Dimensional Analysis

ME 305 Fluid Mechanics I

• Consider that we are interested in determining how the drag force acting on a
smooth sphere immersed in a uniform flow depends on other fluid and flow
variables.
Part 7 • Important varibles of the problem are shown below (How did we decide on these?).
𝑉
𝜇, 𝜌
Dimensional Analysis and Similitude 𝐹𝐷
𝐷
These presentations are prepared by
Dr. Cüneyt Sert
Department of Mechanical Engineering • Drag force 𝐹𝐷 is thought to depend on the following variables.
Middle East Technical University
𝐹𝐷 = 𝑓(𝐷, 𝑉, 𝜇, 𝜌)
Ankara, Turkey
[email protected] • In order to find the actual functional relation we need to perform a set of
experiments.
You can get the most recent version of this document from Dr. Sert’s web site.
• Dimensional analysis helps us to design and perform these experiments in a
Please ask for permission before using them to teach. You are NOT allowed to modify them.
7-1
systematic way. 7-2

Dimensional Analysis (cont’d) Dimensional Analysis (cont’d)

• The following set of controlled experiments should be done. • It is possible to simplify the dependency of drag force on other variables by using
nondimensional (unitless) parameters.
• Fix 𝐷, 𝜇 and 𝜌. Change 𝑉 and measure 𝐹𝐷 .
𝐹𝐷 𝜌𝑉𝐷
• Fix 𝑉, 𝜇 and 𝜌. Change 𝐷 and measure 𝐹𝐷 . = 𝑓1
𝜌𝑉 2 𝐷 2 𝜇
• Fix 𝐷, 𝑉 and 𝜇 . Change 𝜌 and measure 𝐹𝐷 .
Note : These are only illustrative
Nondimensional drag Nondimensional
• Fix 𝐷, 𝑉 and 𝜌. Change 𝜇 and measure 𝐹𝐷 . figures. They do not correspond to
force (Drag coefficient)
any actual experimentation. Reynolds number (𝑅𝑒)

𝐹𝐷 Constant 𝐹𝐷 Constant 𝐹𝐷 Constant 𝐹𝐷 Constant

𝐷, 𝜇, 𝜌 𝑉, 𝜇, 𝜌 𝐷, 𝑉, 𝜇 𝐷, 𝑉, 𝜌
𝐹𝐷
𝜌𝑉 2 𝐷 2 Illustrative figure
𝑉 𝐷 𝜌 𝜇

• We need to perform too many experiments.

• Also there are major difficulties such as finding fluids with different densities, but 𝑅𝑒
same viscosities.
Flow over a sphere at 𝑅𝑒 = 15000
7-3 7-4

1
 Dimensional Analysis (cont’d) Buckingham Pi Theorem
• To find this new relation, we only need to change the Reynolds number. • Buckingham Pi theorem can be used to determine the nondimensional groups of
variables (Pi groups) for a given set of dimensional variables.
• We can do it in any way we want, e.g. the simplest way is to change the speed of air
flow in a wind tunnel. • For the flow over a sphere problem studied previously, dimensional parameter set is
(𝐹𝐷 , 𝐷, 𝑉, 𝜇, 𝜌) and this theorem helps us to find two Pi groups as
• All 𝑅𝑒 = 15000 flows around a sphere will look like the same and they all provide
the same nondimensional drag force. It does not matter what fluid we use or how 𝐹𝐷 𝜌𝑉𝐷
big the sphere is (be aware of very extreme cases). Π1 = and Π2 =
𝜌𝑉 2 𝐷 2 𝜇

• Dimensional analysis is used to formulate a physical phenomenon as a relation • Let’s explain how this works using “the drag force acting on a sphere” problem.
between a set of nondimensional (unitless) groups of variables such that the
number of these groups is less than the number of dimensional variables. • Step 1 : List all the dimensional variables involved in the problem.

• It is important to develop a systematic and meaningful way to perform experiments. • 𝑛 is the number of dimensional variables. 𝑛 = 5 for our example.
• These variables should be independent of each other. For example if the diameter of
• Nature of the experiments are simplified and the number of required experiments is
a sphere is in the list, frontal area of the sphere can not be included.
reduced.
• If body forces are important in a problem, gravitational acceleration should be in the
list, although it is a constant.

7-5 7-6

Buckingham Pi Theorem (cont’d) Buckingham Pi Theorem (cont’d)

• Step 2 : Express each of the variables in tems of basic dimensions, which are • Step 3: Determine the repeating variables that are allowed to appear in more than
one Pi group.
𝐿 : length , 𝑇 : time , 𝑀 : mass
• There should be 𝑟 many repeating variables.
• For problems involving heat transfer Θ (temperature) can also be a basic dimension.
• If 𝐿 is a primary dimension of the problem, we should select one geometric variable
• For the example we are studying basic dimensions of variables are
as a repeating variable.
𝑀𝐿 𝐿 𝑀 𝑀
[𝐹𝐷 ] = 𝑇2
, 𝐷 = 𝐿 , [𝑉] = 𝑇
, [𝜌] = 𝐿3
, [𝜇] = 𝐿𝑇 • If 𝑇 is a primary dimension of the problem, we should select one kinematic variable
as a repeating variable.
• Our example involves 𝑟 = 3 primary dimensions. For most fluid mechanics problems
• If 𝑀 is a primary dimension of the problem, we should select one dynamic variable
𝑟 will be 3.
as a repeating variable.
• Variables having only 𝐿 in their dimension are called geometric variables. • Note that this selection is not unique and the resulting Pi groups will depend on our
selection. Certain selections are ‘‘better’’ than others.
• Variables having only 𝑇 or both 𝐿 and 𝑇 are called kinematic variables.
• For the problem of interest we can select 𝐷 , 𝑉 and 𝜌 as repeating variables.
• Variables having 𝑀 in their dimension are called dynamic variables.
• If there is an obvious dependent variable in the problem, do not select it as a
• For our example 𝐷 is a geometric, 𝑉 is a kinematic and 𝐹𝐷 , 𝜇 and 𝜌 are dynamic
repeating variable. In our example 𝐹𝐷 is a dependent variable. We are trying to
variables.
understand how it depends on other variables.
7-7 7-8

2
 Buckingham Pi Theorem (cont’d) Buckingham Pi Theorem (cont’d)
• Step 4: Determine (𝑛 − 𝑟) many Pi groups by combining repeating variables with • Now determine the second Pi group which has 𝜇 as the nonrepeating variable.
nonrepeating variables and using the fact that Pi groups should be nondimensional.
Π2 = 𝜇 𝐷 𝑎 𝑉 𝑏 𝜌 𝑐
• For our example we need to find 5 − 3 = 2 Pi groups. Each Pi group will include only
one of the nonrepating variables. 𝑀 𝐿 𝑀 𝑐
• Π2 should be unitless : − = [𝐿]𝑎 [ ]𝑏
𝐿𝑇 𝑇 𝐿3

Π1 = 𝐹𝐷 𝐷 𝑎 𝑉 𝑏 𝜌 𝑐
We need to determine Π2 should have no 𝐿 dimension : 0 = −1 + 𝑎 + 𝑏 − 3𝑐
𝑎, 𝑏 and 𝑐. 𝑎 = −1 𝜇
A nonrepeating Π2 should have no 𝑇 dimension : 0 = −1 − 𝑏 𝑏 = −1 Π2 =
parameter Unknown combination 𝜌𝐷𝑉
𝑐 = −1
of repeating parameters Π2 should have no 𝑀 dimension : 0 = 1 + 𝑐

𝑀𝐿 𝐿 𝑀 𝑐 • Therefore the relation of nondimensional groups that we are after is

• Π1 should be unitless : − = [𝐿]𝑎 [ ]𝑏
𝑇2 𝑇 𝐿3
𝐹𝐷 𝜇
Π1 = 𝑓1 Π2 → = 𝑓1
Π1 should have no 𝐿 dimension : 0 = 1 + 𝑎 + 𝑏 − 3𝑐 𝜌𝑉 2 𝐷 2 𝜌𝑉𝐷
𝑎 = −2 𝐹𝐷
Π1 should have no 𝑇 dimension : 0 = −2 − 𝑏 𝑏 = −2 Π1 = 𝜌𝑉𝐷
𝜌𝐷 2 𝑉 2 • It is better to write the second Pi group as because it is the well known
𝜇
𝑐 = −1
Π1 should have no 𝑀 dimension : 0 = 1 + 𝑐 Reynolds number.
7-9 7-10

Exercises for Buckingham Pi Theorem Important Nondimensional Numbers of Fluid Mechanics

Exercise : Consider the flow of an incompressible fluid through a long, smooth- • Following nondimensional numbers frequently appear as a Pi group.
walled horizontal, circular pipe. We are interested in analyzing the pressure drop, 𝜌𝑉𝐿 𝑉𝐿
• Reynolds number : 𝑅𝑒 = = . Ratio of inertia forces to viscous forces.
∆𝑝, over a pipe length of 𝐿. Other variables of the problem are pipe diameter (𝐷), 𝜇 𝜈
average velocity (𝑉) and fluid properties (𝜌 and 𝜇). Determine the Pi groups by a) Δ𝑝
• Euler number : 𝐸𝑢 = 1 . Ratio of pressure forces to inertia forces.
selecting 𝜌 as a repeating parameter, b) selecting 𝜇 as a repeating parameter. 2
𝜌𝑉 2
𝑉
• Froude number : 𝐹𝑟 = . Squareroot of the ratio of inertia forces to gravitational
Exercise : In a laboratory experiment a tank is drained through an orifice from initial 𝑔𝐿
liquid level ℎ0 . The time, 𝜏, to drain the tank depends on tank diameter, 𝐷, orifice forces.
diameter, 𝑑, gravitational acceleration, 𝑔, liquid properties, 𝜌 and 𝜇. Determine the 𝑉 𝑉
• Mach number : 𝑀𝑎 = = . Squareroot of the ratio of inertia forces to
Pi groups. 𝐸𝑣 /𝜌 𝑐
compressibility forces.
Exercise : The diameter, 𝑑, of the dots made by an ink jet printer depends on the ink
𝜌𝑉 2 𝐿
properties, 𝜌 and 𝜇, surface tension, 𝜎, nozzle diameter, 𝐷, the distance, 𝐿, of the • Weber number : 𝑊𝑒 = . Squareroot of the ratio of inertia forces to surface
𝜎
nozzle from the paper and the ink jet velocity, 𝑉. Determine the Pi groups. tension forces.
𝜔𝐿
Exercise : The power, 𝒫, required to drive a propeller is known to depend on the • Strouhal number : 𝑆𝑡 = . Used for flows with oscillatory (periodic) behavior.
𝑉
following variables: freestream speed, 𝑉, propeller diameter, 𝐷, angular speed, 𝜔, 𝑝−𝑝𝑣
• Cavitation number : 𝐶𝑎 = 1 . Used for possibly cavitating flows.
fluid properties, 𝜌 and 𝜇, and the speed of sound 𝑐. Determine the Pi groups. 2
𝜌𝑉 2
7-11 7-12

3
 Model and Prototype Three Basic Laws of Similitude
• In experimental fluid mechanics we sometimes can not work with real sized objects, • A similitude analysis is done to make sure that the results obtained from an
known as prototypes. experiment can correctly be transferred to the real flow field.
• Instead we use scaled down (or up) versions of them, called models. • Three basic laws of similitude must be satisfied in order to achieve complete
• Also sometimes in experiments we use fluids that are different than actual working similarity between prototype and model flow fields.
fluids, e.g. we use regular tap water instead of salty sea water to test the 1. Geometric similarity : Model and prototype must be the same in shape, but
performance of a marine propeller. can be different in size. All linear dimensions of the model must be related to
corresponding dimensions of the prototype by a constant length ratio, 𝐿𝑟 .
• It is usually impossible to establish 100 % geometric similarity due to very small
details that can not be put into the model. Modeling surface roughness exactly
is also impossible.

𝐿𝑝 𝐷𝑝
𝐿𝑟 = = Model
Prototype 𝐿𝑚 𝐷𝑚
www.boeing.com www.reuters.com
𝐷𝑝
Wind tunnel tests of an airliner Race car being tested in a water tunnel 𝐷𝑚
http://www.reuters.com/news/video/story?videoId=131255095 𝐿𝑝 𝐿𝑚
7-13 7-14

Three Basic Laws of Similitude (cont’d) Three Basic Laws of Similitude (cont’d)
2. Kinematic similarity : Model and prototype flow fields are kinematically similar if 3. Dynamic similarity : Two flow fields should have force distributions such that
the velocities at corresponding points are the same in direction and differ only by a identical types of forces are parallel and are related in magnitude by a constant factor
constant factor of velocity ratio, 𝑉𝑟 . of force ratio.
• This also means that the streamline patterns of two flow fields should differ by a • If a certain type of force, e.g. compressibility force, is highly dominant in the
constant scale factor. prototype flow, it should also be dominant in the model flow.
• How suitable would it be to use a water tunnel to study the aerodynamic
forces acting on a supersonic missile ?

𝑉𝑝 𝑉𝑝 • If a certain type of force, e.g. surface tension force, is negligibly small in the
𝐴 𝐵
𝑉𝑟 = = Model prototype flow, it should also be small in the model flow.
B 𝑉𝑚 𝐴 𝑉𝑚 𝐵
A B • How suitable would it be to use a very light and very small model to test
A
Prototype the forces acting on a ship ?

• To establish dynamic similarity we need to determine the important forces of the

prototype flow and make sure that the nondimensional numbers related to those
forces are the same in prototype and model flows.

7-15 7-16

4
 Similitude (cont’d) Similitude (cont’d)
Exercise : The aerodynamic drag of a new car is to be predicted at a speed of • The important question is how to decide on the important force types for a given
100 km/h at an air temperature of 25 oC. Engineers build a one fifth scale of the car problem? In other words equality of which nondimensional numbers should be
to test in a wind tunnel. It is winter time and the air in the tunnel is 5 oC. Determine sought?
how fast the engineers should run the wind tunnel in order to achieve similarity
• Reynolds number similarity is important for almost all flows.
between the model and the prototype. How each Newton of drag force measured
on the model be transferred to the prototype ? • Froude number similarity is important for flows with free surfaces, such as ship
resistance, open channel flows and for flows driven by the action of gravity.
Prototype Model • Euler number similarity is important mostly for turbomachinery flows with
𝑉𝑝
𝜇𝑝 , 𝜌𝑝 considerable pressure changes, for which cavitation may be a concern.
𝑉𝑚 𝜇𝑚 , 𝜌𝑚
𝐹𝑝 𝐹𝑚 • Mach number similarity is important for high speed flows.
• Weber number similarity is important for problems involving interfaces with
density changes and light weighted objects.
𝐿𝑝 𝐿𝑚
• Strouhal number similarity is important for flows with an oscillatory (time
𝑅𝑒𝑝 = 𝑅𝑒𝑚 periodic) flow pattern, such as Karman vortices shed from bodies.

𝜌𝑝 𝑉𝑝 𝐷𝑝 𝜌𝑚 𝑉𝑚 𝐷𝑚 • After deciding on the important parameters of the problem (which is not always
=
𝜇𝑝 𝜇𝑚 an easy task) Buckhingham Pi theorem will also end up with these numbers.
7-17 7-18

Similitude Exercises Similitude Exercises (cont’d)

Exercise : The drag force on a submarine, which is moving well below the free Exercise : The model described in the previous problem will now be used to
surface, is to be determined by a test on a model, which is scaled down to one- determine the drag force of a submarine, which is moving on the surface. The
twentieth of the prototype. The test is to be carried in a water tunnel. The density properties of the sea water are as given above. The speed of the prototype is 2.6
and kinematic viscosity of the seawater are 1010 kg/m3 and 1.3x10-6 m2/s. The m/s.
water in the tunnel has a density of 988 kg/m3 and a kinematic viscosity of
0.65x10-6 m2/s. If the speed of the prototype is 2.6 m/s, then determine the a) Determine the speed of the model.

a) speed of the model. b) Determine the kinematic viscosity of the liquid that should be used in the
experiments.
b) ratio of the drag force in the prototype to the one in the model.
c) If such a liquid is not available, sea
water will be used in the experiments.
Neglecting the viscous effects, determine
the ratio of the drag force due to the
surface waves. This is called incomplete
similarity.

http://www.aviationweek.com

7-19 7-20

5
 Similitude Exercises (cont’d) Similitude Exercises (cont’d)
Exercise : A long structural component of a
Exercise : A model of a harbor is to be made with a scale ratio of 300. Storm waves
bridge has an elliptical cross section. It is
having amplitude of 2 m occur on the breakwater of prototype harbor at a speed of
known that when an unsteady wind blows
10 m/s.
past this type of bluff body, vortices may
a) Neglecting the frictional effects, determine the amplitude and speed of the develop on the downwind side that are shed
waves in the model. in a regular fashion at some definite frequency.
Since these vortices can create harmful periodic forces acting on the structure, it is
b) If the tidal period in the prototype is 12 h, determine the tidal period in the important to determine the shedding frequency. For the specific structure of
model. interest, 𝐷 = 0.1 m, 𝐻 = 0.3 m, and a representative wind velocity is 50 km/hr.
Standard air can be assumed. The shedding frequency is to be determined through
Exercise : An airplane travels in air at a velocity of 200 m/s at an altitude of 5 km. the use of a small scale model that is to be tested in a water tunnel. For the model,
Pressure and temperature of air are 55 kPa and -20 oC, respectively. A model of this 𝐷𝑚 = 20 mm and the water temperature is 20 ℃.
airplane with a length scale of 10 is tested in a wind tunnel at 20 oC. The specific
Determine the model dimension, 𝐻𝑚 , and the velocity at
heat ratio and gas constant for air are 1.4 and 287.1 J/kgK, respectively. Taking the
which the test should be performed. If the shedding
effect of compressibility into account, determine the
frequency for the model is found to be 49.9 Hz, what is
a) velocity of air in the wind tunnel, the corresponding frequency for the prototype? Movie
Collapse of Tacoma Narrows Bridge
b) density of air in the wind tunnel.
http://www.youtube.com/watch?v=j-zczJXSxnw
7-21 7-22