESS ANALYSIS
The use of dimensional analysis can be summarized as follows:

• It helps to find whether an equation of any flow phenomenon is rational or not. The
equation is called rational if it is dimensionally homogeneous. This is to say that
dimensional analysis can be applied only when a phenomenon can be expressed by
a dimensionally homogeneous equation.
• The relationship between various physical quantities in an equation, governing a
particular flow phenomenon, can be known.
• It helps the reduction of a number of variables involved in a flow phenomenon
with which the performance of experiments becomes easy.
• It helps in making suitable smaller-sized models in which experiments can be
performed to predict the performance of prototypes.
• Dimensional analysis, on the whole, facilitates planning of reliable scientific
experiments.
FUNDAMENTAL
AND DERIVED
QUANTITIES
DIMENSIONAL HOMOGENEITY
An equation is called dimensionally homogeneous if the dimensions have
identical powers on both sides. Such an equation is also known as a rational
equation. A dimensionally homogeneous equation would essentially be
independent of the system of units (i.e. SI, Metric or English).
Let us consider the most common equation of hydraulics
Q = AV
Dimensionally,
L3T-1 = L2 x LT-1 =
L3T-1
So, both sides of the equation have the same dimensions. Therefore, the
equation is dimensionally homogeneous or rational.
DIMENSIONAL HOMOGENEITY
The principle of dimensional homogeneity can be used
(i) to determine the dimension of a physical quantity,
(ii) to check the dimensional homogeneity of an equation, and
(iii) to change the coefficient of an equation while using it in other systems
of units.

Methods of dimensional analysis:

• Rayleigh's method
• Buckingham's method, generally known as Buckingham's π-theorem
RAYLEIGH'S METHOD
This method was originally proposed by Lord Rayleigh in 1899. In
this method, the dependent variable is expressed as a function of the
exponents (or powers) of the independent variables. The dependent variable
is one for which information is required, whereas the independent variables
are those which govern the variation of the dependent variable. If y is the
dependent variable and x1, x2, x3,… are the independent variables involved
in a physical process, then the functional relation may be written as

y = f(x a, x b, x c,…..)
1 2 3

where a, b, c, … are the exponents or powers of x1, x2, x3,… ,

respectively The value of a, b, c, ..... are obtained using the principle of
dimensional homogeneity.
BUCKINGHAM'S Π-THEOREM
This method was originally proposed by Edgar Buckingham in 1915.
"If there are n variables in a dimensionally homogeneous equation, and if these
variables contain m fundamental dimensions, they may he grouped into n-m non-
dimensional independent parameters."
Mathematically, if a variable x1 depends upon a number of independent variables x2,
x3, x4,…,xn, the functional equation may be written as
x1 = φ(x2, x3, x4,…,xn)
where φ is a function. This equation may be written in its general form as
f(x1, x2, x3, x4,…,xn)= 0, where f is another function.
If Π1, Π2, Π3,… represent dimensionless groupings of the variables x1, x2, x3,
x4,…,xn with m dimensions involved, then according to Buckingham's π-
theorem, an equation the form
F(Π1, Π2, Π3,…, Πn-m) = 0 exists, where F is still another function.
BUCKINGHAM'S Π-THEOREM
Systematic Approach: Example – Pipe Flow

The steps of the Buckingham's method are as follows:

Step 1. List all the variables that are involved in the problem.
ΔP/ꙇ = f (D, ρ, µ, V)

Step 2. Express each of the variables in terms of basic dimensions.

V = LT-1
µ = FL-2T
ΔP/ꙇ = FL-3
D=L
ρ = FL-4T2
BUCKINGHAM'S Π-THEOREM
Step 3. Determine the required number of pi-terms.
Number of variables, n = 5 (ΔP/l , D, ρ, µ, V)
Number of dimensions, m = 3 (F, L, T)
 Number of pi-terms = n – m = 5 – 3 = 2

Step 4. Select a number of repeating variables, where the number required is

equal to the number of reference dimensions, m. Generally, the repeating
variables selected are D, ρ and V, i.e. the first repeating variable will
represent the geometric property, the second will represent the fluid
property and the third will represent the external effects or flow
characteristics.
BUCKINGHAM'S Π-THEOREM
Step 4 (Continued). For pipe flow problem,
Number of repeating variables = m = 3
We choose three independent variables as repeating variables. Note that,
there can be more than one set of repeating variables.
Repeating variables: D, V, ρ
These repeating variables should be dimensionally independent; you can
not form a dimensionless group with them alone.
You should not choose “µ and ρ” or “µ and ν” or “ρ and ν” at the same time,
since, µ = ρν
BUCKINGHAM'S Π-THEOREM
Step 5. Form a pi-term by multiplying one of the non-repeating variables by
the product of repeating variables, each raised to an exponent that will
make the combination dimensionless. The first group chosen usually
includes the dependent variables.
Here, non-repeating variables are ΔP/l and µ. So, we first form pi-term for
ΔP/l and then we will form pi-term for µ.
Π1 = ΔP/l DaVbρc
To make the product dimensionless, we need to solve for the exponent
values.
(FL-3)(L)a(LT-1)b(FL-4T2)c = F0L0T0
BUCKINGHAM'S Π-THEOREM
Step 5 (Continued).
1 + c = 0 (for F)
- 3 + a + b – 4c = 0 (for L)
- b + 2c = 0 (for T)

Solving the set of algebraic equation, we obtain: a=1, b=-2, c=-1

ΔP/l
Π1 =
ρV2

µ is a remaining non-repeating variable, so we can form another group:

ρVD
Π2 = µ
BUCKINGHAM'S Π-THEOREM
Step 6. Repeat Step 5 for each of the remaining non-repeating variables.
Note that, we could have chosen D, V and µ as another set of repeating
variables.

Step 7. Check all the pi-terms to make sure they are dimensionless.
- 3
ΔP/ꙇ (L)(FL )
Π1 = = - 4 - 1 2 = F 0L 0T 0
ρV2 2
FL T LT
- 4 - 1
2
Π2 = ρVD =
(FL T )(LT )(L)
- 2 = F0L0T 0
µ (FL T)
BUCKINGHAM'S Π-THEOREM
Step 6. Express the final form as relationship among the pi-terms and think
about what it means.
Π1 = φ(Π2, Π3,…, Πn-m)

ΔP/ꙇ = φ(ρVD
For our case, ρV2 µ )

Pressure drop depends on Reynolds number.

BUCKINGHAM'S Π-THEOREM
Example 7.3 The pressure drop ΔP in a pipe depends upon the mean
velocity of flow (V), length of pipe (l), diameter of pipe (D), viscosity
of fluid (µ), average height of roughness element (k) and mass density
(ρ).
(a) By using Buckingham’s pi-theorem, obtain a
dimensionless expression for ΔP.

4flV 2
(b) Show that, hf =
2gD
MODEL TESTING
Since the beginning of the twentieth century, the engineers have
started a new scientific method to predict the performance of their
structures and machines before they made and installed. This is done
by preparing models and testing them in the laboratory, so as to form
some opinions about the working and behavior of the proposed
structures, after the completion or actual installation. The structure, of
which the model is prepared, is known prototype and the model is
known as the physical model or scale model or simply model.
MODEL TESTING
The principal advantages of model testing are:
I. The behavior and working details of a hydraulic structure or machine can
be easily predicted by performing experiments on its model.
2. If the hydraulic structure or machine is made directly, then in case of failure, it
is very difficult to change its design. Moreover, it is very costly to build the
same. Thus, making a model and testing it in the laboratory results in the saving of
human labor and material.
3. With the help of model testing, a number of alternative designs can be studied.
Finally the most economical, accurate and safe design may be selected.
4. In case when an existing structure is not functioning properly, its defects can
be detected and rectified by model testing.
APPLICATION OF MODEL TESTING
(a) Water power engineering
(i) Turbines and pumps
(ii) Hydraulic structures like dams, weirs, spillways, canal falls, etc.
(b) Irrigation engineering
(i) Flood control
(ii)River training
(iii)Investigation of silting-scouring in rivers, irrigation canals, etc.
(c) Design of ships, harbors, etc.
HYDRAULIC SIMILARITY OR SIMILITUDE
The similarity between the prototype and its model is known as
hydraulic similarity or hydraulic similitude.
There are three kinds of similarities which a model should possess in
order that it may reproduce the behavior of its prototype:
(i) geometric similarity,
(ii)kinematic similarity, and
(iii)dynamic similarity.
SIMILITUDE: GEOMETRIC SIMILARITY
The geometric similarity is said to exist between the model and the prototype if
the ratios of all the corresponding linear dimensions are equal. The scale ratio
(Lr) is the ratio of the linear dimension of the model to that of the prototype.
Let Lp, Bp and Dp are the length, width and depth of the prototype and Lm, Bm
and Dm are the corresponding values for the model. If the geometric similarity
exists between the model and the prototype, then,
SIMILITUDE: KINEMATIC SIMILARITY
The kinematic similarity is said to exist between the model and the
prototype if the ratio of the corresponding velocities, accelerations and
direction of flow of the fluid particles at the corresponding points in
the model and the prototype is the same.
Let V1p and V2p are the velocities of the fluid particles in the prototype
at points 1 and 2, respectively, and V1m and V2m are the corresponding
values for the model. Now, if kinematic similarity exists between the
model and the prototype, then the velocity ratio (Vr) of the model and
the prototype,
V1m V2m V3m
Vr = = = =…
V1p V2p V3p
SIMILITUDE: DYNAMIC SIMILARITY
The dynamic similarity is said to exist between the model and the
prototype if the ratios of the corresponding forces acting at the
corresponding points in the model and the prototype are equal.
Let F1p and F2p are the forces acting in the prototype at points 1 and 2,
respectively, and F1m and F2m are the corresponding values for the
model. Now, if the dynamic similarity exists between the model and
the prototype, then the Force ratio (Fr) of the model and the prototype

Fr = 1m = 2m = 3m =…
1p 2p 3p
DIMENSIONLESS NUMBERS

(i) Reynolds number,

(ii)Froude number,
(iii)Weber number,
(iv)Euler number, and
(v) Mach number
REYNOLDS NUMBER
REYNOLDS NUMBER
FROUDE NUMBER
DIMENSIONLESS NUMBERS
Weber number

Euler number
MACH NUMBER
TYPES OF MODELS
Undistorted Model: A model is said to be undistorted when the
horizontal and vertical scale ratios of model and prototype are same. An
undistorted model is geometrically similar to its prototype. The
prediction of an undistorted model is comparatively easy and the model
results can be easily transferred to the prototype.
Distorted Model: A model is said to be distorted when the vertical and
horizontal scale ratios are different. The distorted model does not have
true geometric similarity. To overcome this difficulty, the vertical scale of
the model is increased relative to the horizontal scale. The prediction of a
distorted model is relatively difficult, and the results of the model
cannot be easily transferred to the prototype.
REFERENCES
Fluid Mechanics with Engineering Application by Franzini
Abdul Halim Sir Fluid Mechanics