ME5129: Principles of Thermal Energy

Conversion
Dr. V. Raghavan
Department of Mechanical Engineering
Thermodynamics and Combustion Engineering Lab
IIT Madras
Phone: 044-22574712
Email: [email protected]

Dr. V. Raghavan, IIT Madras 1

 Fluids and Continuum
• Liquids and gasses move under the action of the shear stress, no matter
how small that shearing stress may be. These are called fluids.
• Each fluid is considered to be continuously distributed throughout a region
of interest. Liquids obey this continuum always.
• Gasses, under certain circumstances obey this continuum, when the
molecules are sufficiently close. A metric of mean free path is used to
determine whether or not the fluid obeys continuum.
• If the fluid is a continuum, the density can be defined as:

m
 = lim
V →0 V
Solid
Dr. V. Raghavan, IIT Madras 2
 Properties of fluids
• Specific weight (  ): Specific weight is the weight of the fluid per unit
volume. It is related to density (ρ) using the local gravity (g) by:
 = g
• Specific gravity (S): Specific gravity is also known as relative density. It is
ratio of density of a substance to the density of water and it is often
specified for a liquid.
 = S  water  = S  water
• The volume of a gas is known to depend on pressure and temperature. In a
liquid too, the volume depends slightly on pressure. This is quantified by
bulk modulus, B
Dr. V. Raghavan, IIT Madras 3
 Properties of fluids
• The bulk modulus is given as:
p p
B=V =
V T  T

• To cause a 1% change in volume of water (and most liquids), a pressure of

21000 kPa is needed. So it is obvious to liquids as incompressible fluids.
The modulus can also be used to determine the speed of sound in water.
c= B 
• Another property of occasional interest in our study is surface tension, σ. It
results from the attractive forces between the molecules. It allows the steel
to float, droplets to form and small droplets and bubbles to be spherical.
Dr. V. Raghavan, IIT Madras 4
 Viscosity of fluid
• Viscosity: Viscosity can be considered to be internal stickiness of the fluid.
It results in shear stresses in a flow and accounts for losses in a pipe or drag
on a rocket etc. It is related to the shear stress and velocity gradient by:
 =  ( du dr )
• du/dr is the velocity gradient, r is measured normal to the surface and u is
tangential to the surface.
• The viscosity, called the dynamic viscosity, µ, has a unit of N.s/m2.
• The kinematic viscosity, ν = µ/ρ, is a transport property and has a unit of
m2/s. A very important effect of viscosity is to cause the fluid to stick to a
surface, the no slip condition.

Dr. V. Raghavan, IIT Madras 5

 Types of fluid
• Newtonian fluid: In Newtonian fluid shear stress varies linearly with
velocity gradient.
• Non Newtonian fluid: In non Newtonian fluids, shear stress does not vary
linearly with velocity gradient.
• Dilatant fluid: A dilatant fluid or a shear thickening fluid increases
resistance to flow with increasing applied stress.
• Pseudoplastic fluid: A pseudoplastic or shear thinning fluid decreases
resistance to flow with increasing stress.
• Ideal plastic fluid: These fluid require a finite yield stress before it begins to
flow. The flow behavior after yield can be linear or nonlinear.

Dr. V. Raghavan, IIT Madras 6

 Types of fluid

• In liquids, viscosity is due to cohesive

forces and in gasses, it is due to collision
of molecules.
• In liquids, viscosity is inversely
proportional to temperature and in
gasses, viscosity is directly proportional
to temperature.

Dr. V. Raghavan, IIT Madras 7

 Types of fluid flow

(a) (b)
One dimensional flow: (a) flow in a pipe; (b) flow in a wide channel

• The flows shown above are also referred to as developed flows; the
velocity profiles do not change with respect to the downstream coordinate.
This demands that the pipe flow shown is many diameters downstream of
any change in geometry, such as an entrance, a valve, an elbow, or a
contraction or expansion.
Dr. V. Raghavan, IIT Madras 8
 Types of fluid flow
• If the flow has not developed, the velocity field depends on more than one
space coordinate, as is the case near a geometry change. The developed
flow may be unsteady, i.e., it may depend on time, such as when a valve is
being opened or closed.
• In a uniform flow, the velocity profile and other properties such as pressure
are uniform across the section of the pipe.
• Internal and External flows: Internal flows are flows that are completely
bounded by solid surfaces. Example: flow through pipe. Flows over bodies
immersed in an unbounded fluid are called external flows. Example: flow
over a sphere. Both internal and external flows may be laminar or
turbulent, compressible or incompressible

Dr. V. Raghavan, IIT Madras 9

 Viscous and inviscid flows
• In an inviscid flow, the effects of viscosity can be completely neglected
with no significant effects on the solution to a problem involving the flow.
• All fluids have some viscosity value, but if the viscous effects can be
neglected, it is an inviscid flow.
• Viscous effects are very important in pipe flows and many other kinds of
flows inside conduit. They lead to losses and require pumps in long pipes.
• Consider an external flow such as flow around an air foil. If the air foil is
moving relatively fast (>1m/s), the flow away from the thin boundary layer
can be considered as inviscid flow.

Dr. V. Raghavan, IIT Madras 10

 Viscous and inviscid flows
• All the viscous effects are concentrated inside the boundary layer and
cause the velocity to be zero at the surface of the airfoil, the no slip-
condition.

Dr. V. Raghavan, IIT Madras 11

 Laminar and turbulent flows
• A viscous flow is either a laminar flow or a turbulent flow.
• In a turbulent flow there is mixing of fluid particles so that the motion of a
given particle is random and highly irregular.
• Statistical averages are used to specify the velocity, the pressure, and other
quantities of interest.

Dr. V. Raghavan, IIT Madras 12

Laminar and turbulent flows

Dr. V. Raghavan, IIT Madras 13

 Compressible and incompressible flows
• Liquid flows are assumed to be incompressible in most situations. In such
incompressible flows the density of a fluid particle as it moves along is
assumed to be constant. The flow of gas can be assumed to be
incompressible if the velocity is sufficiently low.
• Air flow in conduits, around automobiles and small aircraft, and the takeoff
and landing of commercial aircraft are all examples of incompressible
airflows.
• The Mach number M, (V/c), is used to determine if a flow is compressible;
V is the characteristic velocity and c is the speed of sound. If M < 0.3, we
assume the flow to be incompressible

Dr. V. Raghavan, IIT Madras 14

Illustration of motion and deformation of a fluid element
• Pure rotational motion (a) and pure translational motion (b):

(a) (b)
• Pure shear motion (c):

(c)
Dr. V. Raghavan, IIT Madras 15
Illustration of motion and deformation of a fluid element
• Overall deformation due to the combined effects of rotation, translation
and shear:

At time ‘t’ At time ‘t+∆t’

Dr. V. Raghavan, IIT Madras 16

 Deformation of a fluid element
• Overall deformation due to the combined effects of rotation, translation
and shear:

After time ‘t+∆t’

At time ‘t’

Dr. V. Raghavan, IIT Madras 17

 Deformation of a fluid element
• Let velocity components at vertex A at time ‘t’ be u and v along x and y
coordinate respectively. After deformation, the velocity components of
vertices A’, B’, C’ and D’ respectively, taking order of ∆x and ∆y into account
are given as:
( ut , vt ) ,
 u v 
 u + x, v + x  ,
 x x 
 u u v v 
 u + x + y, v + x + y  ,
 x y x y 
 u v 
 u + y, v + y 
 y y 
Dr. V. Raghavan, IIT Madras 18
 Vorticity in terms of velocity gradients
• Rotation of the fluid element is quantified by the angular velocity about
each coordinate axis. The angular velocity of the element (or face) about
the z-axis can be evaluated as the arithmetic mean of the rates of rotation
of edges AD and AB. Thus,
1  d d  
z =  − 
2  dt dt 
v u
xt yt
x v y u
 = = t  = = t
x x y y

1  v u   z
 z =  −  =
2  x y  2
Dr. V. Raghavan, IIT Madras 19
 Estimation of strain rates
• Shear strain rate of the fluid element is defined as the average rate of
decrease of the angle between two initially perpendicular lines (AB and
AD). Thus,
1  d d   1  v u 
 xy =  yx =  + =  + 
2  dt dt  2  x y 
1  w v 
 yz =  zy =  + 
2  y z 
1  u w 
 zx =  xz =  + 
2  z x 

Dr. V. Raghavan, IIT Madras 20

 Dilatational strain
• Dilatational strain of a fluid element is the rate of change of length along
each coordinate direction divided by the original length. The dilatational
strain along the x-direction is given as,
u
x + x t −  x
1 x u
 xx = =
t x x
v w
 yy = ;  zz =
y z   xx  xy  xz 
• The strain rate tensor is finally given as:  =   yx  yy  yz 


  zx  zy  zz 


Dr. V. Raghavan, IIT Madras 21

 Strain rate tensor

 u 1  v u  1  u w  
  +   + 
 x 2  x y  2  z x  
 1  v u  v 1  w v  
 =  +   + 
 2  x y  y 2  y z  
 
 1  u + w  1  w v  w 
 2  z x   +  
 2  y z  z 

Dr. V. Raghavan, IIT Madras 22

 Governing equations
• Differential form of continuity equation:
D  u v w  D
+ + +  = 0  +   V = 0
Dt  x y z  Dt
• Differential form of momentum equation:
 xx  xy  xz Du
+ + +  gx = 
x y z Dt
 xy  yy  yz Dv
+ + + gy = 
x y z Dt
 xz  yz  zz Dw
+ + +  gz = 
x y z Dt
Dr. V. Raghavan, IIT Madras 23
 Governing equations
• The stresses can be expressed in terms of velocity and pressure fields and
viscosity as:
u  u v 
 xx = − p + 2  +   V  xy =   + 
x  y x 
v  u w 
 yy = − p + 2 +   V  xz =   +
y 
 z x 
w  v w 
 zz = − p + 2  +   V  yz =  +
z 
 z y 

Dr. V. Raghavan, IIT Madras 24

 Navier-Stokes equation
• If an incompressible flow is assumed so that ▽.V = 0, the Navier-Stokes
equation is obtained which is given as:

Du p   2u  2u  2u 
 = − +   2 + 2 + 2  +  gx
Dt x  x y z 
Dv p   2v  2v  2v 
 = − +   2 + 2 + 2  +  gy
Dt y  x y z 
Dw p  2w 2w 2w 
 = − +   2 + 2 + 2  +  gz
Dt z  x y z 

Dr. V. Raghavan, IIT Madras 25

 Flow through pipe
• For a laminar flow in a pipe, the development of the velocity profile is as
shown:

r
x

Dr. V. Raghavan, IIT Madras 26

 Flow through pipe
• Considering the Navier-Stokes equation in cylindrical coordinates:
   
    2u 1 u 1  2u
 u v u u u  p  2u 
  vr + +u +  = − +  gx +   2 + + 2 + 
r r  x t x  r r r r  2
x 2
 no no developed steady   developed 
 velocity flow 
symmetric

radial swirl flow
  flow flow 
p   2u 1 u  p    u 
0= − + 2 + 0=− + r 
x  r r r  x r r  r 
Integrating the above equation twice w.r.t r, we get:
r 2 p
u= + A ln r + B
4  x

Dr. V. Raghavan, IIT Madras 27

 Flow through pipe
• In the above equation, when r = 0, A ln r, is not defined. Therefore, for u to
be bounded, A = 0.
• Then, at r = R (the radius of the pipe), u = 0. Applying this boundary
condition, we get:
R 2 p
B=−
4  x

• Applying these conditions in the equation, we get parabolic velocity profile

given as:
1 p 2
u=
4  x
r − R2 ( )

Dr. V. Raghavan, IIT Madras 28

 Flow through pipe
• The entrance length (Le) is given by: LE/D = (0.065)Re
• The shear stress profile and the wall shear stress is given as:
u
 =
r
  1 p 2 2 
 =  
r  4  x
(
r −R  )

1 p
 = ( 2r )
4 x
1  p 
  wall =  r = R   wall =   R
2  x 

Dr. V. Raghavan, IIT Madras 29

 Flow through pipe
• The pressure drop (∆p) and the wall shear stress (τwall) for flow through a
pipe of radius R over a length L is obtained as:
8 LU avg Rp
p = 2
 wall =
R 2L

• The head loss due to friction (hL), is given as:

p fLU avg
2

hL = =
g 2 gD
• The friction factor (f) for laminar and turbulent flow is given as:
64 0.316
Laminar: f = Turbulent: f = 0.25
Re Re
Dr. V. Raghavan, IIT Madras 30
 Head losses in flow through pipe
• Major head loss comprises of loss in head due to viscous effects in straight
pipes. This is nothing but hL.
• Minor head losses comprises of losses in head due to viscous effects in pipe
components. Pipe components affect the flow by changing the direction of
flow, obstructing the flow, changing the flow speed through varying cross
section etc. These will cause flow separation leading to formation of eddies
which will result in loss in head.
• Some typical components in the pipe system includes bends, elbows, tees,
valves etc.
• Sudden or gradual expansion and contractions, entrance into pipe from
reservoirs and exit from pipe into reservoir also causes loss in head.
Dr. V. Raghavan, IIT Madras 31
 Minor head losses in flow through pipe

• These losses are expressed in terms of a loss co-efficient K, and the minor
losses are defined as:
V2
hL ,minor =K
2g
• The value of V to be used is specified in the table depending on the
condition.
• The value of K for various cases are measured experimentally and are
tabulated. The most commonly used table to determine the minor losses is
the ‘CRANE Technical Paper 410, Flow of Fluids, Through Valves, Fittings
and Pipes’.
Dr. V. Raghavan, IIT Madras 32
 Minor head losses in flow through pipe

Flow through a 90° bend

Sudden expansion

Sudden contraction
Dr. V. Raghavan, IIT Madras 33
 Flow through bundle of tubes
• Tube banks/bundles are commonly employed design elements in heat
exchangers. Both plain and finned tube banks are widely found.
• Two basic tube-bank patterns used are in-line tube banks and staggered
tube banks.
#

In-line tube banks Staggered tube banks

# - S.B. Beale, Cross flow over tube banks,
Dr. V. Raghavan, IIT Madras 34
DOI: 10.1615/AtoZ.t.tube_banks_crossflow_over
 Flow through bundle of tubes
• They are characterized by crosswise pitch (ST) to diameter ratio and
streamwise pitch (SL) to diameter ratios, a and b, which are given as:
ST SL
a= b=
D D
• The flow Reynolds number is given as:
 umax D where ūmax is the bulk velocity in the minimum
Re =
 cross section and D is the cylinder diameter
• The overall pressure loss coefficient is expressed in terms of Euler number
(Eu) as:
prow where ∆prow refers to mean pressure drop across a
Eu =
0.5  umax
2
single row
Dr. V. Raghavan, IIT Madras 35
 Flow through bundle of tubes
• From experiments it is found that Eu = f(Re, a, b)
• Using the flow Reynolds number (Re) and the ratio of (a/b), a constant k1
can be determined from empirical charts. Using this, (Eu/k1) can be found
out from a typical graph of (Eu/k1) vs. Re.
• Using Eu, the pressure drop across the bundle of tubes can be evaluated
as:
 umax
2
p = Eu N , where N refers to number of rows of tubes
2

Dr. V. Raghavan, IIT Madras 36

 Types of finned tubes
# #

#
#

Circular fins Spiral fins

Axial fins Plate fins

# - S.B. Beale, Cross flow over tube banks,
Dr. V. Raghavan, IIT Madras 37
DOI: 10.1615/AtoZ.t.tube_banks_crossflow_over
 Pumping power
• Given the density of the fluid (ρ), volumetric flow rate (Q), the pumping
head (HP) and efficiency (η), the pumping power required is evaluated as:
 gQH P
WP =

• The pumping power can also be evaluated using the formula:
WP = Q ( p )
where ∆p is evaluated as:
(
p =  ghL =  g fLU avg 2 2 gD )

Dr. V. Raghavan, IIT Madras 38

 Pumping power
• When pumps are connected in parallel, the net output flow rate Qnet will be
sum of flow rates of individual pumps connected.
Qnet = Q1 + Q2 + Q3 + ...... + Qn

• When pumps are connected in series, the net pumping head HP,net will be
sum of pumping heads of individual pumps connected.
H P , net = H P ,1 + H P ,2 + H P ,3 + ...... + H P ,n

Dr. V. Raghavan, IIT Madras 39

 Similitude
• Prior to construction of prototype, a model is built and tests are conducted
on it. The results from these tests will hold true for the prototype only if it
conforms with the geometric, kinematic and dynamic similarity.
• Geometric similarity: All the length ratios of the model (LM) to the protype
(LP) should be a constant. It is given as:
LM
=C
LP
• Kinematic similarity: All the velocity ratios of the model (VM) to the protype
(VP) should be a constant. It is given as:
VM
=C
VP
Dr. V. Raghavan, IIT Madras 40
 Non dimensional numbers used in similarity analysis
• Dynamic similarity: All the force ratios of the model (FM) to the protype (FP)
should be a constant. It is given as:
FM
=C
FP
• Reynolds number: This is used to analyze flow in a closed conduit pipe.
Finertia VL
Re = =
Fviscous 
where ρ is the density of the fluid, V is the flow velocity, L is the
characteristic length and μ is the viscosity of the fluid

Dr. V. Raghavan, IIT Madras 41

 Non dimensional numbers used in similarity analysis
• Froude number: This is used where a free surface is present (eg: spillway,
channels etc.) where gravity force is predominant.
Finertia V
Fr = =
Fgravity gL
where V is the flow velocity, g is the acceleration due to gravity L is the
characteristic length
• Euler number: This is used in cavitation studies.
Finertia V where V is the flow velocity, p is the
Eu = = local pressure and ρ is the density of
Fpressure p
 the fluid
Dr. V. Raghavan, IIT Madras 42
 Mach number and Weber number

• Mach number: This is used in applications where the compressibility of the

fluid plays an important role.
Finertia V where V is the flow velocity
M= = and C is the speed of sound
Felastic C
in that medium.
• Weber number: This is mainly used in capillary studies.
where V is the flow velocity, σ is the
Finertia V
We = = surface tension (units: N/m), ρ is the
Fsurface tension  density of the fluid and L is the
L characteristic length

Dr. V. Raghavan, IIT Madras 43

 Similitude: Model laws for turbines and pumps
• Similarity of model (m) and prototype (p) turbines or pumps are based on
the assumption that the efficiency of the model is equal to that of the
prototype.
• If Q is the volumetric discharge (m3/s), N is the speed of rotation of pump
or turbine (rpm), D is the runner diameter (m), H is the head of water
under which the turbine operates or the head against which the pump
operates (m) and P is the power developed by the turbine or the power
required by the pump (in kW), then the three parameters are defined as:
1. Flow coefficient (CQ):
Q  Q   Q 
CQ =  3 
= 3 
 ND  m  ND  p
3
ND
Dr. V. Raghavan, IIT Madras 44
 Similitude: Model laws for turbines and pumps
2. Head coefficient (CH):
H  H   H 
CH = 2 2   2 2  =  2 2 
N D  N D m  N D  p
3. Power coefficient (CP):
P  P   P 
CP = 3 5   3 5  =  3 5 
N D  N D m  N D  p

Dr. V. Raghavan, IIT Madras 45

 Turbulence
• Unsteady, irregular (chaotic) motion in which transported quantities
(mass, momentum, scalar species) fluctuate in time and space.
• Fluid properties exhibit random variations. Statistical averaging of results
is accountable.
• Turbulence motion carries a wide range of eddy sizes (scales).
• They are typically identifiable swirling patterns where larger eddied
‘carry’ smaller eddies
• Larger, higher energy eddies transfer energy to smaller eddies via vortex
stretching.

Dr. V. Raghavan, IIT Madras 46

 Turbulence
• Larger eddies derive energy from mean flow.
• Smallest eddies convert kinetic energy into thermal energy via viscous
dissipation.
• The rate at which the energy is dissipated is set by the rate at which they
receive energy from the larger eddies at the start of the energy cascade.
• The three eddy scaled (from largest to smallest) are:
1. Integral scale
2. Taylor microscale
3. Kolmogorov scale

Dr. V. Raghavan, IIT Madras 47

 References
1. F.M. White, Fluid Mechanics, fourth edition, WCB McGraw Hill
Publication
2. V. Babu, Fundamentals of Incompressible Fluid Flow, Anne Books
Private Limited
3. R.W Fox, A.T. McDonald, P.J. Pritchard, Introduction to Fluid Mechanics,
John Wiley and Sons, Inc.
4. S.B. Beale, Cross flow over tube banks,
DOI: 10.1615/AtoZ.t.tube_banks_crossflow_over
5. M. Potter, D.C. Wiggert, Schaum’s Outlines Fluid Mechanics,
McGraw Hill Publications

Dr. V. Raghavan, IIT Madras 48