MEC551

THERMAL ENGINEERING

3.0 Convection

1
 3.0 SYLLABUS CONTENT

3.1` Convection Principles – heat transfer coefficient

3.2 Convection boundary layer theory
3.3 Forced convection over exterior surface (laminar
flow)
3.4 Forced convection in turbulent flow (Reynold’s
analogy.
3.5 Principle of dynamic similarity applied to forced
convection
3.6 Free convection and laminar profile over vertical
plates
2
 Convection Principles
(Nusselt Number)

• Nusselt Number
– Developed by Wilhelm Nusselt
(1882-1957) from Germany
– In convection analysis, it is
common practice to non-
dimensionalized the governing
equations and combine the
variables, which group together in
dimensionless numbers – to reduce
the number of variables.

3
 Convection Principles
(Nusselt Number)

– The Nusselt number is a non-dimensionalized h,

defined as:

hLc
Nu  Lc - Characteristic Length
k k - Thermal conductivity of fluid

4
 Convection Principles
(Nusselt Number)

• Since:
– Heat transfer by conduction occurs when the fluid
is motionless and …
– Heat transfer by convection occurs when the fluid
involves some motion.
– In either case, the heat flux is the rate of heat
transfer per unit time per unit surface area.

q conv  h  T
T
q cond k
L 5
 Convection Principles
(Nusselt Number)
• Taking the ratio of these two equations:

q conv h  T h  L
 k T   Nu
q cond L k
• Thus Nu represents the enhancement of heat transfer through a
fluid layer as a result of convection relative to conduction
across the same fluid layer. The larger Nu, the more effective
the convection.

• Nu = 1 for a fluid layer, represents pure conduction.

6
 Convection Principles
(Viscosity)

• Dynamic viscosity () – The shear force per unit area

required to drag on layer of fluid with unit velocity passed
another layer a unit distance away from the fluid.


 du 
 dy 
• Kinematic viscosity () – The ratio of dynamic viscosity
to density.



7
3.2 Convection boundary
layer theory

8
 Convection Principles
(Velocity Boundary Layer)

• Velocity boundary development on a flat plate:

• The boundary layer thickness (d) is normally defined

as where:
u  0.99  u 9
 Convection Principles
(Velocity Boundary Layer)

• The dashed line, divides the flow

y
over the plate into two regions: u∞
Inviscid
• Boundary layer region Flow
– In which the viscous effects and
velocity changes are significant. Boundary
Layer
• Inviscid flow region
x
– In which the friction effects are
negligible and the velocity Heated Surface

remains constant.

10
 Convection Principles
(Velocity Boundary Layer)

• Flow regions in velocity boundary of a flat plate:

11
 Convection Principles
(Velocity Boundary Layer)

• Comparison of a laminar and turbulent velocity

boundary layer profile:

12
 Convection Principles
(Thermal Boundary Layer)

• Likewise there is a thermal

boundary layer
y T∞
• No temperature jump condition
– Because velocity of the fluid x
is zero at the point of contact Heated Surface
with the solid surface, the
fluid and solid surface must
have the same temperature
at the point of contact.

13
 Convection Principles
(Thermal Boundary Layer)

• Thermal boundary development on a flat plate:

Ts+0.99·(T-Ts)

• The thickness of the thermal boundary layer (dt) at any location

along the surface is defined as the distance from the surface at
which:

dT=T-Ts=0.99·(T-Ts) 14
 Convection Principles
(Prandtl Number)

• Prandtl Number
– Developed by Ludwig Prandtl (1875-1953) of
Germany.
– The relative thickness of the velocity and
thermal boundary layers is best described by
a dimensionless Prandtl number (below):

Molecular Diffusivit y of Momentum

Pr 
Molecular Diffusivit y of Heat
  Cp
 
 k
15
 Prandtl Number ?

• The Prandtl numbers of gases are about 1, which

indicates that both momentum and heat dissipate
through the fluid at about the same rate.

• Heat diffuses very quickly in liquid metals (Pr << 1) and

very slowly in oils (Pr >> 1) relative to momentum.

• Consequently the thermal boundary layer is much thicker

for liquid metals and much thinner for oils relative to the
velocity boundary layer,
16
 Convection Principles
(Reynolds Number)

• Reynolds Number
– Derived by Osbourne Reynolds (1842-1912)
of Britain

– The transition from laminar to turbulent flow

depends on the surface geometry, surface
roughness, free stream velocity, surface
temperature, and type of fluid (among other
things).
– However, the flow regime primarily depends
upon the ratio of inertia forces to viscous
forces in a fluid. This is a dimensionless
quantity, known as Reynolds number (Re).
17
 Convection Principles
(Reynolds Number)

• The Reynolds number is defined as:

Inertia Forces V  L   V  L
Re   
Viscous Forces  

V – upstream velocity
L – characteristic length
n = m/r – kinematic viscosity of fluid

18
 Convection Principles
(Reynolds Number)

Inertia Forces V  L   V  L
Re   
Viscous Forces  

• A large Re (inertia forces large)

– Means that the viscous forces cannot contain random and
rapid fluctuations (turbulent).

• A small Re (viscous forces large)

– Keeps the fluid in-line (laminar).
• The Reynolds number where the flow becomes turbulent is
called the critical Reynolds number (Recrit)

19
 Convection Principles
(Reynolds Number)

• For flow over a flat plate, the generally accepted

value of Recrit is:

  u  xcrit
Flat Plate: Recrit   5 105


where: xcrit= Distance between the leading edge

of the plate to the transition point from laminar to
turbulent flow takes place.

20
3.3 Forced convection over an
exterior surface
(laminar and turbulent flow)

21
 External Flow
• The convection equations for an external flow can be
derived from the conservation of mass, conservation
of energy, and the conservation of momentum
equations.

22
 External Flow Equations
(Conservation of Mass)

• Conservation of Mass

m x    u  dy  1
Unit Area dv
m y    v  dx  1 v   dy
dy
Unit Area

dx
u
u dy u   dx
x

v 23
 External Flow Equations
(Conservation of Mass)

Rate of mass Rate of mass

flow into = flow out of
control volume control volume

 u   v 
  u  dy    v  dx     u  dx   dy     v  dy   dx
 x   y 
u v
  u  dy    v  dx    u  dy     dx  dy    v  dx     dx  dy
x y
u v
 0 ~ 2-D Continuity Equation
x y 24
 External Flow Equations
(Conservation of Momentum)

• Conservation of Momentum

  dy
Σm·a = Net Force y

P
P P dx
x
dy
dx

 25
 External Flow Equations
(Conservation of Momentum)
• In the x-direction:

 u u  P   2u  2u 
   u  v        2  2     g x
 x y   x x y   
       Body force
       Net
 
du per unit
Net effect of viscous
pressure volume
force and shear forces
• In the y-direction:

 v v  P   2v  2v 
   u  v        2  2     g y
 x y   y  x y    
           Body force
dv Net Net effect of viscous per unit
pressure and shear forces volume
force 26
 External Flow Equations
(Conservation of Energy)

• Conservation of Energy
Eheat out, y Emass out, y
E in  E out  0

Eheat in, x dx Eheat out, x

dy
Emass in, x Emass out, x

Eheat in, y Emass in, y 27

 External Flow Equations
(Conservation of Energy)

• General 2-D energy equation

 T T   T  T 
2 2  2
 u   v 
2
  u v 
2

 Cp u
 v 
  
 k  2  2   2              
 
 x y   x y   x   y    y x 

• For 2-D inviscid flow:

 T T    2T  2T 
  C p  u v   k  2  2 
 x y   x y 

28
 Convection over a Flat Plate
T, u Boundary layer

y dy
dx
x
u(x,0)= 0
v(x,0)= 0
T(x,0)= Ts

• Consider laminar flow over a flat plat. When viscous

dissipation is negligible, the convection equations
reduce for steady, incompressible laminar flow (with
constant properties) over a flat plate.
29
 Convection over a Flat Plate
• Consider elemental control volume for force balance
in the laminar boundary layer.

u v
Continuity:  0
x y

u u  2u
Momentum: u v  2
x y y

T T  2T
Energy: u v  2
x y y
30
 Convection over a Flat Plate
• Boundary conditions:

– At x= 0: u(0,y)= u, T(0,y)= T

– At y= 0: u(x,0)= 0, v(x,0)= 0, T(x,0)= Ts
– At y= : u(x,)= u, T(x,)= T

• Define a dimensionless similarity variable:

u
y
 x

31
 Convection over a Flat Plate
• Recall, that the stream function is defined as:

 
u ; v
y x
• Dependent variable:

  
f    
  x u  y
u
u

32
 Convection over a Flat Plate
• Therefore:

     x df u df
u    u    u
y  y u d   x d

     x df u 
v    u    f
x  x u dx 2 u x
1 u   df 
    f 
2 x  d 

33
 Convection over a Flat Plate
• So:
u u  2
d f 
     2 
x 2 x  d 

u u  d2 f 
 u   2 
y  x  d 

 2u u2 d 3 f
  3
y 2
  x 
34
 Convection over a Flat Plate
(Momentum Equation)
• Substituting these into the momentum equation and simplifying
gives:
d3 f d2 f
2 3f 0 EQN 6-49
d d 2
text

• A 3rd order non-linear differential equation. Therefore the system

of partial differential equations is transformed into a single
ordinary differential equation by use of a similarity variable.

35
 Convection over a Flat Plate
(Momentum Equation)
• Using the definitions for f and η, the boundary equations in
terms of the similarity variables can be found.

f  0  0 • However, the transformed equation

with its similarity variable cannot be
df solved analytically.
0
d   0
• Therefore, an alternative solution is
df necessary.
1
d   

36
 Convection over a Flat Plate
(Momentum Equation)
• The non-dimensional velocity profile can be obtained by
plotting u/u vs. η. The results agree experimentally.

df u
• A value of:   0.992 corresponds to:   5.0
d u
• Recall that the definition of a velocity boundary layer is when:

u
 0.99
u

37
 Convection over a Flat Plate
(Momentum Equation)
• So substituting these values into the definition for η, gives the
boundary layer thickness for a flat plate.
  5.0; y  

u
y
 x
u
5
 x
For laminar 5.0 5.0  x u x
  where : Re 
flat plate: u Re 
EQN 6-51
 x text 38
 Convection over a Flat Plate
(Energy Equation)

Energy Equation
• Knowing the velocity profile, we can now solve the energy
equation.
• Introduce dimensionless temperature:

T  x, y   Ts
  x, y  
T  Ts

• Note: both Ts and T are constant.

39
 Convection over a Flat Plate
(Energy Equation)
• Substituting θ into the energy equation gives:

   2
u v  2
x y y
• Again using the similarity variable, η, so θ= θ(η)

u
y
 x
• So the energy equation becomes:

2
df d d 1 u   df  d d d   d  2
u      f    
2 

d d dx 2 x  d  d dy d  dy 40
 Convection over a Flat Plate
(Energy Equation)

u
• Since:   y
 x
3

d 1  u  2 1  u
   y      y
dx 2   x  2 u   x
d u

dy  x

41
 Convection over a Flat Plate
(Energy Equation)

  
• and: f    
  x u  y
u
u
df 

d u   y

42
 Convection over a Flat Plate
(Energy Equation)
• Substituting these in gives:

2
df d d 1 u   df  d d d   d 
2
 u       f    
2 

d d dx 2 x  d  d dy d  dy 

   d  y  u  1 u   u     d u
 u        
 y     
 u y  d  2u   x  2 x    x u  y u   d
y  x
d 2  u 
  
d 2    x 

43
 Convection over a Flat Plate
(Energy Equation)

1 d   u  u  u 1 u    u
  
      
2 d  u   x  x   x u 
x u  y    x
d 2  u 
 2  
d    x 

d   u  1 u     x d 2
        2
d u
    x  x   x xy u
  d 2

44
 Convection over a Flat Plate
(Energy Equation)

d    x u  u    d 2
      2
d  u    x u    x u y
  d  2

Prandtl number  d    d 2
   2 2
Pr    d  u y  d
 Pr
f

d 2 d
2 2  Pr  f  0 EQN 6-58
d d text

45
 Convection over a Flat Plate
(Energy Equation)
• A closed form solution cannot be obtained for this boundary
layer problem, and it must be solved numerically.

• If this equation is solved for numerous values of Pr, then for

Pr > 0.6, the non-dimensional temperature gradient at the
surface is found to be (reference Table 6-3, p. 378 in text):

d 1
 0.332  Pr 3
d   0

46
 Convection over a Flat Plate
(Energy Equation)
• The temperature gradient at the surface is:

T   
  T  Ts     T  Ts  
y y 0
y y 0
  0 y y 0

u  u
• Since: y 
then: 
 x y  x

• Therefore substituting these values in gives:

T u
 0.332  Pr  T  Ts  
1
3

y y 0
 x 47
 Convection over a Flat Plate
(Energy Equation)
• Therefore the local convection coefficient and Nusselt number
become:

hx 
q s

k  T
y y  0   k  0.332  Pr 1
3
  Ts  T   u
 x 
Ts  T Ts  T Ts  T

u
hx  0.332  Pr  k
1
3

 x

48
 Convection over a Flat Plate
(Laminar Flow)
• The local Nusselt number is the dimensionless temperature
gradient at the surface. This is defined as:

hx  x
Nu x 
k

• Thus for Pr > 0.6, the local Nusselt number for laminar flow is:

1
Nu x  0.332  Pr  Re
1
3 2

49
 Convection over a Flat Plate
(Laminar Flow)

• The local friction coefficient (CFx) can also be determined.

• Since the wall shear stress is:

u u d f2
 wall      u   2
y y 0
  x d  0

• From Table 6-3 (pp. 378 in text) this is found to be:

0.332    u2
 wall 
Re x
50
51
 Convection over a Flat Plate
(Laminar Flow)
• Therefore the local skin friction coefficient is:

1
 wall   C F , x    u2
2

2  wall  12
CF , x   0.664  Re x
  u2

52
 Convection over a Flat Plate
(Laminar Flow)
• The average heat transfer coefficient over the entire plate can be
obtained by integrating over its length:
L L
0.332  Pr  k
1
1 3
u
h   hx  dx h  dx
L0 L 0
 x

 
1
0.332  Pr 3  k u L
   2 x
L  0
1
0.664  Pr 3  k u  L
 
L 
1 1
0.664  k  Pr  Re 3 2


L 53
 Convection over a Flat Plate
(Laminar Flow)

• So the average Nusselt number for laminar flow over

the entire plate is:

hL 1
Nu   0.664  Re L  Pr 3
0. 5

54
 Convection over a Flat Plate
(Laminar Flow)

• Solving numerically for temperature profile for

different Prandtl numbers, and using the definition of
the thermal boundary layer, it is determined that for
laminar flow over a flat plate:

  13  13
t   Pr    Pr
1.026

55
 Convection over a Flat Plate
(Example 3.1)

• Example 3.1a Calculate the heat transfer and the

thermal boundary layer thickness ¼ of the way along
a flat plate that is 50 m long. Liquid (Tsat = 40 ºC)
flows over it at 4 m/s. The plate is kept at a surface
temperature (Ts= 80 ºC).
Ts= 80ºC
40 ºC
4 m/s
y
x
50 m
56
 Convection over a Flat Plate
(Example 3.1)
40 ºC
y
80 ºC

x
50 m

• The first step is to calculate the mean film temperature of the

fluid flowing along the plate.
• This is just the average of the surface temperature and the fluid
bulk temperature.

Ts  T 80C  40C
T film    60C
2 2
57
 Convection over a Flat Plate
(Example 3.1)
40 ºC
y
80 ºC

x
50 m

• For liquid water at 60 ºC from Table

  983.3 kg
m3

  4.67 kg
m s

k  0.654 W
m C

Pr  2.99 58
 Convection over a Flat Plate
(Example 3.1)
• First calculate the Reynolds number to determine whether the
flow is laminar or turbulent.

 x  983.3    4     50 m 
kg
  u m3
m
s
1
4
Re    10,527.8
 4.67 kg
m s

• Since Re < Recrit = 5x105 or 500,000 ~ Flow is laminar

 
• Therefore:
5 x 5  504m
     0.609 m
Re 10,527.8

t 
  Pr

 0.609 m    2.99 
 13  13
 0.412 m
1.026 1.026 59
 Convection over a Flat Plate
(Example 3.1)
• Example 3.1b Now calculate the convective heat transfer.
• First we must check to see whether the entire plate is in a
laminar boundary layer or not.

 L  983.3    4    50 m 
kg
  u m3
m
s
Re    42,111 .3
 4.67 kg
m s

• Since Re < Recrit = 5x105 or 500,000 ~ Flow is laminar over the

entire plate

60
 Convection over a Flat Plate
(Example 3.1)
• Therefore we can use the following equation to find h:

u u  
h  0.332  Pr  k  0.332  Pr  k
1 1
3 3

 x x

 0.332   2.99    0.654 mWC  

1  4 m
s    983 . 3 kg
m3 
 4.67 mkgs    50 m
3

 0.619 W
2
m  C

61
 Convection over a Flat Plate
(Example 3.1)
• Using this h, we can now find the convection heat transfer:

q  h  (Ts  T )

 0.619 W
m 2  C
   80C  40C 
 24.8 mW2

62
Convection over a Flat Plate
(Turbulent and Mixed Flows)

Completely Turbulent Flow

Turbulent

Mixed Laminar/Turbulent Flow

63
64
 Convection over a Flat Plate
(Turbulent and Mixed Flows)
• Note: if it had been found that the boundary layer was not
completely laminar another equation for h could have been
used instead.
• For turbulent flow (all over the plate):

1 0.6  Pr  60
Nu  0.037  Re  Pr0.8
L
3

5 105  Re  107

• For a mixed combination of laminar and turbulent flow over the

plate:

Nu   0.037  Re  871  Pr
0.8 1 0.6  Pr  60
3
L 5  105  Re L  107
65
 Convection over a Flat Plate
(Example 3.2)

• Example 3.2 Oil flows over a 40-m long heated plate at free
stream conditions of 5 m/s and 25ºC. If the plate is held at 45ºC.
Ts= 45ºC
u= 5 m/s
T= 25ºC

40 m

a) Determine the velocity and thermal boundary layer

thicknesses at the middle of the plate.

b) Calculate the distance where the laminar change to

turbulence flow
66
 Convection over a Flat Plate
(Example 3.2)

• First calculate the film temperature (Tf)

T  Ts 25C  45C
T film    35C
2 2
• From Tables for oil at 35ºC, the fluid properties are:

Pr  3,711
  3.5 10 4 m2
s

k  0.2864 W
m C

  1,255 m3 kg

67
 Convection over a Flat Plate
(Example 3.2)

a) At the middle of the plate:

40 m
x  20 m
2
u  x  5 ms    20 m 
Re mid    2 . 86  10 5

po int  3.5  10 s 4 m2

Since the critical Reynolds number is 5x105, then:

Re mid  Recrit
po int

The flow at the mid-point of the plate is laminar.

68
 Convection over a Flat Plate
(Example 3.2)

• The hydrodynamic (or velocity) boundary layer is:

5 x 5   20 m 
 x  20    0.187 m or 18.7 cm
Re 2.86  10 5

• The thermal boundary layer is:


t   Pr 
1
3

1.026
0.187 m
  3,711  0.0118 m or 11 .8 mm
 1
 3

1.026
69
 Convection over a Flat Plate
(Example 3.2)

b) At the end of the plate:

u  L  5 ms    40 m 
Reend    5 .714  10 5

 3.5 10 s 4 m2

Since Re > Recrit the flow is turbulent at the end

• The critical distance (transition point from laminar to

turbulent is:
Recrit 
xcrit 
u
 5 10   3.5  10
5 4 m2
  35 m
 s
70
5 m
s
 Convection over a Flat Plate
(Example 3.2)

c) Using the mixed Nu equation for a flat plate:

 
Nu  0.037  Re  871  Pr
0 .8 1
3

  0.037   5.7110  
L
5 0.8
 871   3,711
1
3

 9,600.7

Nu  k
h 
L
 9,600.7    0.2864 W
  68.7
 m C W
m 2  C
40 m
71
 Convection over a Flat Plate
(Example 3.2)

• The total heat transfer per is:

Q  h  As   Ts  T 

 68.7 W
m 2  C
   40 m 1 m   45C  25C 
 54,960 W

72
• Example 7-1 7-2, 7-3 pp 404-407

73
 Forced Convection
(on Cylinders and Spheres)

• Flows across cylinders and

spheres, in general, involve
flow separation which is
difficult to handle analytically.
• Thus these must be studied
empirically or experimentally
• Several correlations have
been developed for the heat
transfer coefficient (h).

74
 Forced Convection
(on Cylinders and Spheres)

• Churchill and Bernstein developed this empirical

equation for flow over a cylinder (Eqn. 7-35 in text):
4

0.62  Re  Pr   Re  
1 1 5 5
hD 2 3 8

Nucyl   0.3   1    
k 
1   0Pr.4  3
2
1


4
  282,000  

• Whitaker developed this empirical equation for flow

over a sphere (Eqn. 7-36 in text):
1

  0. 4    
4
hD
Nusph   2  0.4  Re 2  0.06  Re 3  Pr   
1 2

k  s 
75
 Forced Convection
(over Circular and Non-Circular Cylinders)
• Additionally the following empirical correlations have been made
by Zukauskas and Jakob for the average Nusselt number for flow
over circular and non-circular cylinders (Table 7-1 in text):

76
 Forced Convection
(over Circular and Non-Circular Cylinders)

77
 Forced Convection
(Example 3.3)

• Example 3.3 A long 10-cm diameter hexagonal steam pipe

whose external surface temperature is 110°C passes through
some open area that is not protected against the wind.
Determine the rate of heat loss when the air is at 1 atm
pressure and 10°C and the wind is blowing across a 1-m length
of pipe at a velocity of 8 m/s.

V = 8 m/s
Ts=110°C
T = 10°C

10 cm

78
1m
 Forced Convection
(Example 3.3)

• The properties of air at the average film temperature

of:
Ts  T 110 C  10C
T film    60C
2 2
can be found from Table A-15 as:

k  0.02808 W
m C ; Pr  0.7202
  1.896  10 5 m 2
s

79
 Forced Convection
(Example 3.3)

• The Reynold’s number is:

V  D  8 ms    0.10 m 
Re    4 .219  10 4

 1.896 10 s 5 m 2

• The Nusselt number can be determined from Table 7-1

in the text book:

Nu  0.153  Re  Pr
0.638 3


 0.153  4.219 10 4 0.638
   0.7202 
1
3

 122.5
80
 Forced Convection
(Example 3.3)

• Therefore:
k
h  Nu
D
0.02808 mWC
  122.5  34.4 W
m 2  C
0.10 m
• The surface area of the hexagon is: 60°
D/2

D
As  6  L
2  sin  60
3   0.10 m   1 m 

sin  60
81
 0.346 m 2
 Forced Convection
(Example 3.3)

• Therefore, the heat transfer is:

Q  h  As   Ts  T 

 34.4 W
2
m  C
   0.346 m   110 C  10C 
2

 1,191.7 W

82
• Example 7-5 pp 414

83
3.4 Principle of dynamic similarity
and dimensional analysis
(applied to forced convection)

84
 Non-dimensionalized
convection equations
• The continuity , momentum, and energy equations for steady,
incompressible, laminar flow of a fluid with constant properties
can be non-dimensionalized by dividing all the dependent and
independent variables, as follows:

x y
x  ; y  ;
* *

L L
u v
u 
*
; v 
*

V V Free stream velocity

P T  Ts
P 
*
; T 
*

 V2
T  Ts Surface temperature
Free stream temperature
85
• Note: the asterisks denote non-dimensional variables.
 Non-dimensionalized
convection equations
• Introducing these variables the equations become:

u * v *
• Continuity:  * 0
x y
*

u *
u *
1  2 *
u dP *
• Momentum: u * *  v* *   *2  *
x y Re L y dx

• Energy: T *
T *
1  2
T
u *
v *
  *2
x *
y *
Re L  Pr y
86
 Non-dimensionalized
convection equations
• For a plate, the boundary conditions are:

u  0, y   1
* *
v  x ,0   0
* *
T  0, y   1
* *

u  x ,0  0
* *
T  x ,0   0
* *

u  x ,   1
* *
T  x ,   1
* *

u, T

y*
Ts
x*
87
 Similarity
• Where:
V  L 
Re L  Pr 
 

• For a given geometry, the solutions of problems with

the same Re and Nu are similar, thus Re and Nu are
called similarity parameters.
• Two physical phenomena are similar if they have the
same dimensionless forms of the governing
differential equations and boundary conditions.
88
 Similarity
• A major advantage in non-dimensionalizing is the
significant reduction in the number of similarity
parameters.
– Original equations have 6 parameters: (L, V, T, Ts,
a, and n)

– The non-dimensionalized equations have only 2

parameters (ReL and Pr).

89
 Similarity
• For a given geometry, problems that have the same
values of similarity parameters (ReL and Pr) have
identical solutions.

90
Fig 6-28 (text)
 Similarity
• Example: Determining the convection heat transfer
coefficient (h) for flow over a given surface will
require numerical solutions or experiments with
several sets of:
– Velocities (V)
– Surface lengths (L)
– Wall temperatures (Ts)
– Free stream temperatures (T).

• The same information can be determined with far

fewer experiments or investigations by grouping the
data into the dimensionless:
– Reynolds number (Re)
– Prantdl number (Pr) 91
Similarity

Fig 6-29 (text)

92
 Similarity
• Another advantage is that data from a large group of
experiments can be conveniently reported in the
terms of the similarity parameters.

93
3.5 Reynold’s Analogy

94
Forced Convection
(Drag Force)

95
 Forced Convection
(Reynold’s Analogy)

• In forced convection analysis, we are primarily

interested in the determination of quantities of:
– The coefficient of friction (CF) (to calculate the
shear stress at the wall)
– Nusselt number (Nu) ( to calculate the heat
transfer rates).
• Therefore, it is desirable to have a relation between
CF and Nu, so that we can calculate one when the
other is available.

96
 Forced Convection
(Reynold’s Analogy)

• Since:
u *  f1  x* , y * , Re L 

• The shear stress at the surface becomes:

u  V u *   V
s      f 2  x * , Re L 
y y 0
L y * y* 0
L

97
 Forced Convection
(Reynold’s Analogy)

• Substituting this into its definition gives the local

friction coefficient:
 V
 u *
s L y * y * 0 2 u*
C f ,x     *
 V2
2
 V2
2
Re y y * 0

f 2  x * , Re L 
2

Re L
 f 3  x , Re L 
*

98
 Forced Convection
(Reynold’s Analogy)

• Similarly, solving the energy equation for the

dimensionless temperature (T*) for a given geometry
gives:
*

T  g1 x , y , Re L , Pr
* *

• Using this definition, the convection heat transfer
coefficient (h) becomes:

k  Ty
y 0
h
Ts  T
99
 Forced Convection
(Reynold’s Analogy)
• Since:
T  Ts y y
T 
*
y  
*

Ts  T L x
for local

• Then:



T  Ts  T   T *  Ts


 Ts  T  T *
 *
y 
 y x
*
 x y

100
 Forced Convection
(Reynold’s Analogy)
• Therefore:

 k  T  Ts  T *
h   *
x   Ts  T  y y * 0

k T *
  *
L y y * 0

101
 Forced Convection
(Reynold’s Analogy)

• Substituting this into the local Nusselt number

equation gives:
h  x x  k T *  T *

Nu x     *
k k  x y y *0  y *
       y * 0

h
• We previously determined that:

T *  g1  x * , y * , Re L , Pr 
• Therefore:
T *
Nu x  *  g 2  x * , Re L , Pr 
y y * 0 102
 Forced Convection
(Reynold’s Analogy)

• Note: the Nusselt number is equivalent to the

dimensionless temperature gradient at the surface, and
this is why it is sometimes called the dimensionless heat
transfer coefficient (h).

Fig 6.30 (text) 103

 Forced Convection
(Reynold’s Analogy)

• The average friction and heat transfer coefficients

are determined by integrating the local CF,x and Nux
over the surface of the given body with respect to x*
(from 0 to 0.1), which removes the dependence on x*
and thus gives:

C F  f 4  Re L  and Nu  g 3  Re L , Pr 

• These relations allow experimenters to study a

problem with a minimum amount of experiments and
report their results in terms of just Re and Pr.
104
 Forced Convection
(Reynold’s Analogy)

• The experimental data for heat transfer is often

represented (with reasonable accuracy) by a simple
power law relation of the form:

Nu  C  Re mL  Pr n

– Where m and n are constant exponents (normally between 0

and 1), and the value of C depends on geometry.

105
 Forced Convection
(Reynold’s Analogy)

• Summary, so far (Fig 6-31 in text book):

106
 Forced Convection
(Reynold’s Analogy)

• Now if we simplify the momentum and energy

equations by assuming:

 Pr
P=*1 (which is approximately true for gases)
0
x *

 (true when u = u = V = constant)

For Pr = 1, the
thermal and
velocity boundary
layers coincide

107
 Forced Convection
(Reynold’s Analogy)

• The equations then become:

u *
u *
1  2 *
u
 Momentum: u *
v *
  *2
x *
y *
Re L y

T *
T *
1  2 *
T
 Energy: u *
v *
  *2
x *
y *
Re L y

• Note: These two equations are exactly in the same

form for u* and T*.
108
 Forced Convection
(Reynold’s Analogy)

• Since the boundary conditions are also identical:

– Recall: u *  0, y *   1 T *  0, y *   1
u *  x * ,0   0 T *  x * ,0   0
u *  x* ,    1 T *  x* ,    1

u * T *
 *
• Then: y * y Equation
y * 0 y * 0
*
109
 Forced Convection
(Reynold’s Analogy)

• Since as previously derived:

2 u * k T *
CF   * and h  *
Re y L y

• Rearranging these equations gives:

u * C F , x  Re T * hL
 and   Nu x
y * y * 0
2 y * y * 0
k

110
 Forced Convection
(Reynold’s Analogy)

• Therefore substituting these values into Equation

*
gives: u *
T
*
 *
y *
y * 0
y y * 0

C F , x  Re Reynold’s Analogy for

Nu x  Pr = 1
2
or
CF , x
St x 
2 111
 Forced Convection
(Stanton Number)

• Reynold’s Analogy can also be

expressed in terms of the Stanton
number (St).

• This was derived by Sir Thomas

Edward Stanton (1865-1931) from
England

h Nu
St  
  C P  V Re Pr
112
 Forced Convection
(Reynold’s Analogy)

• Reynold’s Analogy is important because it allows us

to determine the heat transfer coefficient (h) for fluids
where Pr = 1, from knowledge of the friction
coefficient (which is easier to measure).

113
 Forced Convection
(Chilton-Colburn Analogy)

• However, the Reynold’s number is of limited use

because of the restrictions:

 Pr = 1

P*
 x *  0

• Therefore it is desirable to have an analogy that is

applicable over a wide range of Pr.

• This is done by adding a Prandtl number correction.

114
 Forced Convection
(Chilton-Colburn Analogy)

• Recall as previously derived:

 12 1
C F , x  0.664  Re Nu x  0.332  Pr  Re x2
1

x and 3

• Taking their ratio and rearranging give the relation

known as the Chilton-Colburn analogy or the
modified Reynold's analogy:

CF , x  13 1 Colburn j-factor
 Nu x  Pr  Re  jH
L
2
For 0.6 < Pr < 60
CF , xhx 2
jH    Pr 3
2   C p  V 115
 Forced Convection
(Chilton-Colburn Analogy)

• The Chilton-Colburn Analogy is derived using:

– Laminar flow
– Over a flat plate ( P  0 )
x

– However, experimental studies however show that it is also

approximately applicable to turbulent flow over a surface in
the presence of pressure gradients.
– For laminar flow it is not applicable unless it is a flat plate,
therefore it cannot be applied to laminar flow in a pipe.
– Also the analogy above can be used for local or average
quantities. 116
 Forced Convection
(Example 3.4)

• Example 3.4 Laminar flow profile Air Flow

over a vertical plate. A 2 x 3 m plate
T= 15ºC
is suspended in a room and subject V = 7 m/s
to air flow parallel to its surfaces
along its 3 m side. The total drag
force acting on the plate is 0.86 N.
Determine the average heat transfer
coefficient (h) for the plate: 3m

• The properties of air at 1 atm (Table A-15 in Ts=25C

text book) at Tfilm= 20C:
kg kJ 2m
  1.204 3 ; C p  1.007
m kg
117
Pr  0.7309
 Forced Convection
(Example 3.4)

• Set L= 3 m ~ Characteristic length

• Since both sides of the plate are exposed to the air (and
considering the thickness negligibly small) the total surface area
is:

As  2  w  L
 2   2 m    3 m   12 m 2

118
 Forced Convection
(Example 3.4)

• For all flat plates:

Drag = Friction Force

 F friction  D   C F    As  V
1
2 
2

• Therefore:

2 D 2   0.86 N 
CF    0.00243
2
 
  As  V 1.204 mkg3  12 m 2    7 s
m 2

119
 Forced Convection
(Example 3.4)

• Then from the modified Reynold’s analogy (Chilton-

Colburn) the average heat transfer coefficient (h) can
be calculated:

C F   V  C p
h   2
2 Pr 3



0.00243 1.204

kg
m3
   7   1007
m
s
J
kg  C 
2
2 0.7309 3

 12.7 m 2WC

120
3.6 Convection in an
internal flow

121
 Internal Flow
• Internal flow relates to flow through fixed conduits
such as pipes or ducts.

122
Internal Flow
(Non-Circular Tubes)

• For flow through non-

circular tubes Re and Nu,
are based on the hydraulic
diameter Dh.
4  Ac
Dh 
p
  Vm  Dh
Re 

• Where p is the perimeter, Vm is
the mean velocity, and Ac is the
cross-sectional area.
123
 Internal Flow
(Mean Velocity)

• Because the velocity varies over the cross-section it

is necessary to work with a mean velocity (Vm) when
dealing with internal flows.

m    Ac  Vm

m
Vm 
  Ac

124
 Internal Flow
(Circular Tubes)

• In a circular tube:
 D2
4  Ac 4 
Dh   D 4
p D
 Vm D
Re 


Re < 2,300 laminar flow

2,300 < Re < 10,000 transitional flow
Re > 10,000 turbulent flow

125
Internal Flow
(Entrance Region)

126
 Internal Flow Equations
(Example 3.5)

• Example 3.5 - Temperature rise of oil in a bearing

(a) Find the temperature and velocity distributions
(b) Find the maximum temperature in the oil
(c) Find the maximum heat flux in the oil

Upper plate moving

V= 12 m/s

L= 2 mm u(y) Oil
k= 0.145 W/(m·K)
μ= 0.8 kg/(m·s)

Lower plate stationary 127

 Internal Flow Equations
(Example 3.5)
• Assumptions:
– Steady operating conditions
– Oil is incompressible with constant properties
– Body forces such as gravity are negligible
– The plates are large, so no variation in the z-direction

Upper plate moving

V= 12 m/s

L= 2 mm u(y) Oil
k= 0.145 W/(m·K)
μ= 0.8 kg/(m·s)

Lower plate stationary

128
 Internal Flow Equations
(Example 3.5)

(a) Find the temperature and velocity distributions

• Solution:
– Flow only in the x-direction v = 0
0
• Continuity Equation: u v
 0
x y
u
0
x
 u  u( y)
P
• The x-component of velocity does not change. Since  0 also, the
x
flow is maintained by the upper plate and not the pressure gradient.
129
 Internal Flow Equations
(Example 3.5)
• x-momentum equation:
0 0 0 0 0
 u v    u  u  P
2 2
  u  v     2  2    g x
 x y   x y  x

 u 2
0
y 2

• This is a 2nd order differential equation. So integrating twice

gives:

u  C1 y  C2
130
 Internal Flow Equations
(Example 3.5)
• The boundary conditions are:

u(0)= 0
u(L)= V= 12 m/s
• Using these boundary conditions to solve for the constants C1
and C2 gives:

0  C1  0  C2 V  C1  L   0
C2  0 V
C1 
L

131
 Internal Flow Equations
(Example 3.5)
• Therefore the equation becomes:

y
u  V
L

• Frictional heating due to viscous dissipation in this case is

significant because of the high viscosity of oil and large plate
velocity. The plates are isothermal and there is no change in
flow direction, so the temperature changes with y only T= T(y).

132
 Internal Flow Equations
(Example 3.5)
• So the energy equation for this system is:

0 0 0
 T T    2T  2T 
  C p  u v   k  2  2   
 x y   x y 
0 0 02
  u   v    u v 
2 2

   2            
  x   y    y x  

2
 T 2
 u 
0k    
y 2
 y  133
 Internal Flow Equations
(Example 3.5)

• Since: y
u  V
L
u V

y L
• Therefore the equation becomes:

2
T2
V 
k 2    
y L
134
 Internal Flow Equations
(Example 3.5)
• Now integrating the equation twice:
2
 y 
T    V   C3  y  C 4
2k  L 
• Applying boundary conditions:

T(0) = T0 y  0 : T0  C4
T(L) = T0 2
 L 
y  L : T0     V   C3  L  T0
2k  L 
 2
C3  V
2kL
135
 Internal Flow Equations
(Example 3.5)
• Substituting these constants in to the equation gives:

 y 2
2 y 2
T    2  V    V  T0
2k  L  2kL
 V 2  y y 2 
 T0    2 
2k  L L 

136
 Internal Flow Equations
(Example 3.5)

(b) Find the maximum temperature in the oil

The temperature gradient is found by differentiating T(y) with
respect to y.
T  V 2  y
 1  2    0
y 2kL  L
Now to find the maximum temperature, maximize T by setting
the above equation equal to 0.
y
1 2
L
L 0.002 m
y   0.001 m
2 2
137
 Internal Flow Equations
(Example 3.5)
• This means that the maximum temperature will occur at the mid-
plane (y= 1 mm), which is not surprising since both planes are
maintained at the same temperature.
• The maximum temperature at y= 1 mm is:

 V 2  L2  L2  
2

Tmax  T0   
2k  L L2 

 V 2
 T0 
8k

 20C 
 0.8 m 2   12 s   1 W 
N s m 2
    119 C
8  0.145 mWC  1 Nsm 
138
 Internal Flow Equations
(Example 3.5)

(c) Find the maximum heat flux in the oil

The heat flux at the plates is determined from the definition of a
heat flux.

dT  V 2  y
q 0  k  k  1  2  
dy y 0
2kL  L


 V


2 0.8   12  1 W 
N s
m2
m 2
s
 N m
2L 2   0.002 m  1 s 
W
 28,800 2
m
139
 Internal Flow Equations
(Example 3.5)
• As a check, we can also calculate the heat flux at y= L (should
be equal but opposite sign).

dT  V 2  L
q L   k  k  1  2  
dy yL
2kL  L


 V


2 0.8   12  1 W 
N s
m2
m 2
s
 N m
2L 2   0.002 m  1 s 
W
 28,800 2
m Correct !
140
 Internal Flow Equations
(Example 3.5)

• Discussion of example
T=20ºC Upper plate moving
V= 12 m/s

L= 2 mm T=119ºC

Lower plate stationary

T=20ºC

• A temperature rise of 99ºC confirms that viscous dissipation is

very significant
141
 Internal Flow Equations
(Example 3.5)
• Discussion of example
kW
q   28.8
m2
V= 12 m/s

L= 2 mm

kW
q   28.8
m2
• Heat flux is equivalent to the mechanical energy rate of
dissipation. Therefore, mechanical energy is being converted
into thermal energy to overcome friction in oil. This accounts
for the temperature flux. 142
3.7 Free (natural) convection

143
 Free Convection

• Hot air rises due to the

Warm air
buoyancy effect.
• This causes fluid motion
(possibly in a circulating pattern)
that causes natural or free
convection
Heat
Transfer

Cold
can Cold air 144
 Free Convection
(Volume Expansion Coefficient)

• In heat transfer, the primary variable is the

temperature, so it is desirable to express the net
buoyancy force in terms of a temperature difference.
– This requires knowledge of a property that represents the
variation of the density of a fluid with temperature at constant
pressure.
– This is called the volume expansion coefficient (β) which is
defined as:

1    1   
       
  T  P   T  P
145
 Free Convection
(Volume Expansion Coefficient)

• In natural convection studies, the condition of the fluid

sufficiently far from the hot or cold surface is indicated by the
subscript “” to indicate that the presence of the surface is not
felt.

• In such cases, β can be expressed approximately by replacing

the differential equations by differences, such as:

1  1     
    
 T   T  T 

          T  T 
146
 Free Convection
(Volume Expansion Coefficient)

• For an ideal gas:

P

R T

• Thus for an ideal gas the discharge coefficient

becomes:

1    1  RTP  1
     P  
  T  P RT  T  T

147
 Free Convection
(Grashof Number)

• The velocity and temperature for natural

convection over a vertical plate are
shown in the figure.
– As in forced convection, the
boundary layer thickness increases
in the flow direction

– Unlike forced convection, the fluid

velocity (u) is 0 at the outer edge of
the boundary layer as well as the
surface of the plate.

– This is expected since the fluid

beyond the boundary layer is 148

motionless.
 Free Convection
(Grashof Number)

• Recall that the x-momentum equations is:

 u u   2u P
  u  v    2   g
 x y  y x
• Now the momentum equation outside the boundary layer can be
obtained from this relation as a special case by setting u = 0,
giving:
P
    g
x

149
 Free Convection
(Grashof Number)

• Since: P  P x   P  x 
P P
     g
x x
• Then the momentum equation becomes:
 u v   2u
  u  v    2        g
 x y  y
 u v   2u
  u  v    2   g   T  T 
 x y  y
u v  2u
EQN 9-13 u  v   2  g   T  T 
in text x y y 150
 Free Convection
(Grashof Number)

• If we now non-dimensionalize this x-momentum

equation, we get:

u *
v
u *


g    T  T   L3
 T *
1  2 *
u
 2   *2
* * s  c
u
x y          Re L Re L y
* * 2

Grashof Number

151
 Free Convection
(Grashof Number)

• The Grashof number is derived by

Franz Grashof (1826-1893) from
Germany.

g    Ts  T   L3c
GrL 
2

152
 Free Convection
(Grashof Number)

• Gr is a measure of the relative

magnitudes of the buoyancy force
and the opposing viscous force
acting on the fluid

153
 Free Convection
(Raleigh Number)

• Lord Raleigh (1842-1919) from

England derived the Raleigh Number

Ra  Gr  Pr

g    Ts  T   L
3
RaL  c
 Pr
 2

154
 Free Convection
(Example 3.6)

• Example 3.6 A 6-m long section of 8-cm diameter

horizontal hot water pipe passes through a large
room. The pipe surface temperature is 70 ºC.
Determine the heat loss from the pipe by natural
convection.
Ts= 70 ºC T= 20 ºC

D= 8 cm

L= 6 m

155
 Free Convection
(Example 3.6)

• Assume:
– Steady operating conditions
– Air is an ideal gas
– The local atmospheric pressure is 1 atm

Ts  T  70C  20C 
T film    45C
2 2
• From Table A-15, the properties of air are:

k  0.02699 W
m C ;   1.749 10 5 m 2
sec ; Pr  0.7241
156
 Free Convection
(Example 3.6)

• The volumetric expansion coefficient (β) is:

1 1 1
  
T f 45C  273 318 K
• The characteristic length is the outer diameter of the
pipe:

Lc  D  0.08 m

157
 Free Convection
(Example 3.6)

• Therefore the Raleigh Number is:

g    Ts  T   D 3
RaD   Pr
 2


 
m 1

9.81 s 2  318 K   343 K  293 K   0.08 m   0.7241
  3

 
5 m 2 2
1.749  10 s
 1.869  10 6

158
 Free Convection
(Example 3.6)

• Table 9-1 in the text book gives average Nusselt

numbers for natural convection over surfaces.

• For a horizontal cylinder:

159
 Free Convection
(Example 3.6)

• Thus Nu is:
2
 
 
 0.387  RaD6
1

Nu D  0.60  8 
   0.559  1 6  
9 27

 1     
   Pr   
2
 
 

 0.60 

0.387  1.869  10 6 6 
1
  17.4
8

   0.559  1 6  
9 27

 1     
   0.7241    160
 Free Convection
(Example 3.6)

• Then:

k
h   Nu 
 0.02699 mWC 
 17.4  5.869 W
D  0.08 m  2
m  C

• The surface area of the cylinder is:

As    D  L
    0.08 m    6 m   1.508 m 2

161
 Free Convection
(Example 3.6)

• Therefore the heat transfer is:

Q  h  As   Ts  T 
  5.869 W
m C
  1.508 m    70C  20C 
2

 442.5 W

162