Principles of Convection 7

7.1. Mechanism of Heat Convection. 7.2. Classification of Convection. 7.3. Convection Heat Transfer Coefficient. 7.4. Convection
Boundary Layers—Velocity boundary layer—Thermal boundary layer—Significance of boundary layers. 7.5. Laminar and Turbulent
Flow—Laminar boundary layer—Turbulent boundary layer. 7.6. Momentum Equation for Laminar Boundary Layer. 7.7. Energy Equation
for the Laminar Boundary Layer. 7.8. Boundary Layer Similarities—Friction coefficient—Nusselt number. 7.9. Determination of
Convection Heat Transfer Coefficient—Dimensional analysis—Exact mathematical solutions—Approximate analysis of boundary layers—
Analogy between heat and momentum transfer—Numerical analysis. 7.10. Dimensional Analysis—Primary dimensions and dimensional
formulae—Dimensional homogeneity—Rayleigh’s method of dimensional analysis—Buckingham π theorem—Dimensional analysis for forced
convection—Dimensional analysis for natural convection. 7.11. Physical Significance of the Dimensionless Parameters—Reynolds
number—Critical reynolds number Recr—Prandtl number—Grashof number—Nusselt number—Stanton number—Peclet number—Graetz
number. 7.12. Turbulent Boundary Layer Heat Transfer—Prandtl mixing length concept—Turbulent heat transfer. 7.13. Reynolds Colburn
Analogy for Turbulent Flow Over a Flat Plate. 7.14. Mean Film Temperature and Bulk Mean Temperature. 7.15. Summary—Review
Questions—Problems—References and Suggested Reading.

The objective of this chapter is to give basic 7.1. MECHANISM OF HEAT CONVECTION
understanding of physics of convection heat transfer and
to present them in the form of general equations, which As discussed in chapter one, the heat convection involves
are applied in subsequent chapters for the particular two mechanism, simultaneously. One is energy transfer
cases. from a hot surface to a adjacent fluid by random
In the previous chapters, we dealt with heat molecular motion, it is called diffusion. The other one is
conduction, which is a mechanism of heat transfer due advection, i.e., the transport of energy by bulk movement
to random molecular activities through a stationary of the fluid from higher temperature region to lower
medium, solid or fluid. The convection heat transfer was temperature region. Such motion in presence of tem-
restricted to the boundary conditions only and the rate perature gradient will enhance the heat transfer rate.
of heat convection at the boundaries was considered The molecules in aggregate retain their random motion
constant so far. and the fluid motion brings the hotter and colder fluid
The convection heat transfer is of importance to chunks in contact, thus initiating the high rate of
practical problems in industrial application. The flow conduction at a large number of sites in the fluid.
of a liquid or a gas through a heat exchangers, two phase Therefore, the rate of heat transfer in the convection is
flow in the boilers and condensers, cooling of electronic due to superposition of energy transfer by random
chips, heat removal from the condenser of a refrigerator molecular motion (conduction) at the surface as well as
are some common examples of convection heat transfer. the energy transfer by bulk motion of fluid.
The convection heat transfer is recognised closely
related to the fluid flow. Hence understanding of 7.2. CLASSIFICATION OF CONVECTION
convection should start with basic knowledge of fluid
dynamics, momentum transfer, energy transfer, shear The convection heat transfer is classified as natural (or
stress, pressure drop, friction coefficient and the nature free) or forced convection, depending on how the fluid
of fluid flow like laminar or turbulent etc. motion is initiated. The natural or free convection is a

234
 PRINCIPLES OF CONVECTION 235

process, in which the fluid motion results from heat Consider the heating of a cold iron block as
transfer. When a fluid is heated or cooled, its density shown in Fig. 7.2. If there is no significant velocity of
changes and the buoyancy effects produce a natural hot air surrounds the block, the heat will be
circulation in the affected region, which causes itself transferred from hot air to block by natural convection.
the rise of warmer fluid and the fall of colder fluid : If a fan blows air over the block, the heat will be
Therefore, energy transfers from hotter region to colder transferred from hot air to cold block by forced
convection. If the speed of the air over the block surface
region and such process is repeated as long as the
increases, the block will be heated up faster. If air is
temperature difference in the fluid exists.
replaced by water, the heat transfer rate by convection
In the forced convection, the fluid is forced to flow will be increased several times.
over a surface or in a duct by external means such as a
pump or a fan. A large number of heat transfer Air Relative
T¥ = 100°C
applications utilize forced convection, because the heat velocities of
u¥ = 5 m/s fluid layers
transfer rate is much faster than that in free convection.
Zero
velocity
Air Q
at the
20°C surface
5 m/s Q

Heated Hot iron block

plate at 70°C Ts = 20°C

(a) Forced convection Fig. 7.2. Heating of cold block by forced convection
Experience shows that the convection heat
Warm transfer strongly depends on fluid properties, dynamic
Q
Air air rising viscosity µ, thermal conductivity kf, density ρ, and
Heated specific heat Cp, as well as on the fluid velocity. It also
plate depends on geometry and roughness of the solid surface,
in addition to type of fluid flow. Thus the convection
heat transfer relations are rather complex, because of
(b) Natural convection dependence of convection on so many variables.

Q
Stagnant No convection 7.3. CONVECTION HEAT TRANSFER
air current
Heated
COEFFICIENT
plate
The rate of heat transfer per unit surface area from a
surface to a fluid is proportional to temperature
(c) In absence of fluid motion, heat transfer in the difference and it is expressed as
fluid is by conduction only
qconv ∝ (Ts – T∞)
Fig. 7.1. The heat transfer from a hot surface qconv = h (Ts – T∞) ...(7.1)
to the surrounding fluid
where h = constant of proportionality and is called
Further, the convection heat transfer is also heat transfer coefficient,
classified as external convection or internal convection. Ts = temperature of the surface, °C
In external convection, the fluid surrounds a surface T∞ = temperature of free stream fluid, °C.
such as flow over a flat or curved surface, while in Based on the interpretation, the convective heat
internal convection, the fluid is surrounded by a surface transfer coefficient is expressed as
such in a pipe carrying steam or water filled cooling
qconv
passage in an internal combustion engine. The fluid h= ...(7.2)
flows can also be stated as laminar, turbulent or (Ts − T∞ )
translatory (transition from laminar to turbulent). or it is defined as the rate of convection heat transfer
Forced and natural convection have separate criteria per unit surface area per unit temperature difference.
for distinctions of these regims. It is measured in W/m2.K or W/m2. °C.
 236 ENGINEERING HEAT AND MASS TRANSFER

Consider the flow of a hot fluid at temperature the wall surface is obtained over entire distance x = 0 to
T∞ over a cold surface at a constant temperature of Ts x = L and width w as
as shown in Fig. 7.2. It is observed that the fluid layer Q = h (wL) (T∞ – Ts) ...(7.6)
in contact with the solid surface sticks to the surface. It
Example 7.1. Experimental results for local heat
is very thin layer of fluid and has zero velocity (no slip
transfer coefficient hx for flow over a flat plate with an
condition). Therefore, the heat transfer from wall surface
to the adjacent, fluid layer is by pure conduction, and extremely rough surface were found as
the conduction heat flux q(x) at the wall surface y = 0 is hx = ax–0.1
given by where a is a constant and x is a distance from the leading
∂T( x, y) edge of the plate.
q(x) = kf ...[7.3(a)] Develop an expression for ratio of average heat
∂y y=0
where T(x, y) is temperature distribution of fluid transfer coefficient h for a plate of length x to the local
heat transfer coefficient hx at x.
∂T
= temperature gradient at the surface, Solution
∂y y=0
Given : The variation of local heat transfer
kf = thermal conductivity of the fluid.
coefficient as hx = ax–0.1
The negative sign is omitted from the eqn. [7.3(a)]
because the heat flow from fluid to wall, i.e., in negative To find : The ratio of average heat transfer
y direction. coefficient to local heat transfer coefficient.
The heat transfer rate between the fluid and the Analysis : The average heat transfer coefficient
wall surface is related to the local heat transfer is given by eqn. (7.5), over a distance 0 to x is
coefficient hx, defined as
q(x) = hx (T∞ – Ts) ...[7.3(b)]
where Ts and T∞ are the wall surface and free stream
h=
1
x z0
x
hx dx

Boundary
fluid temperatures, respectively. T∞ layer hx = ax
– 0.1

In steady state conditions, the heat flow rate is

constant, thus equating eqn. [7.3(a)] with eqn. [7.3(b)],
hx (T∞ − Ts ) ∂T( x, y) Ts

  = k
convection ∂y y=0
x

 
conduction
Fig. 7.3. Schematic
∂T( x , y )
k Using hx = ax–0.1 and integrating, we get

zL z
∂y y=0
or hx = ...(7.4) 1 x a x
T∞ − Ts h= ax – 0.1 dx = x – 0.1 dx
x 0 x 0

MN OPQ
where hx is local heat transfer coefficient at a certain
position x in flow direction and for given temperature a x 0.9
= = 1.11 ax–0.1
distribution in the flow. It is calculated from eqn. (7.4). x 0.9
It is used to obtain the heat flux at any location in the = 1.11 hx. Ans.
fluid flow.
Example 7.2. Experimental results for heat transfer over
The local heat transfer coefficient may vary along a flat plate with an extremely rough surface were found
the length of flow as a result of changes in the velocity to be correlated by an expression of the form
and other parameters in the flow direction. We are Nux = 0.04 Re0.9 Pr1/3
usually interested for the heat transfer rate from the
where Nux is the local value of Nusselt number at a
entire surface. Which can be obtained by using average position x measured from the leading edge of the plate.
heat transfer coefficient over a distance x = 0 to x = L, Derive an expression for ratio of average heat transfer
determined from coefficient to local heat transfer coefficient hx.
1 L
h=
L 0 z
hx dx

With use of average value of heat transfer

...(7.5) Solution
The local Nusselt number for flow over a flat plate
is given by
coefficient h, the heat transfer rate Q from the fluid to
 PRINCIPLES OF CONVECTION 237

Nux = 0.04 Re0.9 Pr1/3 (b) Thermocouple reading, when gas velocity is
hx x 20 m/s.
where Nux = ,
kf Wall

ρ u∞ x µCp Thermocouple
Rex = and Pr =
µ k u gas
The local heat transfer coefficient is expressed
as
FG ρ u x IJ kf ∞
0.9 Fig. 7.4. Schematic
x H µ K
hx = 0.04 × Pr1/3
Assumptions :
F ρ u IJ x
= 0.04 k G
∞
0.9 (i) Steady state conditions,
H µK
–0.1 Pr1/3 ...(i)
f (ii) Constant properties.
The average heat transfer coefficient is obtained Analysis : The heat transfer coefficient is given
as by

z F ρ u IJ x Pr dx 0.9
h ∝ u0.8 or h = au0.8
0.04 k G
1 x
∞
H µK
h= –0.1 1/3
f and the heat transfer rate by radiation
x 0

0.04 k F ρu I
z Qrad ∝ ∆T or Qrad = b ∆T
0.9

GH µ JK Pr x dx
x
f ∞ 1/ 3 − 0.1
= where a and b are constants of proportionality under
x 0

0.04
k G
F ρu IJ Pr LM x OP ∞
0.9
1/3
0.9
steady state conditions.
Rate of convection heat transfer from thermo-
=
x HµK f
N 0.9 Q couple to gas
0.04
k G
F ρu x IJ Pr ∞
0.9
1/3 = Rate of heat radiation from walls
=
0.9 x H µ K f ...(ii)
to thermocouple
h 1 h A(T – T∞) = Qrad
and hx = 0.9 = 1.11. Ans. a u0.8 A(T – T∞) = b(Tw – T)
Example 7.3. A bare thermocouple is used to measure b (Tw − T) Tw − T FG IJ
the temperature of a gas flowing through a hot duct. The
heat transfer coefficient between a gas and thermocouple
or u0.8 =
a A (T – T∞ )
=d
T − T∞ H K
is proportional to u0.8, where u is the gas velocity and b
where d = , a new constant.
heat transfer rate by radiation from the walls to the aA
thermocouple is proportional to temperature difference. (a) (i) At u1 = 5 m/s and T1 = 323 K
When the gas is flowing at 5 m/s, the thermo- FG T − 323 IJ
w
couple reads 323 K and when it is flowing at 10 m/s, it (5)0.8 = d
H 323 − T K ∞
reads 313 K. Calculate the appropriate wall temperature
3.624 × G
F 323 − T I
H T − 323 JK
at a gas temperature of 298 K. or d=
∞
...(i)
What temperature will the thermocouple indicate w
when the gas velocity is 20 m/s ? (ii) At u2 = 10 m/s, T2 = 313 K
Solution FG T − 313 IJ
w

Given : Thermocouple is exposed to gas stream

(10)0.8 = d
H 313 − T K ∞
Using d from eqn. (i), we get
h ∝ u0.8
F 323 − T I × FG T − 313 IJ
Qrad ∝ ∆T
u1 = 5 m/s, T1 = 323 K
6.309 = 3.624 × GH T − 323 JK H 313 − T K
w
∞ w

u2 = 10 m/s, T2 = 313 K (iii) At T∞ = 298 K

F 323 − 298 I × FG T − 313 IJ
u3 = 20 m/s, T∞ 2 = 298 K. 6.309 = 3.624 × GH T − 323 JK H 313 − 298 K
w
w

To find :
Tw − 313
(a) Wall temperature, when gas temperature is 1.0446 =
298 K. Tw − 323
 238 ENGINEERING HEAT AND MASS TRANSFER

or 1.0446 Tw – 337.4 = Tw – 313 The retardation of fluid motion in the boundary

or 0.0446 Tw = 337.41 – 313 = 24 layer is due to the shear (viscous) stresses acting in
Tw = 548 K. Ans. opposite direction. With increasing the distance y from
(b) At u3 = 20 m/s the surface, shear stress decreases, the local velocity u

FG 548 − T IJ increases until approaches u∞. With increasing the dis-

tance from the leading edge, the effect of viscosity
(20)0.8 = d
H T − 298 K penetrates further into the free stream and boundary
10.98 = 3.624 layer thickness grows (δ increases with x).
FG 323 − 298 IJ × FG 548 − T IJ In fluid mechanics, the surface shear stress τs in
×
H 548 − 323 K H T − 298 K terms of skin friction coefficient Cf is expressed as
or 10.98 T – 3272 = 220.66 – 0.40 T Cf
τs = ρ u∞ 2 ...(7.7)
11.37 T = 3488.7 2
or T = 307 K. Ans. The surface shear stress may also determined
from knowledge of velocity gradient of the fluid at the
7.4. CONVECTION BOUNDARY LAYERS surface
LM du OP
7.4.1. Velocity Boundary Layer
Consider the flow of fluid over a flat plate as shown in
τs = µ
N dy Q y=0
...(7.8)

Fig. 7.5. The fluid approaches the plate in x direction 7.4.2. Thermal Boundary Layer
with a uniform velocity u∞. The fluid particles in the If the fluid flowing on a surface has a different
fluid layer adjacent to the surface get zero velocity. This temperature than the surface, the thermal boundary
motionless layer acts to retard the motion of particles layer is developed in similar manner to velocity
in the adjoining fluid layer as a result of friction between boundary layer.
the particles of these two adjoining fluid layers at two
Consider a fluid at temperature T∞ flows over a
different velocities. This fluid layer then acts to retard
surface at a constant temperature Ts. The fluid particles
the motion of particles of next fluid layer and so on,
in adjacent layer to the plate get the same temperature
until a distance y = δ from the surface reaches, where
that of surface. The particles exchange heat energy with
these effects become negligible and the fluid velocity u
particles in adjoining fluid layer and so on. As a result,
reaches the free stream velocity u∞. As a result of
the temperature gradients are observed in the fluid
frictional effects between the fluid layers, the local
layers and a temperature profile is developed in the fluid
fluid velocity u will vary from x = 0, y = 0 to y = δ.
flow, which ranges from Ts at the surface to fluid
temperature T∞ sufficiently far from the surface in y
Y
Velocity direction. The flow region over the surface in which the
Boundary
layer
temperature variation in the direction normal to surface
Velocity profile is observed is called thermal boundary layer. The
u¥ d(x)
u(x, y) thickness of thermal boundary layer δth at any location
along the length of flow is defined as a distance y from
x
the surface at which the temperature difference (T – Ts)
Fig. 7.5. Velocity boundary layer on a flat plate equals 0.99 of (T∞ – Ts).
The region of the flow over the surface bounded or Ts – T = 0.99(Ts – T∞) if Ts > T∞
by δ in which the effects of viscous shearing forces caused where T is local temperature in thermal boundary layer,
by fluid viscosity are observed, is called the velocity a function of x and y directions.
boundary layer or hydrodynamic boundary layer With increasing the distance from leading edge
or simply the boundary layer. The thickness of the effect of heat transfer penetrates further into the
boundary layer δ is generally defined as a distance from free stream and the thermal boundary layer grows as
the surface at which local velocity u = 0.99 of free stream shown in Fig. 7.6 (a) and Fig. 7.6 (b).
velocity u∞.
 PRINCIPLES OF CONVECTION 239
Thermal
T¥
boundary t.b.l d
Y layer dth th
— >> 1, Pr << 1
d
u¥ v.b.l
u∞ d
T∞ Ts
T∞
Plate at δth
Ts (a) Liquid metals
O X
x T = f(x, y) t.b.l
T¥
Fig. 7.6. (a) Thermal boundary layer for flow of d
th
a cold fluid over a hot plate u¥ v.b.l — <1
dth d
d
T = Ts + 0.99 (T∞ – Ts)
Thermal Ts
Y boundary
layer (b) Gases
u∞
T∞ Temperature
T∞ v.b.l
profile
δth T = f(x, y)
u¥ d T¥
X dth
O Ts t.b.l — << 1, Pr >> 1
x dth d
Ts
Fig. 7.6. (b) Thermal boundary layer for flow
of hot fluid on a cold plate Thermal boundary layer (t.b.l)
Velocity boundary layer (v.b.l)
The convection heat transfer rate any where along
the surface is directly related to the temperature (c) Oils
gradient at that location. Therefore, the shape of the Fig. 7.7. Relative thickness of thermal and velocity
temperature profile in the thermal boundary layer leads boundary layers for different types of fluid
to the local convection heat transfer between surface
and flowing fluid. 7.5. LAMINAR AND TURBULENT FLOW
7.4.3. Significance of Boundary Layers
The analysis of convection problems requires the
The velocity boundary layer, is of extent δ(x) and is knowledge of the type of boundary layer developed,
characterised by the presence of velocity gradients and whether it is laminar or turbulent. The type of boundary
fluid friction. The thermal boundary layer is of extent strongly influences the skin friction and heat transfer
δth(x) and is characterised by temperature gradient and coefficient.
heat transfer.
The developed boundary layer may consist of
For flow over a heated (or cold) surface, both laminar boundary, transition region and turbulent
velocity and thermal boundary layers are developed boundary layer as shown in Fig. 7.8.
simultaneously. If the effects of fluid viscosity (viscous The velocity boundary layer δ(x) is characterized
shear stress) is stronger than thermal effects, then the by the presence of velocity gradients and shear stresses.
velocity boundary layer will be thicker than the thermal
The thermal boundary layer δth(x) is characterized
boundary layer and vice versa. The schematic
by temperature gradients and heat transfer.
illustrations of relative thickness of δ(x) and δth(x) for
liquid metals, gases and oils are shown in Fig. 7.7. For 7.5.1. Laminar Boundary Layer
liquid metals, thermal effects are much stronger than The velocity boundary layer starts at the leading edge
viscous effects, and therefore, thermal boundary layer of the plate as a laminar boundary layer, in which the
(t.b.l) is much thicker than the velocity boundary layer fluid motion is highly ordered and it is possible to identify
(v.b.l) Fig. 7.7 (a). For gases, the viscous effects are the stream lines along which particles move. The fluid
slightly weaker than thermal effects, thereore, thermal motion along a stream line is characterized by the
boundary layer is a little thicker than velocity boundary
velocity components u and v in both x and y directions
layer, Fig. 7.7 (b). Similarly for oils, greese etc, the
and it influences the momentum and energy transfer
viscous effects are much stronger than thermal effects
through the boundary layer. The velocity profile in
and thus, the velocity boundary layer is much thicker
laminar boundary layer is approximately parabolic.
than the thermal boundary layer Fig. 7.7 (c).
 240 ENGINEERING HEAT AND MASS TRANSFER

7.5.2. Turbulent Boundary Layer and heat transfer mechanisms involve the fluid lumps
The fluid motion in the turbulent boundary layer has moving randomly.
very large disturbances and is characterized by velocity The characteristic of fluid flow is governed by
fluctuations. The fluctuations increase the momentum dimensionless quantity called the Reynolds number as
and heat transfer. Due to fluid mixing, the turbulent
u∞ x
boundary layer thickness is larger and velocity profiles Rex = ...(7.9)
ν
are flatter with the sharp drop near the surface.
where, u∞ = free stream velocity, m/s,
Laminar boundary Transition Turbulent boundary ν = kinematic viscosity, m2/s,
layer region layer Turbulent
Boundary u¥ layer x = distance from the leading edge for flow
layer u(x, y) u
¥
over a flat plate, m.
y thickness
d(x) The Reynolds number at which the transition
x from laminar to turbulent boundary layer takes place
xcr
u¥xcr
d(x) Buffer
Viscous is called the critical Reynolds number and for flow along
Boundary layer
Recr = v
layer thickness
sublayer a flat plate, the transition begins at critical Reynolds
number
Fig. 7.8. Boundary layer concept for flow along a flat plate
Recr ≈ 5 × 105. ...(7.10)
At some distance from the leading edge, the small
disturbances in the flow begin to be amplified and the Example 7.4. Water flows at 20°C at 8 kg/s through
fluid fluctuations begin to develop, it is transition from the diffuser having 3 cm diameter at the entrance and
laminar to turbulent boundary layer as shown in 7.0 cm diameter at its exit. Calculate the fluid velocity
Fig. 7.8. The transition to turbulence is attained by and Reynolds number at the inlet and exit of the diffuser.
significant increase in boundary layer thickness, wall Solution
shear stress, and heat transfer coefficient. These effects Given : Flow of water through a diffuser.
are shown in Fig. 7.9.
D1 = 3.0 cm = 0.03 m, D2 = 7.0 cm = 0.07 m
h m
= 8 kg/s.
Cfx To find :
(i) Velocity of water at inlet and exit of diffuser.
hx or Cfx
(ii) Reynolds number at the inlet and exit of
d d(x)
diffuser.
T¥
Assumptions :
(i) Steady flow conditions.
u¥
(ii) Constant properties of fluid.

xcr Flow
x 8 kg/s
Laminar Transition Turbulent

Fig. 7.9. Variation of velocity boundary layer δ(x), local Inlet Exit
heat transfer coefficient hx and local friction coefficient
Cfx for flow over a flat plate Fig. 7.10. Schematic of diffuser

The turbulent boundary layer has three

different regions. A laminar sublayer is very thin layer Analysis : (i) The properties of water at 20°C
next to wall in which flow is laminar. Adjacent to laminar δ = 1000 kg/m3, µ = 1007.4 × 10–6 kg/m/s
sublayer, there is a buffer layer in which small The flow cross-sectional area at inlet
disturbances exist. The buffer layer is followed by the π 2 π
A1 = D = × (0.03 m)2 = 7.069 × 10–4 m2
turbulent layer with larger turbulences. The momentum 4 1 4
 PRINCIPLES OF CONVECTION 241

π 2 π (i) The Reynolds number is expressed as

At exit A2 = D = × (0.07 m)2
4 2 4 ρu∞ x Re x µ
= 3.848 × 10–3 m2 Rex = or x =
µ ρ u∞
Using continuity equation, the velocities.
For Rex = 108,
m

At inlet u1 = 10 8 × 184.6 × 10 − 7
ρ1 A 1 x= = 31.82 m Ans.
1.16 × 50
8 kg/s The minimum length of the plate for Re = 108 is
=
3 –4 2 31.82 m.
(1000 kg/m ) × (7.069 × 10 m )
= 11.32 m/s. Ans. (ii) For transition to occur at Recr = 5 × 105
m
5 × 10 5 × 184.6 × 10 − 7
At exit u2 = xcr =
ρ2 A 2 1.16 × 50
(8 kg/s) = 0.159 m. Ans.
=
(1000 kg/m ) × (3.848 × 10 − 3 m 2 )
3 The transition from laminar to turbulent will
occur at x = 0.159 m.
= 2.08 m/s. Ans.
(ii) The respective Reynolds numbers
7.6. MOMENTUM EQUATION FOR LAMINAR
ρu1D 1 1000 × 11.32 × 0.03
Re1 = = BOUNDARY LAYER
µ 1007.4 × 10 − 6
= 337105. Ans. Considering two dimensional control volume as shown
in Fig. 7.11. The equation of motion for the laminar
ρu2 D 2 1000 × 2.08 × 0.07 boundary layer can be obtained by equating force and
and Re2 = =
µ 1007.4 × 10 − 6 momentum transfer on the element. The assumptions
= 144530. Ans. made in the analysis are

Example 7.5. A fan provides air speed upto 50 m/s, is Y

used in low speed wind tunnel with atmospheric air at
rvdx + rdx ¶v dy
27°C. If this wind tunnel is used to study the boundary mdx [ ¶¶uy+ ¶y¶ ( ¶¶uy (dy [ ¶y u¥ x
layer behaviour over a flat plate upto Re = 108. What dy
should be the minimum plate length ? At what distance dx
from the leading edge would transition occur, if critical dx
Reynolds number is Recr = 5 × 105 ? rudy ¶u
r udy + rdy dx
Control ¶x
Solution dy
Volume
pdy pdy + ¶ (pdy) dx
Given : Flow over a flat plate ¶y
¶u
m dx
u∞ = 50 m/s, T∞ = 27°C ¶y

Rex = 108, Recr = 5 × 105. rvdx

To find : (i) Minimum length of a flat plate for
Fig. 7.11. Force and momentum analysis for
Rex = 108. laminar boundary layer
(ii) Distance from the leading edge for
1. The flow is incompressible and steady ;
Recr = 5 × 105. 2. No pressure variation in perpendicular
Analysis : The properties of air at 27°C from direction of plate ;
Table A-4 3. Viscosity is constant ;
ρ = 1.16 kg/m3 4. Negligible shear forces in y direction ;
5. Unit depth in z direction.
µ = 184.6 × 10–7 kg/m/s
 242 ENGINEERING HEAT AND MASS TRANSFER

According to Newton’s second law of motion ∂

= ρvudx + ρdx (uv) dy
d (mv) x ∂y
ΣFx = ...(7.11)
dt
where ΣFx = sum of applied forces in x direction, and = ρuvdx + ρdxdy u
RS dv
+v
du UV
d (mv) x T dy dy W
= rate of increase in momentum flux in The net momentum transfer in x direction
dt
x direction. d(mv) x RS
= ρdxdy u
∂u
+u
∂u
+u
∂v
+v
∂u UV
The momentum flux in x direction is product of
mass flow rate through a particular side of control
dt T ∂x ∂x ∂y ∂y W
volume and x directional velocity component at that L F ∂u + ∂v IJ + RSu du + v du UVOP
= ρdxdy Mu G
point.
N H ∂x ∂y K T dx dy WQ
volume
The rate of mass entering the left face of control
or
d(mv)
= ρdxdy Su
x R ∂u + v ∂u UV ...(7.13)
= ρudy
dt T ∂x ∂y W
The rate of momentum entering the left face of R ∂u + ∂v UV = 0 (continuity equation)
Since u S
control volume T ∂x ∂y W
= ρudyu = ρu2dy. The forces acting in x direction are viscous and
The rate of mass leaving the right face pressure forces.
∂u The pressure force on the left face
= ρudy + ρdy dx
∂x = pdy
The rate of momentum leaving the right face The pressure force on the right face
∂ ∂
= ρu2 dy + ρdy dx (u2) = – pdy – (pdy) dx
∂x ∂x
LM
∂u
+u
∂u OP (∵ in opposite direction)
= ρu2 dy + ρdydx u
N
∂x ∂x Q The viscous force at bottom face
The rate of mass entering the bottom face = ρvdx ∂u
The rate of mass leaving the top face =–µ dx
∂y
∂v The viscous force at top face
= ρvdx + ρdx dy
∂y
The mass balances on the control volume =µ
∂u
dx + µ
RS UV
∂ ∂u
dxdy
∂u ∂v
∂y T W
∂y ∂y
ρudy + ρvdx = ρudy + ρdy dx + ρvdx + ρdx dy Net forces in x direction
∂x ∂y
Rearranging we get ; ∂p ∂ 2u
ΣFx = – dxdy + µ 2 dxdy ...(7.14)
R U
ρ S ∂u + ∂v V dxdy = 0
∂x ∂y

T ∂ x ∂ y W Substituting eqns. (7.13) and (7.14) in eqn. (7.11),

RS ∂u + ∂v UV = 0 we get
or
T ∂x ∂y W ...(7.12)
ρ u
RS ∂u
+v
∂u UV
∂ 2u ∂p
=µ 2 −
It is the mass continuity equation for the laminar T ∂x ∂y ∂yW ∂x
...(7.15)

boundary layer. The eqn. (7.15) is the momentum equation for the
laminar boundary layer with constant properties. If the
The rate of momentum in x direction associated
pressure changes on two side of control volume is
with mass entering the bottom face
negligible then above equation reduces to
= ρvudx
∂u ∂u µ ∂ 2u ∂ 2u
The rate of momentum in x direction leaves the u +v = = ν ...(7.16)
top face ∂x ∂y ρ ∂y2 ∂y2
 PRINCIPLES OF CONVECTION 243

7.7. ENERGY EQUATION FOR THE LAMINAR BOUNDARY LAYER

Consider the element control volume as shown in Fig. 7.12

u¥ x T¥

dy

Ts dx –kfdx [ ¶¶Ty + ¶y¶ ( ¶¶Ty (dy [

¶(vT)
(rvdx)CpT + (rCpdx)
¶y dy
2
¶u
Net viscous work m ( (
¶y
dx dy

(rudy) CpT Control

dy ¶(uT)
Volume (rudy)CpT + (rCpdy) dx
¶x

dx
¶T
–kf dx ¶y (rvdx)CpT

Fig. 7.12. Energy analysis for laminar boundary layer

The assumption made to simplify the analysis : Energy convected out the right face
1. Incompressible steady flow ; ∂
= ρCp (uT) dy + ρCp dy (u T) dx
∂x
2. Constant properties ; Energy conducted out the top face
3. Negligible heat conduction in direction of fluid ∂
flow. = ρCp (vT) dx + ρCp dx (v T) dy
∂y
The energy balance on the control volume can be Energy conducted out the top face
expressed as LM ∂T + ∂ T OP dx dy
= – kf
2

Energy convected at the left face

+ energy convected at the bottom face
N ∂y ∂y Q 2

Net energy change rate,

+ heat conducted in the bottom face + net
viscous work done on the element R∂
E′ = ρC dxdy S (u T) +
∂ U
(v T)V
= Energy convected out the right face
net p
T ∂x ∂y W
+ energy convected out the top face ∂2T
− kf dx dy ...(7.17)
+ heat conducted out the top face. ∂y 2
Writing the each quantity separately ; The net viscous force
∂u
Energy convected in the left face FD = µ dx
∂y
= (ρudy)CpT = ρCp(uT)dy Element moves through a distance per unit time
Energy convected in the bottom face ∂u
= dy
= (ρvdx)CpT = ρCp(vT)dx ∂y
Energy conducted in the bottom face Net viscous work done on the element,

∂T FG ∂u IJ 2

= – kf dx
∂y
WD = µ
H ∂y K dx dy
 244 ENGINEERING HEAT AND MASS TRANSFER

Writing the energy balance on the element, we To find :

get (i) An expression for local friction coefficient Cfx,

ρC S
R ∂(uT) + ∂(vT) UV = µ FG ∂u IJ + k ∂ T 2 2 (ii) An expression for average friction coefficient

T ∂x ∂y W H ∂y K
p ∂y
f 2
Cf .
Analysis : (i) The shear stress at the wall is
LF ∂T + v ∂T IJ + T FG ∂u + ∂vIJ OP
ρC MG u
expressed as
or
NH ∂x ∂y K H ∂x ∂y K Q
p
τs = µ
∂u
...(i)
∂y
R ∂u U ∂ T ...(7.18)
= µS V +k
2 2 y=0

T ∂y W ∂y
f 2 and shear stress in terms of local friction coefficient as
C fx
Using the continuity eqn. (7.12), τs = ρu∞ 2 ...(ii)
2
∂u ∂v
+ =0 Equating two equations for shear stress at the
∂x ∂y surface, we get
Rearranging eqn. (7.18), we get
2 µ ∂u
∂T
+v
∂T kf ∂ 2 T
+
µ RS ∂u UV 2
Cfx =
ρ u∞ 2 ∂y
u
∂x ∂y
=
ρC p ∂y 2
ρC p T ∂y W For given velocity profile
y=0

∂T
+v
∂T ∂2T
=α 2 +
ν RS ∂u UV 2
LM 3 × 1 − 1 . 3 y OP
T ∂y W
2
or u ...(7.19) ∂u 3u∞
∂x ∂y ∂y Cp = u∞ =
For low velocity flow, viscous forces are negligibly
∂y y=0 N2 δ 2 δ Q 3
y=0
2δ

small in comparison to conduction term, then Then

∂T ∂T ∂2T 2µ 3 u∞ 3ν
u +v =α 2 ...(7.20) Cfx = 2
× =
∂x ∂y ∂y ρ u∞ 2 δ u∞ δ
Introducing the expression for δ(x), we get,
Example 7.6. The velocity profile u(x, y) for a boundary
layer flow over a flat plate is given by 3ν 13 u∞ ν
Cfx = × = 0.646
u( x, y) 3 y 1 y LM OP 3 u∞ 280 νx u∞ x
u∞
= −
2 δ 2 δ N Q 0.646
where the boundary layer thickness δ(x) is the function = . Ans.
Re x
of x and is given by
(ii) The average friction coefficient Cf is given by
280 νx
δ(x) =
13 u∞
(i) Develop an expression for local drag coefficient
Cf =
1
L z 0
L
C fx dx =
1
L z0
L
0.646
ν
u∞ x
dx

z
Cfx.
(ii) Develop an expression for average drag 0.646 ν L
= × x–1/2 dx
coefficient Cf over a distance x = L from the leading edge L u∞ 0
of the plate.
0.646 ν Fx I
− 1/ 2 + 1
L

Solution =
L u∞ GH − 1/2 + 1JK 0
Given : The velocity profile for the boundary layer
as
ν 2 × 0.646
u 3 y 1 y F I 3 = 2 × 0.646
u∞ L
=
Re L
u∞
= −
2 δ 2 δ H K
280 νx = 2 C fx Ans.
and δ(x) =
13 u ∞ x=L
 PRINCIPLES OF CONVECTION 245

Example 7.7. The temperature profile in a thermal hx x

boundary layer for flow over a flat plate is given by Nux = = 0.332 Rex1/2 Pr1/3
kf
T ( x, y) − Ts 3 y 1 y FG IJ 3
where Nux is called the local Nusselt number.
T∞ − Ts
= −
2 δ th 2 δ th H K and the thickness of
The average heat transfer coefficient
thermal boundary layer δth is the function of x and is
h=
1 L
z
L 0 x
h dx

z
x µ Cp
given by δth(x) = 4.53 where, Pr = Pr 1/3 L 1 ρ u∞ x
Rex 1/2 Pr 1/3 kf = 0.332 kf × dx
L 0 x µ

z
ρ u∞ x
and Re = . Develop the expressions for local and ρ u∞ 1 L
µ = 0.332 kf Pr1/3 × x − 1/2 dx
µ L 0
average heat transfer coefficients.
Solution
kf
Pr 3
1 ρ u∞ LM L OP
1/ 2

Given : The temperature profile in a thermal

= 0.332
L µ MN 1/2 PQ
boundary layer as kf ρu ∞ L
= 2 × 0.332 Pr1/ 3
L µ
T − Ts 3 y FG IJ
1 y FG IJ 3
kf
T∞ − Ts
=
2 δ th
−
H K
2 δ th H K or h = 2 × 0.332
L
Re L 1/2 Pr 1/3 = 2hx
x=L
x Average Nusselt No. Nu = 2Nux. Ans.
and δth = 4.53
Re x 1/ 2 Pr 1/3
ρ u∞ x
7.8. BOUNDARY LAYER SIMILARITIES
µC p
with Pr = , and Re =
kf µ The eqn. (7.16) and eqn. (7.20) derived earlier for low
To find : The expressions for local and average speed forced convection flow are of practical importance
heat transfer coefficients. in many engineering applications. By close examination
of above two equations, we find that the two equations
Analysis : The local heat transfer coefficient, hx
is expressed as are of same form. Each equation is characterised by
advection term on the left hand side and a diffusion term
∂T on right hand side : These similarity may be extended
kf
∂y y=0 in a rational manner by non dimensionalizing the
hx =
T∞ − Ts governing equations.
For given temperature profile The boundary layer equations are normalized by
∂T LM 3 1 −
1 3 y2
×
OP defining dimensionless indenpedent variables of the
following form.
∂y y=0
= (T∞ – Ts)
MN 2 δ th 2 δ 3th PQ y=0 x y
x* = , y* = ...(7.21)
3 (T∞ − Ts ) L L
= where L is characteristic length for the surface of interest
2 δ th
such as length of a flat plate. The dependent
3 kf (T∞ − Ts ) 3 kf dimensionless variables may be defined as
Then hx = =
2 (T∞ − Ts ) δ th 2 δ th u v T − Ts
u* = , v* = and T* =
Introducing the expression for δth u∞ u∞ T∞ − Ts
1/2 1/3
...(7.22)
3 kf × Re x Pr where u∞ and T∞ are velocity and temperature of free
hx = ×
2 4.53 x stream fluid respectively, and Ts is the temperature of
the surface. Substituting eqns. (7.21) and (7.22) in
kf
Rex1/2 Pr1/3. Ans.
= 0.332 eqns. (7.12), (7.16) and (7.20) to obtain the corresponding
x
boundary layer equations in non dimensional form as
This expression can be arranged in the
shown in Table 7.1.
dimensionless form as
 246 ENGINEERING HEAT AND MASS TRANSFER

TABLE 7.1. Convection transfer equations and their boundary conditions in non dimensional form

Boundary Transfer equation Boundary conditions Similarity

layer parameter(s)
wall free stream

∂u∗ ∂v∗
Continuity + =0 ...(7.23) — — —
∂x∗ ∂y∗
∂u∗ ∂u∗ ν ∂ 2 u∗
Velocity u∗ + v∗ = u∗ (x∗, 0) = 0
∂x∗ ∂y∗ u∞ L ∂y∗
...(7.24) v∗ (x∗, 0) = 0 u∗(x∗, ∞) = 1 ReL
2
∂T∗ ∂T∗ α ∂ T∗
Thermal u* + v∗ = T∗ (x∗, 0) = 0 T∗ (x∗, ∞) = 1 ReL Pr
∂x∗ ∂y∗ u∞ L ∂y∗2
...(7.25)

From eqns. (7.24) and (7.25), the two similarity 7.8.1. Friction Coefficient
parameters may be concluded. These similarity Shear stress at the surface is given by eqn. (7.8)
parameters are important, because they permit us to
apply solutions from one configuration to another ∂u
τs = µ
geometrical similar configuration under entirely ∂y y=0
different conditions. For example, if the Reynolds Substituting u and y from eqns. (7.21) and (7.22),
number is same, the dimensionless velocity distribution we get
for air, water and glycerine etc. flowing over a flat plate
will be the same at a given value of x∗. u∞ ∂u *
τs = µ ...(7.31)
Eqn. (7.23) indicates that v∗ is related to u∗, x∗ L ∂y * y* = 0
and y∗, thus
Defining local skin friction coefficient using
v* = f1(u∗, x∗, y∗) ...(7.26)
eqn. (7.7)
Similarly, from eqn. (7.24), u∗ can be expressed
in the form τs
Cfx = ...(7.32)
u∗ = f2(x∗, y∗, ReL) ...(7.27) ρ u∞2 /2
ν Substituting eqn. (7.31) for τs, we get
The eqn. (7.24) has a quantity on its right
u∞ L 2 µ u∞ ∂u∗ 2 ∂u∗
Cfx = =
hand side. This quantity is a dimensionless group and ρ u∞2 L ∂y∗ y∗ = 0
Re L ∂y∗ y∗ = 0
its reciprocal is well known Reynolds number.
...(7.33)
u∞ L
ReL = ...(7.28) From eqn. (7.33) it is also evident that
ν
Cfx = f3 (x∗, ReL) ...(7.34)
α
From eqn.(7.25) the term is also a It indicates that for flow over bodies of similar
u∞ L
dimensionless group and it may be expressed as shape the local friction coefficient is function of x∗ and
α νFG α IJ FG IJ
1 1
ReL and it is independent of fluid or free stream velocity.

u∞ L
=
u∞ LH×
ν K H K
=
ReL Pr
...(7.29) 7.8.2. Nusselt Number
In convection heat transfer, the local heat transfer
The ratio of two properties (α/ν) is also a
coefficient is expressed as
dimensionless property and its reciprocal is referred as
Prandtl number (Pr) FG ∂T IJ
ν
hx = –
kf
H ∂y K y=0
or Pr = ...(7.30)
α Ts – T∞
 PRINCIPLES OF CONVECTION 247

In non dimensional form Solution

(T∞ − Ts ) ∂T * kf ∂T * Given : Operating conditions of an internally
hx = – kf =
(Ts − T∞ )L ∂y * L ∂y * y* = 0
cooled turbine blade as shown in Fig. 7.13 (a).
y∗ = 0
...(7.35) To find :
In appropriate dimensionless form (i) Heat flux to the blade, when surface
temperature is lowered to 700°C.
hx L ∂T *
Nux = = = f4(x∗, ReL, Pr) (ii) Heat flux to larger identical turbine blade
kf ∂y *
y* = 0 with reduced velocity to 80 m/s.
...(7.36) Schematic :
The quantity hx L/kf is called the local Nusselt
q1
Number. It is equal to dimensionless temperature
Ts = 700°C
gradient at the surface and thus it provides a measure
Air
of convection heat transfer occurring at the surface. The
role of local Nusselt number in thermal boundary layer
T¥ = 1150°C
is same as that of local friction coefficient in velocity
L=
boundary layer. 40
u¥ = 160 m/s mm
The average value of Nusselt number gives
average value of heat transfer coefficient h, that is
independent of x∗ Case 1
hL
Nu = = f5 (ReL , Pr) ...(7.37)
kf q2
Ts = 800°C
Example 7.8. Experimental test on a portion of a turbine Air
blade as shown in Fig 7.13 (a) indicates a heat flux of
95000 W/m2. T¥ = 1150°C

2 L=
q = 95000 W/m u¥ = 80 m/s 80
mm
Ts = 800°C

Case 2
T¥ = 1150°C Coolant
Fig. 7.13 (b)

u¥ = 160 m/s Assumptions :

40
mm (i) Steady state conditions.
(ii) Constant air properties.
Analysis : (i) From eqn. (7.36) for given geometry
Fig. 7.13. (a) Original condition
hx L
The blade is cooled at inside in order to maintain Nux = = f(x∗, ReL, Pr)
kf
its temperature constant at 800°C.
Since there is no change in dimension and envi-
(i) Determine the heat flux to the blade if its
ronmental conditions, thus x∗, ReL, Pr will remain same
temperature is reduced to 700°C by increasing the
even with change in Ts. The Nusselt number is also
coolant flow,
unchanged, thus the local heat transfer coefficient hx
(ii) Calculate the heat flux at same dimensionless
also remains same.
location for a similar turbine blade having a chord
For case 1. The heat flux can be given by
length of 80 mm when the blade operates in an air flow
at T∞ = 1150°C and u∞ = 80 m/s with Ts = 800°C. q1 = hx1 (T∞ – Ts),
 248 ENGINEERING HEAT AND MASS TRANSFER

q1 95000 information about the phenomenon, but facilates the

where hx1 = h = = interpretation of variables to obtain them in certain
T∞ − Ts 1150 − 800
dimensionless groups. The dimensional analysis for
= 271.43 W/m2.K forced and natural convection is discussed in next
and q1 = 271.43 × (1150 – 700) articles.
= 122143 W/m2.K. Ans.
7.9.2. Exact Mathematical Solutions
(ii) For case 2. The blade size is increased to 2L
This method requires solution of simultaneous equations
and free stream air velocity is reduced to one half,
therefore related with fluid motion and energy transfer in moving
fluid. The complete mathematical equations describing
1/2 u∞ (2L)
Re L 2 = = ReL fluid flow and heat transfer can only be written for
ν
Environment is same, thus Pr remains laminar flow. Even for laminar flow, the equations are
unchanged therefore, Nu also remains same quite complicated and their solution is very tedious.
Nu2 = Nu Exact solutions are important, because they serve
as basis for comparison and a check on approximated
hx2 L 2 hL
or = x solutions.
kf kf
7.9.3. Approximate Analysis of Boundary Layers
L 1
or hx2 = hx = 271.43 × In these methods, the detailed mathematical formula-
2L 2
tion for flow in the boundary layer is avoided. But simple
= 135.7 W/m2.K equations are used to approximate momentum and
and the heat flux energy transfer of fluid flow and they are solved by in-
q2 = hx2 (T∞ – Ts) tegral method. The method is relatively simple and
= 135.7 × (1150 – 800) yielding solutions are well agree with exact solutions
= 47500 W/m2. Ans. within certain range. This technique is used to laminar
flow as well as to turbulent flow.

7.9. DETERMINATION OF CONVECTION HEAT 7.9.4. Analogy between Heat and Momentum Transfer
TRANSFER COEFFICIENT It is very useful tool for analysis of the turbulent flow
process. Because our knowledge of turbulent exchange
There are five general methods, that may be used for mechanism is quite limited and thus we cannot write
determination of heat transfer coefficient : equations completely.
1. Dimensional analysis combined with 7.9.5. Numerical Analysis
experimental data.
2. Exact mathematical solution of boundary It can approximate the exact equations. It requires to
layer equations. express the field variables at descrete points in time
and space coordinate. However, the solution can be made
3. Approximate analysis of boundary layer
sufficiently accurate with proper descretization of
equations by integral methods.
problem field. It has one advantage that once the
4. Analogy between heat and momentum solution procedure is programmed, the solution for
transfer. different boundary conditions, property variables and
5. Numerical analysis. so on can easily be handled.
All methods can evaluate the heat transfer
coefficient, but no single method can solve all types of 7.10. DIMENSIONAL ANALYSIS
problems, because each method has its own limitations
that restrict its scope of applications. Dimensional analysis differs from the conventional
methods of approach in which certain equations are
7.9.1. Dimensional Analysis solved for a resulting equation. Instead, it combines
It is a mathematically simple method and has a wide several variables affecting a phenomenon in
range of applications. Its main limitation is that, the dimensionless group, such as Nusselt number, which
obtained results are incomplete and useless without facilitates the interpretation and extends its application
experimental validation. It does not provide any to experimental data.
 PRINCIPLES OF CONVECTION 249

The dimensional analysis does not give any homogeneous i.e., fundamental dimensions of various
information about the nature of phenomenon, hence the terms on two sides of the equation are identical. A
success or failure of the method depends on proper dimensionally homogeneous equation is independent of
selection of affecting variables. It is therefore, important fundamental units, if their units are consistent. It states
to understand the physics of phenomenon before that to a dimensionally homogeneous equation, the
applying dimensional analysis. quantities of the same units can be added, substracted
or equated. Consistency of units demands, if one term
7.10.1. Primary Dimensions and Dimensional Formulae
of equation is measured in a particular unit say m/s
In SI system of units the primary dimension of length then all term in the equation must be measured in m/s.
L, time t, temperature T and mass M are used. The dimensional homogeneity is useful in the following
The dimensional formula in primary dimensions ways :
for a physical quantity is obtained from its definition of (i) It facilates the determination of the
physical laws. For an example, the dimensional formula dimensions of a physical quantity.
for the length of a rod is (L) by definition, for velocity (ii) It helps to check the dimensional consistency
(distance/time) is Lt–1 and so on. of an equation.
The symbols, units and dimensions of commonly (iii) It facilates the conversion of units from one
used quantities in heat transfer analysis are listed in system to another.
Table 7.2. (iv) It provides a step towards dimensional
analysis.
TABLE 7.2. Important physical quantities
with their symbols, units and primary 7.10.3. Rayleigh’s Method of Dimensional Analysis
dimensions In this method, a functional relationship of the
quantities that may influence a dependent variable is
Sr. Quantity Symbol Primary expressed in form of an exponential equation.
No. and units dimensions If y is dependent variable and it depends on
1. Length, diameter L or D, m L variables x1, x2, x3, ..., then the equation in an
2. Time t, s t exponential form
3. Mass m, kg M y = C (x1a, x2b, x3c, ...,) ...(7.38)
4. Temperature T, °C T Where C is dimensional constant, which may be
5. Area A, m2 L2 evaluated from physical characteristic of problems or
6. Volume V, m3 L3 through experimentation, a, b, c,.... are arbitrary
7. Density ρ, kg/m3 ML–3 exponents and are obtained by comparing the exponents
8. Velocity u, v or u∞, m/s Lt –1 of the primary dimensions on two sides of an equation.
9. Acceleration a or g, m/s2 Lt–2 The Rayleigh method does not provide any
10. Force F, N(kg m/s2) MLt–2 information regarding the number of dimensionless
11. Pressure p, N/m2 ML–1 t–2 groups to be obtained as a result of dimensional analysis.
12. Shear stress τ, N/m2 ML–1 t–2 Further, this method can only be used for dimensional
13. Heat transfer rate Q, W(Nm/s) ML2 t–3 analysis of a dependent variable which depends upon
14. Specific heat Cp, J/kg. K L2 T–1 t–3 maximum four independent variables. If number of
15. Dynamic viscosity µ, kg/m.s ML–1 t–1 independent variables exceeds then it becomes tedious
16. Kinematic viscosity ν, m2/s L2 t–1 work to obtain an expression for the dependent variable.
17. Thermal conductivity k, W/m.K ML T–1 t–3 Therefore, this method has become obsolete and is not
18. Thermal diffusivity α, m2/s L2 t–1 favoured for use.
19. Coefficient of expansion β, K–1 T –1 Example 7.9. Find the functional relationship for
20. Heat transfer coefficient h, W/m2.K MT–1 t–3 pressure drop for a fluid flowing through a tube diameter
21. Mass flow rate m
, kg/s M t–1 D, fluid density ρ, fluid velocity u and fluid viscosity µ.
Solution
7.10.2. Dimensional Homogeneity The pressure drop for a flowing fluid in the
A physical equation is the relationship between two or functional form.
more physical quantities. The principle of dimensional ∆p = f(D, ρ, u, µ)
homogeneity states that all equations, describing the Let ∆p = C [Dp, ρq, ur, µs]
behaviour of a physical system must be dimensionally
where C is non dimensional constant.
 250 ENGINEERING HEAT AND MASS TRANSFER

The dimensional equation of above equation in e = 1 – f, d=–c+f

primary dimensions M, L, and t are b=c a=c–1
ML–1 t –2 = C [L p (ML–3)q (Lt –1)r (ML–1 t –1)s] c–1
Therefore, h = C [D u ρ µc c –c+f k1–f Cpf ]
or ML–1 t –2 = C [Mq+s L p–3q + r – s . t –r– s]
Multiplying D/kf on both sides, we get
For dimensional homogeneity, the exponent of
hD LMF ρ uD I F µ C
c
I f
OP
MNGH µ JK GH k
each dimension on both sides of the equation must be
identical. Thus kf
=C
f
p
JK PQ
For M : 1=q+s
or NuD = C ReDc Prf
For L : – 1 = p – 3q + r – s
For t : –2=–r–s ρ uD
where ReD = , Reynolds numbers,
There are four unknowns (p, q, r and s), but three µ
equations. Therefore, it is not possible to find the
µCp
numerical values of p, q, r and s. However, three Pr = , Prandtl number
unknowns can be expressed in terms of fourth variable. kf
Let choose viscosity with exponent ‘s’ and other
hD
exponents are expressed as and NuD = , Nusselt number.
kf
r=2–s
q = 1 – s and p = – s, Thus in forced convection, the Nusselt number is
Substituting these values, we get a function of Reynolds number and Prandtl number.
∆p = C [D–s ρ1–s u2–s µs] 7.10.4. Buckingham π Theorem
L µ OP
=CM
s
× ρu2 = C
ρu 2 When the large physical quantities are involved in a
N ρuD Q Re d s phenomenon, the Rayleigh method of dimensional
analysis becomes tedious. The Buckingham π theorem
ρ uD
where ReD = . is an improvement of Rayleigh method. It is simple and
µ
more systematic for determining the dimensionless
The values of C and s are to be evaluated by
groups.
experimentation.
Example 7.10. Find the functional expression for forced According to Buckingham π theorem, if n
convection heat transfer between a fluid flowing through variables are influencing a phenomenon and if they can
a tube and its wall. be expressed in m primary dimensions (M, L, T and t),
then the dimensionless independent groups formed will
The heat transfer coefficient in forced convection
be n – m. These dimensionless term are called π terms.
is influenced by tube diameter (D), fluid velocity (u), and
The independent dimensionless groups can be arranged
the fluid properties such as density (ρ), dynamic viscosity
as
(µ), thermal conductivity (kf) and specific heat (Cp).
f(π1, π2, π3, ...) = 0 ...(7.39)
Solution
The heat transfer coefficient in the functional In a particular phenomenon, involving seven
form variables and if they can be expressed in four primary
h = f(D, u, ρ, µ, kf , Cp) dimensions, then the number of dimensionless groups
formed are
Let h = C [Da ub ρc µd kfe Cpf ] n–m=7–4=3
where C is non dimensional constant. Hence f(π1, π2, π3) = 0 ...(7.40)
Expressing quantities in terms of primary
or π1 = φ(π2, π3) ...(7.41)
dimensions
Each dimensionless π term is formed by
Mt–3 T –1 = C [La (Lt –1)b (ML–3)c (ML–1 t –1)d
m variables, along with one selected variable from
(MLt –3 T –1)e (L2 t –2 T –1)f
remaining (n – m) variables i.e., each π term involves
For dimensional homogeneity of equation ;
(m + 1) variables. These m variables (equal to primary
M : 1=c+d+e
dimensions) formed a core group, which appear
L : 0 = a + b – 3c – d + e + 2f
repeatedly in each π term, consequently, called as
t : – 3 = – b – d – 3e – 2f
repeating variables. The repeating variables are selected
T : – 1 = – e – f.
among the n variables in such way that
There are four equations and six unknowns, thus
complexity starts in such case. Let expressing the (i) They together must contain all primary
exponents in terms of c and f. dimensions.
 PRINCIPLES OF CONVECTION 251

(ii) They themselves must not form a On solving, we get

dimensionless group. e = – 1, f = – 2, d = 1
(iii) As far as possible, the dependent variable ∆p D
should not be selected as repeating variable. or π2 = D ρ–1 u–2∆p =
ρu 2
(iv) No repeating variable should have the same F ∆p D , µ I = 0.
dimension. Therefore, φ GH ρu ρuD JK
2
(v) The repeating variables should be chosen in
such a way that one variable contains geometrical 7.10.5. Dimensional Analysis for Forced Convection
property, other one contains dynamic (flow) property The forced convection heat transfer phenomenon can
and third one contains fluid property etc. be influenced by the variables given in Table 7.3.
Example 7.11. Reconsider the example 7.10 and develop
TABLE 7.3
an expression for pressure drop.
Solution Sr. Parameters Symbol Primary
No. and unit dimensions
Given : The pressure drop for flow through tube
as 1. Tube Diameter D, m L
∆p = f(D, ρ, u, µ) or φ(∆p, D, ρ, u, µ) = 0 (Characteristic length)
2. Fluid density ρ, kg/m3 M L–3
There are five (n) variables affecting a
phenomenon and they can be expressed in three primary 3. Fluid viscosity µ, kg/m/s M L –1t–1
dimensions (M, L, and t). Therefore, the number of 4. Fluid velocity u∞, m/s Lt–1
dimensionless π groups to be formed are 5. Fluid thermal kf , W/m.K M Lt–3T–1
conductivity
n – m = 5 – 3 = 2 or φ(π1, π2) = 0
6. Heat transfer h, W/m2.K M t–3T–1
Each π term consists of m = 3 common variable, coefficient
called repeating variables along with one selected 7. Fluid specific heat Cp, J/kg.K L2t–2T–1
variable.
Let π1 = Da ρb uc µ These seven variables are expressed in four
π2 = Dd ρe uf ∆p primary dimensions (M, L, T, t), therefore, according to
Buckingham π theorem, the independent dimensionless
Expressing the each variable in their primary groups formed are :
dimensions for each π term.
= No. of variable affecting the phenomenon
Therefore, π1 = La (ML–3)b (Lt–1)c (ML–1 t –1)
or M0 L0 t0 = Mb+ 1 La–3b+c–1, t –c–1 – No. of primary dimensions used
For dimensional homogeneity the exponent of = 7 – 4 = 3 (i.e., π1, π2, π3)
M, L, t must be equal to zero. Writing these three groups as,
M: 0=b+1 π1 = Da, ρb, µc, kfd, u∞
L : 0 = a – 3b + c – 1 π2 = De, ρf, µg, kfh, Cp
t : 0=–c–1
π3 = Di, ρj, µk, kfl, h
Solving these simultaneous equations, we get
where D, ρ, µ, kf form a core group (repeating variables)
c = – 1, b = – 1, a = – 1 and u∞, Cp and h are as selected variables.
or π1 = D–1 ρ–1 u–1 µ
Since the groups π1, π2, π3 are dimensionless,
µ 1 hence certain exponents are applied on the repeating
or π1 = =
ρuD Re D variable, which are to be determined,
Similarly π2 = Ld (ML–3)e (Lt–1) f (ML–2 t –2) (i) Expressing the variable in their primary
dimensions for π1,
or M0 L0 t0 = Me+1 Ld – 3e+f – 2 t–f– 2
For dimensional homogeneity π1 = La . (ML–3)b . (ML–1 t–1)c . (MLt–3T–1)d . (Lt–1)
M: 0=e+1 or M0L0T0t0 = Mb+c+d . La–3b–c+d+1 . T–d . t–c–3d–1
L : 0 = d – 3e + f – 2 Separating the exponents for dimensional
t : 0=–f–2 homogeneity
 252 ENGINEERING HEAT AND MASS TRANSFER

M: 0=b+c+d 7.10.6. Dimensional Analysis for Natural Convection

L : 0 = a – 3b – c + d + 1 The relations used for the natural convection are based
T : 0=–d on practical observations. Hence the dimensional
t : 0 = – c – 3d – 1 analysis is also useful in natural convection. The
parameters which influence the natural convection
Solvingthese simultaneous equations, we get
phenomenon are listed below with their primary
d = 0, c = – 1
dimensions :
b = 1, a = 1
TABLE 7.4
Hence the dimensionless group formed is,
Dρu∞
π1 = = ReD (Reynolds number) Sr. Parameters Symbol Primary
µ No. and unit dimensions
...(7.42)
(ii) Expressing the primary dimension for 1. Characteristic length Lc, m L
variables of π2, 2. Fluid density ρ, kg/m3 M L–3
π2 = Le . (ML–3) f . (ML–1t–1) g 3. Fluid viscosity µ, kg/m.s M L–1t–1
4. Temperature difference ∆T, °C T
. (MLt–3T–1)h . (L2t–2T–1)
5. Coefficient of volumetric
or M0L0T0t0 = Mf+g+h . Le–3f–g+h+2 . T–h–1 . t–g–3h–2 expansion β, K–1 T–1
Separating the exponents for dimensional 6. Gravitational acceleration g, m/s2 L t–2
homogeneity. 7. Fluid thermal conductivity kf , W/m.K MLt–3T–1
M: 0=f+g+h 8. Heat transfer coefficient h, W/m2.K M t–3T–1
L : 0 = e – 3f – g + h + 2 9. Fluid specific heat Cp , J/kg.K L 2t –2T–1
T : 0=–h–1
t : 0 = – g – 3h – 2 Out of these nine variables, the product of gβ∆T,
Solving these simultaneous equations, we get represents the buoyancy force and considered as a one
h = – 1, g = 1 variable. Thus the variable, affecting the natural
f = 0, e = 0 convection are now remaining only seven and these can
Hence the dimensionless group formed is, be represented by four primary dimensions (M, L, T, t),
therefore, the independent dimensionless groups formed
µC p
π2 = = Pr (Prandtl Number) ...(7.43) are :
kf = No. of variables affecting the phenomenon
(iii) Expressing the primary dimension for – No. of primary dimensions used
variables of π3, = 7 – 4 = 3 (i.e., π1, π2, π3)
π3 = Li . (ML–3) j . (ML–1t–1) k . (MLt–3T–1) l . (Mt–3T–1) Writing these three groups as,
or M0L0T0t0 = Mj+k+l+1 . Li–3j–k+l . T–l–1 . t–k–3l–3
π1 = Lca , ρb, µc, kfd, gβ∆T
Separating the exponents for dimensional
homogeneity π2 = Lce , ρf, µg, kfh, Cp
M: 0=j+k+l+1 π3 = Lci , ρj, µk, kfl, h
L : 0 = i – 3j – k + l where Lc, ρ, µ, kf form a core group and gβ∆T, Cp and h
T : 0=–l–1 are as selected variables.
t : 0 = – k – 3l – 3 (i) Expressing the variable in their primary
Solving these simultaneous equations, we get dimensions for π1,
l = – 1, k = 0 π1 = La.(ML–3)b.(ML–1t–1)c.(MLt–3T–1)d.(Lt–2)
j = 0, i=1 or M L T t0 = Mb + c + d . La – 3b – c + d + 1 . T – d.t – c – 3d – 2
0 0 0
Hence the dimensionless group formed is, Separating the exponents for dimensional
hD homogeneity
π3 = = NuD (Nusselt Number)
kf M: 0=b+c+d
...(7.44) L : 0 = a – 3b – c + d + 1
Hence for forced convection, T : 0=–d
NuD = f(ReD, Pr) ...(7.45)
t : 0 = – c – 3d – 2
 PRINCIPLES OF CONVECTION 253

Solving these simultaneous equations, we get 7.11. PHYSICAL SIGNIFICANCE OF THE

d = 0, c=–2 DIMENSIONLESS PARAMETERS
b = 2, a=3
Hence the dimensionless group formed is, 7.11.1. Reynolds Number
ρ 2
( gβ∆T)L3c ( gβ∆T)L3c It is the ratio of inertia forces to viscous forces in the
π1 = 2
= 2 velocity boundary layer. It is used in forced convection
µ ν
and approximated as :
= GrL (Grashof Number) ...(7.46)
(ii) Expressing the primary dimensions for Inertia forces ρu∞ L c u∞ L c
Re = = =
variables of π2, Viscous forces µ ν
π2 = Le . (ML–3) f . (ML–1t–1) g . (M L t–3T–1) h where, Lc = characteristic length of flow geometry, m ;
. (L2t–2T–1) = x, distance from the leading edge in the
flow direction for a flat plate;
or M0L0T0t0 = M f+g+h . Le –3f–g+h+2 . T –h–1 . t –g–3h–2
= D, diameter for flow through or across a
Separating the exponents for dimensional
cylinder and a sphere;
homogeneity
u∞ = free stream velocity, m/s ;
M: 0=f+g+h
ρ = fluid density, kg/m3 ;
L : 0 = e – 3f – g + h + 2
µ = dynamic viscosity of fluid, Ns/m2 or kg/m/s ;
T : 0=–h–1
ν = µ/ρ = kinematic viscosity of fluid, m2/s.
t : 0 = – g – 3h – 2
Solving these simultaneous equations, we get The Reynolds number is a dimensionless
quantity. It characterises the type of flow, whether it is
h = – 1, g = 1
laminar or turbulent flow.
f = 0, e = 0
Hence the dimensionless group formed is, 7.11.2. Critical Reynolds Number Recr
µC p It is a value of Reynolds number, where boundary layer
π2 = = Pr (Prandtl Number) changes from laminar to turbulent nature. It is denoted
kf
by Recr.
...(7.47)
(iii) Expressing the primary dimensions for For flow over a flat plate, the transition from
variables of π3, laminar to turbulent boundary layer occurs roughly
π3 = Li. (ML–3) j. (ML–1t–1)k. (MLt–3T–1)l when critical Reynolds number is
. (Mt–3T–1) Recr ≥ 5 × 105 ...(7.50)
or 0 0
MLTt =M 0 0 j+k+l+1 Li–3j–k+l T–l–1 t–k–3l–3 In fluid flow through tubes, the Reynolds number
Separating the exponents for dimensional is also used to characterized the fluid flow. The transition
homogeneity from laminar to turbulent boundary layer occurs, when
M: 0=j+k+l+1 u∞ D
ReD, cr = ≥ 2300 ...(7.51)
L : 0 = i – 3j – k + l ν
T : 0=–l–1 These are generally accepted values of critical
t : 0 = – k – 3l – 3 Reynolds numbers, which may vary with surface
Solving these simultaneous equations, we get roughness, level of turbulence and the variation of
pressure along the flow.
l = – 1, k = 0
j = 0, i = 1 7.11.3. Prandtl Number
Hence the dimensionless group formed is, It is defined as the ratio of the momentum diffusivity ν
hL c to the thermal diffusivity α or
π3 = = NuL (Nusselt Number)
kf Momentum diffusivity ν µρC p µC p
Pr = = = =
...(7.48) Thermal diffusivity α ρkf kf
Hence for free convection, It is a dimensionless property, a function of
NuL = f(GrL, Pr) ...(7.49) temperature. It provides a measure of relative
 254 ENGINEERING HEAT AND MASS TRANSFER

effectiveness of momentum and energy transfer in the where, ∆T = temperature difference (Ts – T∞) between
velocity and thermal boundary layers, respectively. wall surface and fluid, K.
For gases Pr ≅ 1 ; i.e., both momentum and heat h = heat transfer coefficient; W/m2.K.
diffusion through the fluid take place at the same rate. kf = thermal conductivity of the fluid ; W/m.K.
For liquid metal Pr << 1 ; indicates heat Lc = characteristic length of fluid flow, m
diffuses in the fluid very quickly, and for oils, Pr >> 1;
indicates heat diffusion is very slow in the fluid relative Based on the interpretation, the value of Nu as
to momentum. unity indicates that there is no convection, the heat
transfer is by pure conduction in the boundary layer.
Consequently, the thermal boundary layer is
much thicker for liquid metals, much thinner for oils Large value of Nu indicates large convection in the fluid.
relative to velocity boundary layer as shown in Fig. 7.7. 7.11.6. Stanton Number
Further, the thicknesses of two boundary layers can be
It is the ratio of the heat transfer at the surface to that
related as
transported by fluid by its thermal capacity.
δ th ( x) Heat flux to the fluid
= Pr n where n is the exponent ...(7.52) Stx =
δ ( x) Heat transfer capacity of fluid
h ∆T h
7.11.4. Grashof Number = = ...(7.54)
ρC pu∞ ∆T ρC pu∞
It is defined as the ratio of the buoyancy forces to the
viscous forces acting in the fluid layer. It is used in free Mathematically, it is the ratio of Nusselt number
convection and its role is same as that of Reynold number and product of Reynolds number and Prandtl number
in forced convection. The Grashof number characterises and it is also expressed as
the type of boundary layer developed in natural convec- Nu x
Stx = ...(7.55)
tion heat transfer. It is denoted by Gr and expressed as Rex Pr

gβ∆TL c3 7.11.7. Peclet Number

Gr = ...(7.53)
ν 2 It is the ratio of heat transfer by convection to heat
where, g = acceleration due to gravity, m/s2, transfer by conduction. It is denoted by Pe and expressed
β = coefficient of volumetric expansion as
= 1/(Tf + 273), K–1, Heat transfer by convection
Pe =
∆T = temperature difference between surface Heat transfer by conduction
and fluid, °C or K, mC p ∆T ρVC pL
= = ...(7.56)
T + T∞ kf ∆T/L kf
Tf = mean film temperature = s °C,
2 Mathematically, the Peclet numbers is product
ν = kinematic viscosity of fluid, m2/s, of Reynolds number and Prandtl number.
Lc = characteristic length of the body, m Pe = Re.Pr ...(7.57)
= height, L for vertical plates and cylinders, 7.11.8. Graetz Number
= diameter, D for horizontal cylinder and It is a dimensionless number used in study of stream
sphere, line fluid flow. It is the ratio of fluid stream thermal
capacity of fluid flowing per unit length thermal con-
Surface area A s
= = , for other geometries ductivity of fluid. It is denoted by Gz and expressed as
Perimeter P
Thermal capacity of fluid per unit length
For free convection, the transition from laminar Gz =
to turbulent occurs, when Grcr ≈ 10 9. Thermal conductivity

Cp
m π D
7.11.5. Nusselt Number = = Re . Pr . ...(7.58)
kf x 4 x
It is defined as the ratio of convection heat flux to
conduction heat flux in the fluid boundary layer or where x = hydrodynamic entry length,
D = inside diameter of the tube.
Convection heat flux h∆T hL c
Nu = = = Generally it is associated with thermal entry
Conduction heat flux kf ∆T/L c kf
length of a fully developed flow through tubes.
 PRINCIPLES OF CONVECTION 255

Example 7.12. Calculate the approximate The density of air at 5 bar and 400°C (= 673 K)
Reynolds numbers and state if the flow is laminar or p 5 × 100 kPa
turbulent for the following : ρ=
RT
=
b g
0.287 kJ / kg . K × (673 K)
(i) A 10 m long yatch sailing at 13 km/h in sea = 2.588 kg/m3
water, ρ = 1000 kg/m3 and µ = 1.3 × 10–3 kg/ms. The viscosity at 400°C
(ii) A compressor disc of radius 0.3 m rotating at
15000 r.p.m. in air at 5 bar and 400°C and 1.46 × 10 − 6 × (673) 3 / 2
µ=
1.46 × 10 − 6 T 3/2 (110 + 673)
µ= kg/ms. = 3.3 × 10–5 kg/ms
(110 + T)
(iii) 0.05 kg/s of CO2 gas at 400 K flowing in a (a) The Reynolds number
20 mm dia. pipe and ρ u∞ D 2.588 × 471.23 × 0.6
1.56 × 10 − 6 T 3/2 Re = =
µ= kg/ms. µ 3.3 × 10 − 5
(233 + T)
(N.M.U., May 2002) = 22.17 × 106. Ans.
Solution (b) The Re > 5 × 105, thus flow is turbulent. Ans.
(i) Given : Yatch sails on sea water (iii) Given : CO2 gas
L = 10 m, m = 0.05 kg/s, T = 400 K,
3
13 × 10 Di = 20 mm = 20 × 10–3 m,
u∞ = 13 km/h = = 3.61 m/s
60 × 60
1.56 × 10 − 6 T 3/2
ρ = 1000 kg/m3, and µ= kg/ms.
(233 + T)
µ = 1.3 × 10–3 kg/ms.
To find :
To find :
(a) Reynolds number, (a) Reynolds number,
(b) Type of flow. (b) Type of flow.
Analysis : (a) The Reynolds number can be Analysis : At 400 K, the density of CO2
calculated as
ρ = 1.3257 kg/m3
ρu∞ L 1000 × 3.61 × 10
Re = =
µ 1.3 × 10 − 3 (a) The Reynolds number for pipe flow can also
= 2.78 × 10 7. Ans. be calculated as
(b) The Re > 5 × 105, thus flow is turbulent. Ans. 4m

Re =
(ii) Given : A compressor disc with π Di µ
ro = 0.3 m or D = 0.6 m
1.56 × 10 − 6 T 3/2
N = 15000 rpm, where µ=
(233 + T)
p = 5 bar, T = 400°C
1.46 × 10 − 6 T 3/2 1.56 × 10 − 6 × (400) 3/2
Then µ=
and µ= (233 + 400)
(110 + T)
To find : = 1.97 × 10–5 kg/ms.
(a) Reynolds number, 4 × 0.05
Then Re =
(b) Type of flow. π × 20 × 10 − 3 × 1.97 × 10 − 5
Assumption : For air R = 0.287 kJ/kg.K. = 1.61 × 105. Ans.
Analysis : The equivalent linear velocity of
(b) Re ≥ 2300 for tube flow, thus the flow is
compressor disc
πDN π × 0.6 × 1500 turbulent. Ans.
u∞ = = = 471.23 m/s
60 60
 256 ENGINEERING HEAT AND MASS TRANSFER

Example 7.13. Calculate the approximate Grashof 0.4 × 0.2

number and state if the flow is laminar or turbulent for = = 0.067 m
2 × (0.4 + 0.2)
the following :
(a) A central heating radiator, 0.6 m high with a
9.81 × (854)2 × 0.7 × 10 − 3
surface temperature of 75°C in a room at 18°C,
× (75 − 40 ) × (0.067 )3
(ρ = 1.2 kg/m3, Pr = 0.72, and µ = 1.8 × 10–5 kg/ms). (i) Gr =
(3.56 × 10 − 2 )2
(b) A horizontal oil sump with a surface
temperature of 40°C, 0.4 m long and 0.2 m wide = 4.1 × 104. Ans.
containing oil at 75°C. Take ρ = 854 kg/m3, Pr = 546, (ii) Gr < 109, thus the flow is laminar
β = 0.7 × 10–3 K–1 and µ = 3.56 × 10–2 kg/ms. (c) Air
(c) Air at 20°C (ρ = 1.2 kg/m3, Pr = 0.72 and T∞ = 20°C, ρ = 1.2 kg/m3,
µ = 1.8 × 10 –5 kg/ms) adjacent to a 60 mm dia. Pr = 0.72
horizontal light bulb, with a surface temperature of
µ = 1.8 × 10–5 kg/ms, Lc = D = 60 mm,
90°C. (N.M.U., May 2002)
Ts = 90°C
Solution
90 + 20
The properties are given, thus the Grashof Tf = = 55°C,
2
number for any flow situation can be calculated as
1
g β∆T L3c gρ2 β∆T L3c β= = 3.049 × 10–3 K–1
Gr = = 55 + 273
ν2 µ2
where Lc = significant length of the body. 9.81 × (1.2)2 × 3.049 × 10 − 3
(a) Lc = 0.6 m, ∆T = 75 – 18 = 57°C × (90 − 20) × (60 × 10 − 3 )3
ρ = 1.2 kg/m3, Pr = 0.72, Gr =
(1.8 × 10 −5 )2
–5
µ = 1.8 × 10 kg/ms
Mean film temperature, = 2.0 × 106, Laminar. Ans.
75 + 18 Example 7.14. Calculate the Nusselt number in
Tf = 46.5°C
2 following cases :
1 1 (i) A horizontal electronic component with a
β= = surface temperature of 35°C, 5 mm wide and 10 mm
Tf + 273 46.5 + 273
long, dissipating 0.1 W heat by free convection from its
= 3.13 × 10–3 K–1 one side into air at 20°C. Take for air k = 0.026 W/m.K.
The Grashof number (ii) A 1 kW central heating radiator 1.5 m long
9.81 × (1.2) 2 × 3.13 × 10 − 3 × 57 × (0.6) 3 and 0.6 m high with a surface temperature of 80°C,
GrL = dissipating heat by radiation and convection into room
(1.8 × 10 −5 ) 2
at 20°C (k = 0.026 W/m.K, assume black body radiation
= 1.68 × 109. Ans.
and σ = 5.67 × 10–8 W/m2.K4).
The GrL > 109, the flow is turbulent. Ans.
(iii) Air at 6°C (k = 0.024 W/m.K) adjacent to a
(b) A horizontal oil sump
wall 3 m high and 0.15 m thick made of brick with
Ts = 40°C T∞ = 75°C, k = 0.3 W/m.K, the inside temperature of the wall is
L = 0.4 m, w = 0.2 m, 18°C, the outside wall temperature is 12°C.
ρ = 854 kg/m ,3 Pr = 546,
Solution
–3
β = 0.7 × 10 K –1
(i) Given : A horizontal electronic component
µ = 3.56 × 10–2 kg/ms. w = 5 mm, L = 10 mm,
For horizontal plate, the characteristic length Q = 0.1 W, Ts = 35°C,
As L×w T∞ = 20°C, kf = 0.026 W/m.K.
Then Lc = =
P 2(L + w)
 PRINCIPLES OF CONVECTION 257

Analysis : This heat is also transfered by convection, thus

Q = h As (Ts – T∞) Q
= h(Ts – T∞)
0.1 = h × (10 mm × 5 mm × 10–6) × (35 – 20) A
or h = 133.33 W/m2.K On inner surface
The significant length 12 = h1 × (18 – 6) or h1 = 1 W/m2.K
−6
As 10 mm × 5 mm × 10
Lc = = h1L c 1× 3
P 2 × (10 mm + 5 mm) × 10 −3 and Nu = = = 125. Ans.
kf 0.024
= 1.67 × 10–3 m
The Nusselt number On outer wall surface
hL c 133.33 × 1.67 × 10 −3 12 = h2 × (12 – 6) or h2 = 2 W/m2.K
Nu = =
kf 0.026 h2 L c 2×3
= 8.54. Ans. and Nu = = = 250. Ans.
kf 0.024
(ii) Given : Central heating radiator
Q = 1 kW = 103 W,
w = 1.5 m,
7.12. TURBULENT BOUNDARY LAYER HEAT
L = Lc = 0.6 m,
TRANSFER
Ts = 80°C = 353 K, The flow of fluid in the boundary layer is more often
T∞ = 20°C = 293 K, turbulent rather than laminar as shown in Fig. 7.14. In
kf = 0.026 W/m.K, the turbulent flow, the transport mechanism is added
σ = 5.67 × 10–8 W/m2.K4. by random fluctuation of lumps of fluid. The irregular
Analysis : The radiation heat transfer for black velocity fluctuations are superimposed upon the motion
surface of main stream and these fluctuations are primarily
Qrad = σ As (Ts4 – T∞4) responsible for transfer of heat and momentum. The
rates of momentum and heat transfer in the turbulent
= (5.67 × 10–8) × (1.5 × 0.6)
flow and associated friction and heat transfer coefficients
× (3534 – 2934)
are many times more than the laminar flow, because of
= 416.2 W better mixing in which lumps of fluid collide with one
Heat transfer by convection another randomly and make multidirectional flow and
Qconv = Q – Qrad = 1000 – 416.2 = 583.7 W mix the fluid effectively.
Then Qconv = h As (Ts – T∞) In the turbulent flow, the instantaneous fluid
583.7 = h × (1.5 × 0.6) × (80 – 20) currents are highly torn and fluctuating randomly and
or h = 10.80 W/m2.K it is very difficult to trace the path of an individual fluid
The Nusselt number element. This behaviour is shown in Fig. 7.15, which
hL c 10.80 × 0.6 plots arbitrary flow property P as a function of time at
Nu = = = 249.4 Ans. some location in a turbulent boundary layer. The
kf 0.026
property P could be a velocity component, or fluid tem-
(iii) Given : Air flow adjacent to a wall perature at any instant. The time mean value and
T∞ = 6°C, kf = 0.024 W/m.K, fluctuating component may be represented as P and P′,
Lc = H = 3 m, L = 0.15 m, respectively for steady flow. The instantaneous velocity
k = 0.3 W/m.K, T1 = 18°C, components u and v can be expressed in the form
T2 = 12°C.
u = u + u′ and v = v + v′
Analysis : The heat transfer rate per m2 by steady
state conduction, through the wall Similarily instantaneous temperature can be
expressed as
Q k(T1 − T2 ) 0.3 × (18 − 12)
= =
A L 0.15 T = T + T ′ and so on
= 12 W/m2
 258 ENGINEERING HEAT AND MASS TRANSFER

Y u¥ Time average of instantaneous rate of x

directional momentum transfer per unit area
X

u
Turbulent
u=u
v=–
– + u¢

v + v¢
–
T = T + T¢
τt = –
1 t
t 0 z
(ρ v′ )(u + u′ ) dt ...(7.60)
It is also called “apparent turbulent shear stress
Buffer
or Reynolds stress” and can be rearranged as

z z
–
layer r = r + r¢
etc. 1 t 1 t
Laminar τt = – ρ v′ u dt − ρ v′ u′ dt
sub layer t 0 t 0

Fig. 7.14. Velocity profile in turbulent boundary

= – ρ v′ u − ρ u′ v′ ...(7.61)
layer on a flat plate Since u is mean velocity, thus constant and the
p P¢ time averaged ρv– ′ is zero, therefore

– τt = – ρ u′ v′ ...(7.62)
P
As stated above u′v′ is not zero, but it is negative,
Flow thus the turbulent shear stress is positive and analogous
property
– to laminar shear stress
P = P + P¢
du du
τl = µ = ρν ...(7.63)
time t dy dy
Fig. 7.15. Property variation with time at some Actually, the laminar shear stress τl is true stress
points in turbulent boundary layer whereas the Reynolds stress τt is the stress to account
for the effects of momentum transfer due to turbulence.
Y
Mean velocity u Thus the total shear stress
du
u
l τtotal = τl + τt = µ − ρ u′ v′ ...(7.64)
A A¢ dy
u¢ l
7.12.1 Prandtl Mixing Length Concept
Turbulent lump y Prandtl postulated that the fluctuations of fluid lumps
in turbulent flow on average are analogous to motion of
molecules in a gas. The Prandtl mixing length l is the
distance travelled on an average by the turbulent lumps
X
of fluid in direction perpendicular to mean flow before
Fig. 7.16. Turbulent shear stress and mixing length coming to rest. The Prandtl mixing length l is analogous
Consider a turbulent lump crosses the plane to the mean free path of molecules in a gas.
A – A′ as shown in Fig. 7.16. The fluctuating velocity Let us imagine a turbulent lump which is located
components continuously transport mass and at a distance l above or below plane A – A′ as shown in
therefore, momentum across a plane A – A′ normal to y Fig. 7.16. The fluid lumps move back and forth across
direction. the plane and increase turbulent shearing stress effect.
The instantaneous mass transport per unit area At distance y + l, the velocity of fluid would be
across the plane = ρv′ approximately
Instantaneous rate of transfer of x directional du
u(y + l) = u(y) + l
momentum per unit area represents shear stress. dy
τ′ = – ρv′ ( u + u′ ) ...(7.59) and at distance y – l
The negative sign is inducted, because, when a du
u(y – l) = u(y) – l
turbulent lump moves upward (v′ > 0), it enters the dy
region of higher u , it will tend to slow down the The Prandtl demonstrated that the turbulent
fluctuations in u′, thus u′ < 0 and vice-versa so a positive fluctuation u′ is proportional to mean of above two
v′ is associated with negative u′, therefore, the product quantities, or
u′v′ is a negative quantity.
 PRINCIPLES OF CONVECTION 259

du caused by temperature difference between time mean

u′ = l ...(7.65) temperature of two planes. The turbulent heat transfer
dy
rate per unit area
He also postulated that v′ would be of same order
of magnitude as u′, i.e., Qt ∂u ∂ T
= – ρCp v′ T′ = – ρCp lT2 .
du A ∂y ∂y
v′ = l
dy ∂T
= − ρC p εH ...(7.73)
The turbulent shear stress eqn. (7.62) ∂y
FG du IJ 2
du where εH is eddy or turbulent diffusivity of heat, or
τt = – ρ u′ v′ = ρ l2
H dy K = ρ εM
dy εH = lT2
∂u
∂y
...(7.74)
...(7.66)
where εM is eddy or turbulent viscosity, or ∂T
The minus sign is due to second law, because
du ∂y
εM = l2 ...(7.67) would be negative.
dy
The total rate of turbulent heat transfer per unit
The eddy viscosity εM is analogous to kinematic
area
viscosity ν. But ν is a physical property, while εM is not,
and it depends on dynamics of flow. FG Q IJ
H AK
t
= Molecular conduction/area
Total shearing stress total
+ Turbulent heat transfer through eddies/area
du du
τtotal = µ + ρ εM
dy dy ∂T ∂T
=– k – ρCpεH
du du ∂y ∂y
= ρν + ρ εM ∂T ∂T
dy dy = – ρCp α – ρCpεH
du ∂y ∂y
= ρ(ν + εM) ...(7.68)
dy ∂T
= – ρCp (α + εH) ...(7.75)
For turbulent flow ∂y
where α = k/ρCp.
du
εM >> ν and τtotal ≈ ρεM ...(7.69) The contribution to the total heat transfer rate
dy
by molecular conduction is proportional to α, and
For laminar flow
turbulent contribution is proportional to εH.
du For all fluids except liquid metals
εM = 0 and τtotal ≈ ρν ...(7.70)
dy εH >> α in turbulent flow and
For buffer layer (transition zone),
FQ I ∂T
du
H AK ≈ − ρC p εH ...(7.76)
t
τtotal = ρ(ν + εM) ∂y
dy total
For laminar flow εH = 0 and
7.12.2. Turbulent Heat Transfer
FG Q IJ ∂T
H AK
t
The heat transfer in turbulent flow is analogous to = − ρC p α ...(7.77)
∂y
momentum transfer. The instantaneous turbulent heat total

transfer rate per unit area can be expressed as In transition zone

Qt
A
=
1
t z
0
t
(ρv′ ) C p (T + T ′ ) dt = ρ C p v′ T ′
...(7.71)
FQ I
H AK
t
total
∂T
= − ρ C p (α + ε H )
∂y
The ratio of eddy viscosity to eddy thermal
Using Prandtl mixing length concept, the diffusivity is called turbulent Prandtl number
temperature fluctuations T′ can be related with time ε
mean temperature gradient as Prt = M ...(7.78)
εH
dT This definition is analogous to definition of
T′ ≈ lT ...(7.72)
dy Prandtl number
when a fluid lump in turbulent flow migrates plane ν
Pr =
A – A′ by a distance ± lT, the resulting fluctuation is α
 260 ENGINEERING HEAT AND MASS TRANSFER

But the Prandtl number Pr and turbulent Prandtl The eqn. (7.81) is called Reynolds analogy. It is
number are not same. The Prandtl number Pr is a satisfactory for gases Pr = 1. Colburn had corrected to
dimensionless physical property of fluid. However, the fluids having Prandtl number ranging 0.6 to 50 and it
turbulent Prandtl number Pr is a property of flow field is modified to
more than a field. Various models have been developed C fx
for evaluating of Prt. Reynolds model is simplest one, Stx Pr2/3 = ...(7.82)
he assumed Prt = 1 i.e., εH = εM. 2
where subscript x represents the distance from the
However, the numerical values of Prt may vary
leading edge. The expression (7.82) is referred as
between 1 and 2.
Reynolds Colburn analogy for flow over flat pate and
For Prt = 1, the turbulent heat flux eqn. (7.76) Stx Pr2/3 is called Colburn’s factor.
and turbulent shear stress eqn. (7.69) can be related as
For average properties (average heat transfer
∂T coefficient and friction coefficient), the above equation
ρC p εH
Qt ∂y is also valid in the form
=−
Aτ t ∂u Cf
ρεM
∂y St Pr2/3 = ...(7.83)
2
Qt ∂T valid for all types of flow over a flat plate.
or = − τt C p ...(7.79)
A ∂u
This relation was first introduced in 1874 by
7.14. MEAN FILM TEMPERATURE AND BULK
Reynolds and therefore, called Reynolds analogy for
turbulent flow. This analogy however, does not hold good MEAN TEMPERATURE
in viscous sublayer, where the flow is laminar.
For external flows such as flow over a flat plate, flow
across a cylinder or a sphere, the fluid properties like ρ,
7.13. REYNOLDS COLBURN ANALOGY FOR Cp, kf, and µ are generally evaluated at mean film
TURBULENT FLOW OVER A FLAT PLATE temperature Tf or
Ts + T∞
To obtain the heat transfer rate for turbulent flow Tf = ...(7.84)
2
over a flat plate with Prt = 1, the eqn. (7.79) can be ar-
where Ts = surface temperature, °C and
ranged as
T∞ = free stream temperature of fluid, °C
Qs qs
du = − dT or du = − dT For internal flows such as flow through tubes,
Aτ sC p τ sC p ducts etc, the fluid properties are evaluated at mean of
where subscript s indicates that q and τ are taken at the bulk inlet and outlet temperature, Tm or
surface of the plate. Integrating above equation between
Tb, in + Tb, out
u = 0, T = Ts and u = u∞, T = T∞ yields to Tm = ...(7.85)
2
qs qs τs C p
u∞ = Ts – T∞ or = where Tb, in = Bulk mean inlet temperature, °C, and
τ sC p Ts − T∞ u∞ Tb, out = Bulk mean outlet temperature, °C.
Introducing local heat transfer coefficient and Sometimes, the correlations may specify some
friction coefficient as other temperature ; such as for internal flow it may be
qs C fx the mean of fluid temperature Tm and pipe wall surface
hx = and τs = ρu∞ 2 temperature Ts. If temperature differences (surface to
Ts − T∞ 2
fluid, inlet to outlet) are small enough, then changes in
C fx
Then hx = Cp ρu∞ ...(7.80) the fluid properties are negligible and the choice of
2 particular temperature becomes unimportant, provid-
hx C fx ing consistancy is maintained.
or =
ρ u∞ C p 2 Example 7.15. Atmospheric air at 400 K flows with a
velocity of 4 m/s along a flat plate, 1 m long, maintained
C fx
or Stx = ...(7.81) at an uniform temperature of 300 K. The average heat
2 transfer coefficient is estimated to be 7.75 W/m2.K. Using
Nu x hx Reynolds Colburn analogy, calculate the drag force
where Stx = = is called Stanton number.
Re x Pr ρ C p u∞ exerted on the plate per metre width.
 PRINCIPLES OF CONVECTION 261

Solution direction normal to surface is observed, is called thermal

Given : Flow along a flat plate boundary layer (δth).
T∞ = 400 K, Ts = 300 K, The fluid flow over a flat plate starts as a laminar
L = 1 m, w = 1 m, boundary layer, in which the fluid motion is highly
2
h = 7.75 W/m .K, u∞ = 4 m/s. ordered and fluid flow can be identified in stream lines.
To find : Drag (shear) force exerted on the plate. The fluid flow becomes turbulent after some distance
from the leading edge, in which large velocity
Analysis : Reynolds Colburn analogy for flow over
fluctuations and highly disordered motion of the fluid
a flat plate is given by
are observed. The intense mixing in turbulent flow
h Pr 2 / 3 C f enhances both the drag force and heat transfer. The flow
St Pr2/3 = =
ρ C pu∞ 2 regime depends mainly on Reynolds number, expressed
The physical properties of atmospheric air at as
mean film temperature Inertia forces u∞ x
Re = =
Ts + T∞ 300 + 400 Viscous forces ν
Tf = = = 350 K
2 2 where u∞ is free stream fluid velocity, x is the distance
ρ = 0.998 kg/m3, from leading edge and ν is kinematic viscosity.
Cp = 1009 J/kg.K, Pr = 0.697 The Reynolds number for flow through a circular
Then friction coefficient pipe is calculated as
2 × 7.75 × (0.697) 2 / 3 ρ umD umD
Cf = = 3.025 × 10–3 ReD = =
0.998 × 1009 × 4 µ ν
The average shear stress where um = mean fluid velocity, D is inner diameter of
Cf 3.025 × 10−3 tube and ν is kinematic viscosity.
τ= ρu∞ 2 = × 0.998
2 2 The Reynolds number for non-circular duct is
× (4)2 = 0.0241 N/m2. calculated as
The drage (shear) force u D
Re = m h
F = wLτ = 1 × 1 × 0.0241 ν
= 0.0241 N. Ans. 4 Ac
where Dh = , hydraulic diameter,
P

7.15. SUMMARY Ac = cross-section area of non-circular tube,

P = wetted perimeter.
Convection is the mode of heat transfer that involves The Reynolds number at which the flow turns to
conduction as well as bulk fluid motion. The rate of be turbulent from laminar flow is called critical Reynolds
convection heat transfer is expressed by Newton’s law number, Recr and its value is
of cooling as Recr = 5 × 105 for flow over flat plate
Q = hA(Ts – T∞) (W) = 2300 for flow inside tubes.
where Ts is surface temperature and T∞ is free stream For flow over a flat plate, the momentum and
fluid temperature. energy equations are given as
The convection heat transfer is classified as
∂u ∂u 1 ∂ 2 u ∂T ∂T 1 ∂ 2 Τ
natural or forced convection. The natural or free u +v = and u +v =
∂y ∂y ν ∂y 2 ∂x ∂y α ∂y 2
convection is a process in which fluid motion is set up
due to density difference results from heat transfer. The similarities between velocity and thermal
While in the forced convection, the fluid is forced to flow boundary layer indicate that the local skin friction
over a surface or in a duct by external means. coefficient Cfx and Nusselt number Nux are function of
Reynolds number as
The region of flow in which the effects of viscous
Cfx = f(x∗, Rex)
shear forces caused by fluid viscosity are observed, is
called velocity boundary layer (δ). The flow region over hx L
Nux = = φ(x∗, Rex, Pr)
the surface in which the temperature variation in the kf
 262 ENGINEERING HEAT AND MASS TRANSFER

x µCp The Reynolds Colburn anology for turbulent flow

where x∗ = and Pr = , Prandtl number. over a flat plate indicates that the heat transfer
L kf
coefficient and fluid friction are related as
The local friction coefficient is expressed in terms
of local shear stress τs as hx C
Pr 2 / 3 = fx
ρ u∞ C p 2
τs
Cfx = Here the quantity
ρ u∞ 2 /2
hx Nu
The dimensional analysis is a method of analysis =
ρ u∞ C p Rex Pr
in which certain variable affecting a phenomenon are
= Stx (Stanton number)
combined in dimensionless group such as Nusselt
number, which facilates the interpretation and extends C fx
∴ Stx Pr2/3 =
its application to experimental data. 2

TABLE 7.5. Dimensionless groups used in heat transfer

Groups Definition Interpretation

hL c
Biot number (Bi) Ratio of internal thermal resistance of a solid to the boundary
k
layer thermal resistance.
τs
Coefficient of friction (Cf) Dimensionless surface shear stress.
ρu∞2 /2
Colburn j factor St Pr2/3 Dimensionless heat transfer coefficient.
αt
Fourier number (Fo) Ratio of heat conduction to the rate of thermal energy
L c2
storage in a solid.
∆p
Friction factor (f ) Dimensionless pressure drop for internal flow.
( L/D) ρu∞2 /2

gβ (Ts − T∞ )L c3
Grashof number (GrL ) Ratio of buoyancy to viscous forces of the fluid.
ν2

C p (Ts − Tsat )
Jacob number (Ja) hfg
Ratio of sensible heat to latent energy absorbed during liquid

vapour phase change.

hL
Nusselt number (NuL ) Dimensionless temperature gradient at the surface of fluid.
kf

Peclet number (PeL ) ReL Pr Dimensionless independent heat transfer parameter.

µC p ν
Prandtl number (Pr) = Ratio of momentum and thermal diffusivities.
kf α

u∞L
Reynolds number (ReL ) Ratio of inertia to viscous forces of a flowing fluid.
ν

h Nu L
Stanton number (St) = Modified Nusselt number.
ρu∞C p ReL Pr

where Lc = characteristic length of the geometry.

PRINCIPLES OF CONVECTION 263

REVIEW QUESTIONS 23. Express the similarities of momentum and energy

equations for flow over a flat plate.
1. Define laminar and turbulent flows. What is Reynolds 24. State the method for evaluation of heat transfer
number ? coefficient.
2. Explain velocity and thermal boundary layer. 25. State the scope and application of dimensional
3. Discuss laminar sublayer, buffer layer and turbulent analysis in heat transfer processes. What are the two
layer in a boundary layer. methods for obtaining the dimensionless groups ?
4. What is critical Reynolds number? State its 26. Show by Rayleigh method of dimensional analysis,
approximate values for flow over flat plate and that the Nusselt number is function of Reynolds
through a circular tube. number and Prandtl number.
5. What do you understand by local and average value 27. Explain Buckingham π theorem. What are its merits
of heat transfer coefficient ? and demerits ? What are repeating variables ? How
6. Explain local and average value of skin friction are they selected ?
coefficient. 28. What do you understand by mean value and
7. Show that the Reynolds number for flow through a fluctuating component of velocity and other properties
tube of diameter D can be expressed in turbulent flow ?
 29. Explain the Prandtl mixing length concept to describe
4m
Re = . turbulent flow over a surface.
π Dµ
30. Explain the Reynolds analogy for turbulent flow over
8. Explain the mechanism of convection heat transfer. a surface.
9. What are the differences between natural and forced
convection ? PROBLEMS
10. What is external forced convection ? How does it differ
from internal forced convection ? 1. Calculate Reynolds numbers and state the type of flow,
11. What is physical significance of Prandtl number ? whether it is laminar or turbulent for the following :
12. What is physical significance of Reynolds number ? (a) A 15 m long yatch sailing at 15 km/h in sea water
How is it defined for (a) flow over a flat plate of length (ρ = 1000 kg/m3 and µ = 1.3 × 10–3 kg/ms).
L, (b) flow over a cylinder of diameter D, (c) flow [Ans. 48.07 × 106, turbulent]
through a tube of diameter d, and flow through a (b) A compressor disc of radius 0.5 m rotating at
rectangular tube of cross-section a × b ? 18000 r.p.m in air at 5 bar and 400°C and
13. What is physical significance of Nusselt number ?
How is it defined for (a) flow over a flat plate of 1.46 × 10 − 6 T3/2
µ= kg/ms.
length L, (b) flow over a cylinder of diameter D, (c) flow (110 + T)
through a tube of diameter d, and flow through a [Ans.14.78 × 107, turbulent]
rectangular duct of cross-section a × b ? (c) 0.08 kg/s of CO2, gas at 400 K flowing in a 40 mm
14. When is heat transfer through a fluid layer by diameter pipe. For viscosity take
conduction and when is it by convection ? For what
case, the rate of heat transfer is higher ? 1.56 × 10− 6 T3/2
µ= kg/ms.
15. How does the heat transfer coefficient differ from (233 + T)
thermal conductivity ? [Ans. 1.29 × 105, turbulent]
16. What is no slip condition on a surface ? (d) The roof of a coach 6 m long, travelling at 100 km/h
17. What property is responsible for development of in air (ρ = 1.2 kg/m3 and µ = 1.8 × 10–5 kg/ms).
velocity boundary layer ? What property is for thermal [Ans. 1.11 × 107, turbulent]
boundary layer ? F Pr = µC I
18. Consider laminar flow over a flat plate, will the
friction coefficient change with position ? How about
2. Calculate Prandtl number GH k JK f
p
for the

the heat transfer coefficient ? Explain. following :

19. In the fully developed region of the flow in a circular (a) Water at 20°C : µ = 1.002 × 10 –3 kg/ms,
tube, will the velocity profile change in the flow Cp = 4.183 kJ/kg.K and kf = 0.603 W/m.K.
direction ? [Ans. 6.95]
20. How does surface roughness affect pressure drop and (b) Air at 20°C and 1 bar : R = 287 J/kg.K,
heat transfer in a tube flow ? ν = 1.563 × 10–5 m2/s, Cp = 1005 J/kg.K and
21. Derive an expression for momentum transfer kf = 0.02624 W/m.K. [Ans. 0.719]
equation for flow over a flat plate. (c) Mercury at 20°C ; µ = 1520 × 10 –6 kg/ms,
22. Derive an equation for energy transfer for flow over Cp = 0.139 kJ/kg.K and kf = 0.0081 W/m.K.
a flat plate.
[Ans. 0.0261]
264 ENGINEERING HEAT AND MASS TRANSFER

(d) Engine oil at 60°C ; µ = 8.36 × 10 –2 kg/ms, (b) Develop an expression for average friction
Cp = 2035 J/kg.K and k = 0.141 W/m.K. coefficient over a distance x = L form the leading
[Ans. 1207] edge of the plate.
3. Calculate the appropriate Grashof number and state (c) Calculate the drag force acting on a plate 2 m by
the type of flow for the following : 2 m for the flow of air at atmospheric pressure
and at 350 K with velocity of 4 m/s.
(a) A central heating radiator, 0.8 m high with a
surface temperature of 75°C in a room at 18°C 8. The exact expression for local friction coefficient Cfx
(ν = 1.5 × 10– 5 m2/s, Pr = 0.72) for laminar flow over a flat plate is given by
[Ans. 3.98 × 109, Turbulent] 0.664
Cfx = .
(b) A horizontal oil sump with a surface temperature Re x
of 40°C, 0.5 m long and 0.4 m wide containing oil
at 75°C, (Pr = 546, β = 0.7 × 10–3 K–1 and ν = 4.168 Air at atmospheric pressure and 350 K flows with a
× 10–5 m2/s) [Ans. 18.97 × 104, Laminar] velocity of 30 m/s over a flat plate 0.2 m long. Calculate
the drag force acting per meter width of the plate.
(c) The surface of heating coil 30 mm diameter, having
surface temperature of 80°C in water at 20°C [Ans. 0.222 N]
(ρ = 1000 kg/m3, Pr = 6.95, β = 0.227 × 10–3 K–1 9. The local heat transfer coefficient hx for laminar
and µ = 1.00 × 10–3 kg/ms). boundary layer flow over a flat plate is given by
[Ans. 3.6 × 106, laminar] xhx
(d) Air at 20°C , (Pr = 0.72, and ν = 1.5 × 10–5 m2/s) = 0.332 Rex1/2 Pr1/3.
kf
adjacent to a 75 mm diameter horizontal light bulb
with a surface temperature of 100°C. Develop an expression for average heat transfer
[Ans. 4.41 × 104, laminar] coefficient h over a distance x = L from leading edge
4. Calculate appropriate Nusselt number for the of the plate.
following : 10. Engine oil at 40°C (µ = 0.21 kg/(m/s) ; ρ = 875 kg/m3)
A gas flow (Pr = 0.71, µ = 4.63 × 10–5 kg/ms and flows inside a 2.5 cm diameter, 50 m long tube with a
Cp = 1175 J/kg.K) over a turbine blade of chord length mean velocity of 1 m/s. Determine the pressure drop
20 mm, where the average heat transfer coefficient for flow through the tube. (J.N.T.U., May 2004)
is 1000 W/m2.K. [Ans. 261]
LM Hint. ∆p = f L ρu 2
64 OP
N
∞
,f = [Ans. 537.6 kPa]
5. Calculate the distance from the leading edge of a flat
plate at which the transition occurs from laminar to
D 2 Re PQ
turbulent flow for atmospheric air at 27°C with (a) 2,
(b) 10, (c) 20 m/s. Assume transition at Recr = 5 × 105. 11. For a laminar natural convection from a heated
vertical surface, the local convection coefficient may
[Ans. (a) 4.21 m, (b) 0.842 m, (c) 0.421 m] be expressed as hx = C x–1/4.
6. Assume transition from laminar to turbulent at
where hx is heat transfer coefficient at a distance x
Recr = 5 × 105, calculate the distance from the leading
from leading edge and C is a constant.
edge at which the transition occurs for the flow of
each of the following fluids with a velocity of 2 m/s at Derive an expression for the ratio h/hx, where h is
40°C (a) air at atmospheric pressure, (b) hydrogen at average heat transfer coefficient between leading edge
atmospheric pressure, (c) water, (d) ethylene glycol,
LM Ans. 4  L  −0.25 
(e) engine oil.
7. The velocity profile u(x, y) for a laminar boundary
(x = 0) and x = L location.
N 3 x 

layer flow along a flat plate is given by

u ( x, y)
=2
y
−2
LM OP + LM y OP
y
3 4
REFERENCES AND SUGGESTED READING
u∞ δ( x) N Q N δ(x) Q
δ ( x)
where the boundary layer thickness δ(x) is given by 1. Rehsenow W. M, J. P. Harnett and E.N. Ganic, Eds
‘‘Handbook of Heat Transfer’’, 2/e, McGraw Hill, New
δ( x) 5.83 York 1985.
=
x Re x 2. Kays W.M. and M.E. Crawford, ‘‘Convective Heat and
(a) Develop an expression for local friction Mass Transfer’’, 2nd ed., McGraw Hill, New York,
coefficient. 1980.
PRINCIPLES OF CONVECTION 265

3. Giedt Warren H., ‘‘Investigation of Variation of Point 11. Incropera F. P. and D. P. DeWitt, “Introduction to
Unit-Heat Transfer Coefficient Around a Cylinder Heat Transfer”, 2nd ed., John Wiley & Sons, 1990.
Normal to an Airstream’’. Transaction of ASME, 12. Bayazitoglu Y. and M. N. Ozisik, “Elements of Heat
vol. 71 pp. 375–301, 1949. Transfer”, McGraw Hill, New York, 1988.
4. Schlichting H., “Boundary Layer Theory”, 6th ed., 13. Thomas L.C., “Heat Transfer”, Prentice-Hall,
McGraw Hill, New York, 1968. Englewood Cliffs, NJ, 1982.
5. Zhukauskas A and A, B, Ambrazyavichyus, Int, J. 14. White F.M., “Heat and Mass Transfer”, Addison
of “Heat Mass Transfer”, Vol 3, pp. 305, 1961. Wesley, Reading, MA, 1988.
6. Knudsen J.D. and D.L. Katz, ‘‘Fluid Dynamics and 15. Jacob M., “Heat Transfer”, Vol I, Wiley, New York,
Heat Transfer’’, McGraw Hill, New York, 1958. 1949.
7. McAdams W.M., ‘‘Heat Transmission’’, 3rd ed. 16. Suryanarayana N. V., ‘‘Engineering Heat Transfer’’
McGraw Hill, New York, 1954. West Pub. Co. New York, 1998.
8. Jacob M. and G.A. Hawkins, “Elements of Heat 17. Chapman Alan. J., ‘‘Fundamentals of Heat Transfer’’
Transfer”, 3rd ed., Wiley, New York, 1957. Macmillan, New York.
9. Krieth Frank and M.S. Bohn, “Principles of Heat 18. Christopher Long, ‘‘Essential Heat Transfer’’,
Transfer”, 5th ed., PWS Pub. Company, 1997. Addision Wesley Longman, 2001.
10. Holman J. P., “Heat Transfer”, 7th ed., McGraw Hill, 19. Giedt Warren H., ‘‘Principles of Engineering Heat
New York, 1990. Transfer’’, Van Nostrand Inc. 2nd ed., 1967.