1.

Dimensional Analysis

In engineering the application of fluid mechanics in designs make much of the use of empirical
results from a lot of experiments. This data is often difficult to present in a readable form. Even
from graphs it may be difficult to interpret. Dimensional analysis provides a strategy for choosing
relevant data and how it should be presented.

This is a useful technique in all experimentally based areas of engineering. If it is possible to

identify the factors involved in a physical situation, dimensional analysis can form a relationship
between them.

The resulting expressions may not at first sight appear rigorous but these qualitative results
converted to quantitative forms can be used to obtain any unknown factors from experimental
analysis.

2. Dimensions and units

Any physical situation can be described by certain familiar properties e.g. length, velocity, area,
volume, acceleration etc. These are all known as dimensions.

Of course dimensions are of no use without a magnitude being attached. We must know more
than that something has a length. It must also have a standardised unit - such as a meter, a foot, a
yard etc.

Dimensions are properties which can be measured. Units are the standard elements we use to
quantify these dimensions.

In dimensional analysis we are only concerned with the nature of the dimension i.e. its quality not
its quantity. The following common abbreviation are used:

length = L

mass = M

time = T

force = F

temperature = Θ

In this module we are only concerned with L, M, T and F (not Θ). We can represent all the
physical properties we are interested in with L, T and one of M or F (F can be represented by a
combination of LTM). These notes will always use the LTM combination.

The following table (taken from earlier in the course) lists dimensions of some common physical
quantities:

1
 Quantity SI Unit . Dimension
-1
velocity m/s ms LT-1
acceleration m/s2 ms-2 LT-2
N
force
kg m/s2 kg ms-2 M LT-2
Joule J
energy (or work) N m,
kg m2/s2 kg m2s-2 ML2T-2
Watt W
power N m/s Nms-1
kg m2/s3 kg m2s-3 ML2T-3
Pascal P,
pressure ( or stress) N/m2, Nm-2
kg/m/s2 kg m-1s-2 ML-1T-2
density kg/m3 kg m-3 ML-3
N/m3
specific weight
kg/m2/s2 kg m-2s-2 ML-2T-2
a ratio 1
relative density .
no units no dimension
N s/m2 N sm-2
viscosity
kg/m s kg m-1s-1 M L-1T-1
N/m Nm-1
surface tension
kg /s2 kg s-2 MT-2

1. 3. Dimensional Homogeneity

Any equation describing a physical situation will only be true if both sides have the same
dimensions. That is it must be dimensionally homogenous.

For example the equation which gives for over a rectangular weir (derived earlier in this module)
is,

The SI units of the left hand side are m3s-1. The units of the right hand side must be the same.
Writing the equation with only the SI units gives

i.e. the units are consistent.

2
To be more strict, it is the dimensions which must be consistent (any set of units can be used and
simply converted using a constant). Writing the equation again in terms of dimensions,

Notice how the powers of the individual dimensions are equal, (for L they are both 3, for T both -
1).

This property of dimensional homogeneity can be useful for:

1. Checking units of equations;

2. Converting between two sets of units;
3. Defining dimensionless relationships (see below).

4. Results of dimensional analysis

The result of performing dimensional analysis on a physical problem is a single equation. This
equation relates all of the physical factors involved to one another. This is probably best seen in
an example.

If we want to find the force on a propeller blade we must first decide what might influence this
force.

It would be reasonable to assume that the force, F, depends on the following physical properties:

diameter, d

forward velocity of the propeller (velocity of the plane), u

fluid density, ρ

revolutions per second, N

fluid viscosity, µ

Before we do any analysis we can write this equation:

F = φ ( d, u, ρ, N, µ )

or

0 = φ1 ( F, d, u, ρ, N, µ )

where φ and φ1 are unknown functions.

3
These can be expanded into an infinite series which can itself be reduced to

F = Κ dm up ρθ Nr µσ

where K is some constant and m, p, q, r, s are unknown constant powers.

From dimensional analysis we

1. obtain these powers

2. form the variables into several dimensionless groups

The value of K or the functions φ and φ1 must be determined from experiment. The knowledge of
the dimensionless groups often helps in deciding what experimental measurements should be
taken.

5. Buckingham's π theorems

Although there are other methods of performing dimensional analysis, (notably the indicial
method) the method based on the Buckingham π theorems gives a good generalised strategy for
obtaining a solution. This will be outlined below.

There are two theorems accredited to Buckingham, and know as his π theorems.

1st π theorem:

A relationship between m variables (physical properties such as velocity, density etc.) can be
expressed as a relationship between m-n non-dimensional groups of variables (called π groups),
where n is the number of fundamental dimensions (such as mass, length and time) required to
express the variables.

So if a physical problem can be expressed:

φ ( Q1 , Q2 , Q3 ,………, Qm ) = 0

then, according to the above theorem, this can also be expressed

φ ( π1 , π2 , π3 ,………, Qm-n ) = 0

In fluids, we can normally take n = 3 (corresponding to M, L, T).

4
2nd π theorem

Each π group is a function of n governing or repeating variables plus one of the remaining
variables.

6. Choice of repeating variables

Repeating variables are those which we think will appear in all or most of the π groups, and are a
influence in the problem. Before commencing analysis of a problem one must choose the
repeating variables. There is considerable freedom allowed in the choice.

Some rules which should be followed are

1. From the 2nd theorem there can be n ( = 3) repeating variables.

2. When combined, these repeating variables variable must contain all of dimensions (M, L,
T)
(That is not to say that each must contain M,L and T).
3. A combination of the repeating variables must not form a dimensionless group.
4. The repeating variables do not have to appear in all π groups.
5. The repeating variables should be chosen to be measurable in an experimental
investigation. They should be of major interest to the designer. For example, pipe
diameter (dimension L) is more useful and measurable than roughness height (also
dimension L).

In fluids it is usually possible to take ρ, u and d as the threee repeating variables.

This freedom of choice results in there being many different π groups which can be formed - and
all are valid. There is not really a wrong choice.

7. An example

Taking the example discussed above of force F induced on a propeller blade, we have the
equation

0 = φ ( F, d, u, ρ, N, µ )

n = 3 and m = 6

There are m - n = 3 π groups, so

φ ( π1 , π2 , π3 ) = 0

5
The choice of ρ, u, d as the repeating variables satisfies the criteria above. They are measurable,
good design parameters and, in combination, contain all the dimension M,L and T. We can now
form the three groups according to the 2nd theorem,

As the π groups are all dimensionless i.e. they have dimensions M0L0T0 we can use the principle
of dimensional homogeneity to equate the dimensions for each π group.

For the first π group,

In terms of SI units

And in terms of dimensions

For each dimension (M, L or T) the powers must be equal on both sides of the equation, so

for M: 0 = a1 + 1

a1 = -1

for L: 0 = -3a1 + b1 + c1 + 1

0 = 4 + b1 + c1

for T: 0 = -b1 - 2

b1 = -2

c1 = -4 - b1 = -2

Giving π1 as

And a similar procedure is followed for the other π groups. Group

6
For each dimension (M, L or T) the powers must be equal on both sides of the equation, so

for M: 0 = a2

for L: 0 = -3a2 + b2 + c2

0 = b2 + c2

for T: 0 = -b2 - 1

b2 = -1

c2 = 1

Giving π2 as

And for the third,

For each dimension (M, L or T) the powers must be equal on both sides of the equation, so

for M: 0 = a3 + 1

a3 = -1

for L: 0 = -3a3 + b3 + c3 -1

b3 + c3 = -2

for T: 0 = -b3 - 1

b3 = -1

c3 = -1

Giving π3 as

7
Thus the problem may be described by the following function of the three non-dimensional π
groups,

φ ( π1 , π2 , π3 ) = 0

This may also be written:

8. Wrong choice of physical properties.

If, when defining the problem, extra - unimportant - variables are introduced then extra π groups
will be formed. They will play very little role influencing the physical behaviour of the problem
concerned and should be identified during experimental work. If an important / influential
variable was missed then a π group would be missing. Experimental analysis based on these
results may miss significant behavioural changes. It is therefore, very important that the initial
choice of variables is carried out with great care.

9. Manipulation of the π groups

Once identified manipulation of the π groups is permitted. These manipulations do not change the
number of groups involved, but may change their appearance drastically.

Taking the defining equation as: φ ( π1 , π2 , π3 ……… πm-n ) = 0

Then the following manipulations are permitted:

1. Any number of groups can be combined by multiplication or division to form a new

group which replaces one of the existing. E.g. π1 and π2 may be combined to form π1a =
π1 / π2 so the defining equation becomes
φ ( π1a , π2 , π3 ……… πm-n ) = 0
2. The reciprocal of any dimensionless group is valid. So φ ( π1 ,1/ π2 , π3 ……… 1/πm-n ) = 0
is valid.
3. Any dimensionless group may be raised to any power. So φ ( (π1 )2, (π2 )1/2, (π3 )3………
πm-n ) = 0 is valid.
4. Any dimensionless group may be multiplied by a constant.

8
 5. Any group may be expressed as a function of the other groups, e.g.
π2 = φ ( π1 , π3 ……… πm-n )

In general the defining equation could look like

φ ( π1 , 1/π2 ,( π3 )i……… 0.5πm-n ) = 0

10. Common π groups

During dimensional analysis several groups will appear again and again for different problems.
These often have names. You will recognise the Reynolds number ρud/µ. Some common non-
dimensional numbers (groups) are listed below.

Reynolds number inertial, viscous force ratio

Euler number pressure, inertial force ratio

Froude number inertial, gravitational force ratio

Weber number inertial, surface tension force ratio

Mach number Local velocity, local velocity of sound ratio

11. Examples

The discharge Q through an orifice is a function of the diameter d, the pressure difference p, the

density ρ, and the viscosity µ, show that , where φ is some unknown

function.

Write out the dimensions of the variables

ρ: ML-3 u: LT-1

d: L µ: ML-1T-1

p:(force/area) ML-1T-2

9
We are told from the question that there are 5 variables involved in the problem: d, p, ρ, µ and Q.

Choose the three recurring (governing) variables; Q, d, ρ.

From Buckingham's π theorem we have m-n = 5 - 3 = 2 non-dimensional groups.

For the first group, π1:

M] 0 = c1 + 1

c1 = -1

L] 0 = 3a1 + b1 - 3c1 - 1

-2 = 3a1 + b1

T] 0 = -a1 - 1

a1 = -1

b1 = 1

And the second group π2 :

(note p is a pressure (force/area) with dimensions ML-1T-2)

M] 0 = c2 + 1

c2 = -1

10
L] 0 = 3a2 + b2 - 3c2 - 1

-2 = 3a2 + b2

T] 0 = -a2 - 2

a2 = - 2

b2 = 4

So the physical situation is described by this function of non-dimensional numbers,

The question wants us to show :

Take the reciprocal of square root of π2: ,

Convert π1 by multiplying by this new group, π2a

then we can say

11
12