LECTURE 41

21- APRIL -2017

MODELLING and SIMILARITY

 The dimensional analysis what you are studying is used in modelling studies.
 You saw the Buckingham Π theorem.

 How will you do it practically?

 Selection of variables is important. It requires good experience to decide which

variable is important, etc.
 Based on the requirement, the engineer can judge and suggest
 That viscous flows can be neglected or not.
 Temperature effects can be neglected or not.
 Whether surface tension is important or not.
 Gravitational force is important or not, etc.

 Once the variables are selected, you need to perform analysis. Identify  ’s.
 The experimenters will try to seek similarity between the model and the
prototype designed.

 On any phenomena, the engineer may come up with a non-dimensional relation.

1  f  1 ,  2 ,  3 ,................,  k 
 The flow conditions of a model test will be completely similar to the prototype, if all
the dimensionless parameters of the model (say  2,  3,  4,……  k) is same as the
prototype.

i.e. You require

 2m =  2p ;  3m =  3p ; ……

 It is very difficult to achieve complete similarity between a model and a prototype.

 However, by compromising various features, you can have:
 Geometric similarity between model and prototype
 Kinematic similarity
 Dynamic similarity
 Thermal similarity
 Geometric similarity
If all the length dimensions [L] of the model and the prototype are having same linear
scale ratio, the model and prototype is said to be geometrically similar.

For a body, if the dimensions in the x, y, z directions are Lxp, Lyp, Lzp
Then, if you have a miniature model of dimensions Lxm, Lym, Lzm .

Lxm Lym Lzm

And if    constant,
Lx p Ly p Lz p
Then this model is geometrically similar to the prototype)

 You can visualise homologous points between the model and prototype.

 Kinematic Similarity
 If the model and prototype have same length scale ratio (i.e. geometric
similarity) and in addition have same time scale ratio, then they are
kinematically similar.
 For time scale equilibrium, you may require evidence of Reynolds Number and
Mach Number, etc.
 For fluid flows you have seen that one can neglect or incorporate effects of
viscosity (or friction).
 For a frictionless flow like free surface flow, the model and prototype can be
made kinematically similar. (How???)
 You have to provide the same Froude Number.

U
 Recall earlier, we had described Froude’s Number as Fr 
gL
 For example in a Wave Motion Modelling Studies:
The prototype wave:

(Source :https://en.wikibooks.org/wiki/Alevel_Physics_(Advancing_Physics)/What_is_a_wave%3F)

 In the wave modelling studies, the actual or prototype wave may be large and its wave
length can be large.
 A model of the wave can be developed in laboratory by providing kinematically similar
conditions.
 In kinematic similarity, the Froude number have to be equal.

Vm Vp
i.e. Frm    Fr p
gLm gL p

2 2
Vm Vp
or 
gLm gL p

Vm Lm
or 
Vp Lp

Lm V
If =  , Then m  
Lp Vp

Lm
tm Vm
  
t p Lp
Vp

 If viscosity, and surface tension, etc. play important role in flow of the fluid, then
kinematic similarity may not be sufficient. You may have to go for dynamic similarity.
Dynamic Similarity:
 The model and the prototype are dynamically similar, if they have same length scale
ratio, same time scale ratio and same force ratio.
 From Newton’s law,for any fluid particle-the sum of pressure force , gravity force, and
friction force is equal to the net force or inertial force.

Fp Fg Ff Fi
i.e.       
Pressure Gravity Friction Inertia
force force force force or
Net force

 The dynamic similarity consider that

  
Fp m Fg m Ff m
    
Fp p Fg p Ff p

 In this flow through the sluice gate experiment, the force polygon at homologous points
will have exactly the same shape if they are dynamically similar. This is possible, only
if Rep = Rem and Frm = Frp.

Example (Adopted from FM White’s Fluid Mechanics)

 The pressure drop due to friction in a long circular pipe is a function of average
velocity,density,viscosity,and pipe length and diameter.

i.e. ΔP = f( V,ρ,µ,L,D)

 Try to interpret how ΔP varies with V and use Pi-Theorem to rewrite the above function
in dimensionless form.
 Solution:
You are provided the information ΔP = f( V,ρ,µ,L,D)
 These are six basic variables ΔP,V,ρ,µ,L,D (i.e. n=6)
 The basic dimensions are M,L,T (i.e. k=3)
 We expect j = n-k = 6-3 = 3 ∏ groups.
D L Q ΔP ρ µ V
(cm) (cm) (m3/h) (Pa) (kg/m3) (kg/ms) (m/s)
1.0 5.0 0.30 4680 680 2.92 * 10-4 1.06
1.0 7.0 0.60 22300 680 2.92 * 10-4 2.12
1.0 9.0 1.00 70800 680 2.92 * 10-4 3.54
2.0 4.0 1.00 2080 998 0.0010 0.88
2.0 6.0 2.00 10500 998 0.0010 1.77
2.0 8.0 3.10 30400 998 0.0010 2.74
3.0 3.0 0.50 540 13550 1.56 * 10-3 0.20
3.0 6.0 1.70 9600 13550 1.56 * 10-3 0.67
3.0 4.0 1.00 2480 13550 1.56 * 10-3 0.39
 As we went to see how ΔP varies with V, we are not going to keep it as repeating
variable.

We will keep ρ,µ,D as repeating variable

1   a  b D c p  M 0 L0T 0

2   d  eD f V  M 0 L0T 0

3   g  h Di L  M 0 L0T 0

Solving for exponents, you get:

D 2 P VD L
1  , 2  , 3 
 2
 D

We will get a dimensionless expression

D 2 P  VD L 
 f  , 
   D
2

1  f ( 2 ,  3 )

As our objective is to find ΔP versus V,

We have to plot 1 versus  2 with  3 as a parameter.

 L
i.e.  500 for the first row of the data
D

 D2 P
 3.73 109
 2

VD
 24700

From second row of the data,

L  D2 P
 700 ,  1.78 1010
D  2

VD
 49370

Like this you evaluate for each row. Nine data points you will get.

L
 For same ratio , you will see ΔP is increasing linearly with L.
D