NPTEL Course Developer for Fluid Mechanics Dr.

Niranjan Sahoo
Module 05; Lecture 37 IIT Guwahati

DIMENSIONAL ANALYSIS AND HYDRAULIC SIMILITUDE

Introduction and Objective

 Many practical real flow problems in fluid mechanics can be solved by using
equations and analytical procedures.
 However, solutions of some real flow problems depend heavily on experimental
data. Based on the measurements, refinements in the analysis are made. Hence,
there is an essential link in this iterative process.
 Sometimes, the experimental work in the laboratory is not only time-consuming,
but also expensive. So, the main goal is to extract maximum information from
fewest experiments.
 In this regard, dimensional analysis is an important tool that helps in correlating
analytical results with experimental data.
 Also, some dimensionless parameters and scaling laws are introduced in order to
predict the prototype behavior from the measurements on the model.

Dimensional Analysis

 The analytically derived equations in engineering applications are correct for any
system of units and consequently each group of terms in the equation must have
the same dimensional representation. This is the law of dimensional homogeneity.
 In many instances, the variables involved in physical phenomena are known,
while the relationship among the variables is not known. Such a relationship can
be formulated between a set of dimensionless groups of variables and the groups
numbering less than the variables. This procedure is called dimensional analysis.
This procedure requires less experimentation and the nature of experimentation is
considerably simplified.
The following examples will make the things clear.
Example I
Consider a steady flow of an incompressible Newtonian fluid through a long, smooth
walled, horizontal circular pipe. It is desired to measure the pressure drop per unit length
of the pipe without the use of experimental data.

1
NPTEL Course Developer for Fluid Mechanics Dr. Niranjan Sahoo
Module 05; Lecture 37 IIT Guwahati

 The first step is to list out the variables that affect the pressure drop per unit

length Dpl . These variables may be pipe diameter D , fluid density r , fluid
viscosity m and mean velocity V at which the fluid is flowing through the pipe.
Thus, the relationship can be expressed as,
Dpl = g ( D, r , m , V ) (1)
At this point, the nature of the function is unknown and the experiments are to be
performed to determine the nature of the function.
 In order to perform the experiments in a systematic and meaningful manner, it is
necessary to change one variable at a time keeping the others constant and
measure the corresponding pressure drop. The series of tests would result the data
that can be represented in graphical form as shown in Fig. 1 (a-d).

Fig. 1: Illustration of factors affecting the pressure drop in a pipe flow.

Referring to the Fig. 1(c), it would be difficult to vary fluid density while holding
viscosity constant. Moreover, it would be rather impossible to obtain a general

functional relationship between Dpl , D , r , m and V for any similar pipe

system.

2
NPTEL Course Developer for Fluid Mechanics Dr. Niranjan Sahoo
Module 05; Lecture 37 IIT Guwahati

 A simple approach to this problem is to collect two non-dimensional

combinations of the variables (i.e. dimensionless products/dimensional groups)
such that
DDpl �rVD �
=f� � (2)
rV 2
�m �
Now, the working variables are reduced to only two instead of five. The necessary

experiment would simply consist of varying the dimensionless product ( rVD m )

and determining the corresponding value ( DDP

l rV 2 ) . The results of the

experiment could then be represented by a single, universal curve as illustrated in

Fig. 2. It would be valid for any combination of smooth-walled pipe and
incompressible Newtonian fluid.

Fig. 2: Illustrative plot of pressure drop in a pipe flow using dimensionless parameters.

Example II
Consider a problem to determine the drag force on a smooth sphere moving with
certain velocity in a viscous fluid.
 The first step is to list out the variables that affect the drag force (say F ). These
variables are diameter of sphere D , velocity of the sphere V , fluid density r and
viscosity of the fluid m . It can be represented by some unknown function as,
F = f ( D, V , r , m ) (3)

3
NPTEL Course Developer for Fluid Mechanics Dr. Niranjan Sahoo
Module 05; Lecture 37 IIT Guwahati

 Next step is to determine this relationship experimentally by varying one

parameter at a time in the right-hand side parentheses of Eq. (3) by keeping the
others constant. These results may be represented in graphical form as shown in
Fig. 3. It would also lead to similar kinds of situations, which are extremely time-
consuming and expensive investigation.

Fig. 3: Illustration of factors affecting the drag force on a sphere; (a) Vary density ( r )

holding viscosity ( m ) constant; (b) Vary viscosity ( m ) holding density ( r ) constant;

In order to eliminate these difficulties, the drag force can be expressed as functional
relation between two dimensionless groups as,
F �rVD �
=j� � (4)
rV D
2 2
�m �
where the nature of the function j is not known. However, by experiment, a single curve
may be obtained by relating the dimensionless groups as shown in Fig. 4.

Fig. 4: Illustrative plot of drag force on a sphere using dimensionless parameters.

4
NPTEL Course Developer for Fluid Mechanics Dr. Niranjan Sahoo
Module 05; Lecture 37 IIT Guwahati

Thus, the dimensionless combination of variables helps in building up the functional

relationship for the experimental data, which is independent of the system of units (which
is clear from the above examples). This method is known as dimensional analysis and the
basis for its application to wide variety of problems is found in the Buckingham pi
theorem.

Principle of Dimensional Homogeneity

It is stated as, “If an equation truly expresses a proper relationship between variables in a
physical process, then it will be dimensionally homogeneous”. It means each of its
additive terms will have the same dimension. For example,
1 2
 Displacement of a free falling body is, S = S0 + V0t + gt . Each term in this
2

equation has the dimension of length [ L ] and hence it is dimensionally

homogeneous.
p 1 2
 Bernoulli’s equation for incompressible flow is, + V + gz = C . Each term in
r 2

L2 .T -2 �
this equation including the constant has dimension of velocity squared �
� �
and hence it is dimensionally homogeneous.
Dimensional variables: These are the quantities, which actually vary during a given case
and can be plotted against each other. In the first example, S and t are the variables

where as in the second example, the variables are p, V and t .

Dimensional constants: These are normally held constant during a given run. But, they

may vary from case to case. In above examples, S0 , V0 , r , g and C are the dimensional
constants.
Pure constants: have no dimensions. But, they arise from mathematical manipulation. In

1
above examples, arises from mathematical manipulation. The other common
2
dimensionless constants are, p and e .

5
NPTEL Course Developer for Fluid Mechanics Dr. Niranjan Sahoo
Module 05; Lecture 37 IIT Guwahati

Buckingham pi Theorem
 It states that if an equation involving k variables is dimensionally homogeneous,

it can be reduced to a relationship among ( k - r) independent dimensionless

products, where r is the minimum number of reference dimensions required to
describe the variable.
 The dimensionless products are frequently referred to as pi terms and the theorem
is named accordingly after famous scientist Edgar Buckingham (1867-1940). It is
based on the idea of dimensional homogeneity.
 Mathematically, if a physically meaningful equation involving k variables is
assumed,
y = f ( x1 , x2 .........., xk ) (5)
such that the dimensions of the variables on the left side of the equation are equal
to the dimensions of any term on the right side of equation, then, it is possible to
rearrange the above equation into a set of dimensionless products (pi terms), so
that
P1 = j ( P 2 , P 3 .........., P k - r ) (6)

where j ( P 2 , P 3 .........., P k - r ) is a function of P 2 through P k - r .

 The required number of pi terms is less than the number of original variables by r
, where r is determined by the minimum number of reference dimensions
required to describe the original list of variables. These reference dimensions are
usually the basic dimensions M , L and T (Mass, Length and Time) .

Determination of pi Terms
Several methods can be used to form dimensionless products or pi terms that arise in
dimensional analysis. But, there is a systematic procedure called method of repeating
variables that allows in deciding the dimensionless and independent pi terms. For a given
problem, following distinct steps should be followed.
Step I: List out all the variables that are involved in the problem.

6
NPTEL Course Developer for Fluid Mechanics Dr. Niranjan Sahoo
Module 05; Lecture 37 IIT Guwahati

 The ‘variable’ is any quantity including dimensional and non-dimensional

constants in a physical situation under investigation. Typically, these variables are
those that are necessary to describe the “geometry” of the system (diameter,
length etc.), to define fluid properties (density, viscosity etc.) and to indicate the
external effects influencing the system (force, pressure etc.).
 All the variables must be independent in nature so as to minimize the number of
variables required to describe the system.
Step II: Express each variable in terms of basic dimensions.
Typically, for fluid mechanics problems, the basic dimensions will be either
M , L and T or F , L and T . Dimensionally, these two sets are related through Newton’s

second law ( F = m.a ) so that F = MLT -2 e.g. r = ML-3 or r = FL-4T 2 . It should be

noted that these basic dimensions should not be mixed.
Step III: Decide the required number of pi terms.
It can be determined by means of Buckingham pi theorem which indicates that the

number of pi terms is equal to ( k - r ) , where k is the number of variables in the problem

(determined from Step I) and r is the number of reference dimensions required to
describe these variables (determined from Step II).
Step IV: Select the number of repeating variables.
 Amongst the original list of variables, select those variables that can be combined
to form pi terms.
 The required number of repeating variables is equal to the number of reference
dimensions.
 Each repeating variable must be dimensionally independent of the others i.e. they
cannot themselves be combined to form dimensionless product.
 Since there is a possibility of repeating variables to appear in more than one pi
term, so dependent variables should not be chosen as one of the repeating
variable.
Step V: Formation of pi terms
Essentially, the pi terms are formed by multiplying one of the non-repeating variables
by the product of the repeating variables each raised to an exponent that will make the

7
NPTEL Course Developer for Fluid Mechanics Dr. Niranjan Sahoo
Module 05; Lecture 37 IIT Guwahati

a b c
combination dimensionless. It usually takes the form of xi x1 x2 x3 where the exponents
a , b and c are determined so that the combination is dimensionless.
Step VI: Repeat the Step V for each of the remaining non-repeating variables. The
resulting set of pi terms will correspond to the required number obtained from Step III.
Step VII: Checking of pi terms
Make sure that all the pi terms must be dimensionless. It can be checked by simply
substituting the basic dimension ( M , L and T ) of the variables into the pi terms.
Step VIII: Final form of relationship among pi terms
Typically, the final form among the pi terms can be written in the form of Eq. (6)

where P1 would contain the dependent variable in the numerator. The actual functional
relationship among pi terms is determined from experiment.

Illustration of Pi Theorem

In order to illustrate various steps in the pi theorem, following examples are considered.

Example 1 (Pressure drop in a pipe flow)

Consider a steady flow of an incompressible Newtonian fluid through a long, smooth
walled, horizontal circular pipe. It is desired to measure the pressure drop per unit length
of the pipe without the use of experimental data.
 According to Step I, all the pertinent variables involved in the experimentation of
pressure drop per unit length of the pipe, must be noted and should be expressed
in the form,
Dpl = f ( D, r , m ,V ) (7)
where D is the pipe diameter, r is the fluid density, m is the viscosity of the
fluid and V is the mean velocity at which the fluid is flowing through the pipe.
 Next step is to express all the variables in terms of basic dimensions i.e.
M , L and T . It then follows that

Dpl = ML-2T -2 , D = L , r = ML-3 , m = ML-1T -1 and V = LT -1 (8)

8
NPTEL Course Developer for Fluid Mechanics Dr. Niranjan Sahoo
Module 05; Lecture 37 IIT Guwahati

 Third step is to apply pi theorem to decide the number of pi terms required. Since
there are five variables (including the dependent variable Dp ) and three reference

dimensions, so k = 5 and r = 3 . So, only two ( k - r ) pi terms are required.

 The repeating variables to be used to form pi terms (Step IV) need to be selected
from the list D, r , m and V . It is to be noted that the dependent variable should
not be used as one of the repeating variable. Since, there are three reference
dimensions involved, so we need to select three repeating variable. These
repeating variables should be dimensionally independent i.e. dimensionless
product can not be formed from this set. In this case, D, r and V are chosen as
the repeating variables.
 Now, first pi term is formed between the dependent variable and the repeating
variables. It is written as,
P1 = Dpl D a V b r c (9)
Since this combination need to be dimensionless, it follows that

( ML -2
T -2 ) ( L ) ( LT ) ( ML )
a -1 b -3 c
= M 0 L0T 0 (10)

The exponents a, b and c must be determined by equating the exponents for

each of the terms M , L and T i.e.

1 + c = 0 (For M )
-2 + a + b - 3c = 0 (For L) (11)
-2 - b = 0 (For T )

The solution of this algebraic equations gives a = 1; b = -2; c = -1 . Therefore,

Dpl D
P1 = (12)
rV 2
 The process is repeated for remaining non-repeating variables with other

additional variable ( m ) so that,

P 2 = m .D d .V e .r f (13)
and

( ML T ) ( L ) ( LT ) ( ML ) -1 e
d f
-1 -1 -3
= M 0 L0T 0 (14)

9
NPTEL Course Developer for Fluid Mechanics Dr. Niranjan Sahoo
Module 05; Lecture 37 IIT Guwahati

Equating the exponents,

1 + f = 0 (For M )
-1 + d + e - 3 f = 0 (For L) (15)
-1 - e = 0 (For T )

The solution of this algebraic equation gives d = -1; e = -1; f = -1 . Therefore,

m
P2 = (16)
rVD
 Now, the correct numbers of pi terms are formed as determined in step 3. In order
to make sure about the dimensionality of pi terms, they are written as,

P1 =
Dpl D
=
(
ML-2T -2 ( L ) )
= M 0 L0T 0
rV ( )( )
2 2
ML-3 LT -1

P2 =
m
=
( ML-1T -1 ) ( L )
= M 0 L0T 0
r V D ( ML-3 ) ( LT -1 ) ( L )

 Finally, the result of dimensional analysis is expressed among the pi terms as,
D Dpl � m �
=f� � (17)
rV 2
�r V D �

10