CBE 3322: Heat Transfer Operations

Section 3a – Convection (Introduction)

Dominic Pjontek, Ph.D., P.Eng.

University of Western Ontario
Boundary layers

◼ Boundary layers are an important concept to understand

convective heat transfer. Let’s consider the velocity boundary layer
for flow over a flat plate:

◼ We know from fluid mechanics that the fluid in contact with the
surface can be assumed to have zero velocity (u = 0 m/s).
❑ This is generally referred as the no slip condition.

Convection (Introduction) 2
Boundary layers

◼ The fluid at the surface (u = 0 m/s) slows down the adjacent fluid
layer due to shear stresses, which then slows down the next layer of
fluid, etc… resulting in a velocity profile.
◼ Eventually, the effect becomes negligible past a given distance in
the y-direction. This distance in the y-direction is referred as the
velocity boundary layer thickness (δ) and is based on u = 0.99 u∞.
◼ In fluid flow, we are typically interested in the local friction
coefficient (Cf) as well as the surface shear stress (τs):

 u
C f = 2S S =  where: μ is the dynamic
u 2 y y =0
viscosity (Pa·s)

Convection (Introduction) 3
Thermal boundary layer

◼ These concepts can also be applied to define the thermal

boundary layer. Let’s consider flow over an isothermal flat plate:

◼ We are now focusing on the temperatures as a function of the y and

x locations. In this case, we will define the thermal boundary layer
thickness (δt) as:
Ts − T
t  = 0.99
Ts − T

Convection (Introduction) 4
Thermal boundary layer

◼ Based on the previous figure, as well as our previous knowledge

from this course, we can determine the local heat flux by applying
Fourier’s law to the fluid at the surface of the isothermal flat plate:

" T
q = −k f
s
y y =0

◼ The above can be assumed because of the no slip condition at the

surface, where only conduction occurs with no velocity. By applying
an energy balance on the surface, we obtain the following:

− k f  T y y =0
h=
Ts − T

Convection (Introduction) 5
Local versus average convection

◼ Let’s examine the local convection coefficient various objects or

shapes, shown below:

◼ We have previously defined the local heat flux (q”) as:

q" = h(Ts − T )

◼ Where h (W/m2·K) is the local convection coefficient, TS (K) is the

local surface temperature, and T∞ (K) is the bulk fluid temperature.

Convection (Introduction) 6
Local versus average convection

◼ Since flow conditions will vary as a function of the location on a

studied surface or shape, the local heat flux and thus local
convection coefficient will also vary.

◼ We can thus determine the total heat transfer rate (q) by

integrating the local heat flux over the entire surface:

q=  q" dA s = (Ts − T ) h dA s
As As

◼ It is thus useful for us to define an average convection

coefficient ( h ) for the entire surface area.

Convection (Introduction) 7
Local versus average convection

◼ We can now determine the total heat transfer rate as follows:

q = h A s (Ts − T )

◼ Where the average and local convection coefficients are

related by the following general expression:

1
h=
As As
h dA s

◼ When considering flow over a flat plate, this simplifies as follows:

1 L
h =  h dx
L 0

Convection (Introduction) 8
Local versus average convection (Example)

◼ Air at a free steam temperature of T∞ = 20°C is in parallel flow over

a flat plate of length L = 5 m and temperature Ts = 90°C. However,
obstacles placed in the flow intensify mixing with increasing distance
x from the leading edge, and the spatial variation of temperatures
measured in the boundary layer is correlated by the following:
❑ T(x,y) = 20 + 70 exp (-600xy)
◼ where x and y are in meters and T is in °C.

◼ Let’s determine and plot the manner in which the local convection
coefficient h varies with x.
◼ What is the average convection coefficient for the plate?

Convection (Introduction) 9
Flow regimes

◼ The first step when studying convection problems is to determine

whether the boundary layer is under laminar and/or turbulent
flow conditions. The image below demonstrates the sharp
differences between both flow conditions.

Convection (Introduction) 10
Flow regimes

◼ For laminar flow conditions:

❑ Fluid motion in the boundary layer is highly ordered, which
allows for the identification of streamlines for the fluid.
❑ The streamlines are characterized by velocity components in
both the x (u velocity) and y (v velocity) directions.

◼ For turbulent flow conditions:

❑ Fluid motion in the boundary layer is irregular and undergoes
velocity fluctuations.
❑ This enhances the energy transfer by increasing the convection
transfer rates.

Convection (Introduction) 11
Visualizing flow regimes (Schlieren Optics)

Convection (Introduction) https://www.youtube.com/watch?v=mLp_rSBzteI&t=3s 12

Visualizing flow regimes (Schlieren Optics)

Convection (Introduction) https://www.youtube.com/watch?v=4tgOyU34D44 13

Flow regime transition
◼ When considering a flat plate, it can be demonstrated that the flow
will initially be in the laminar flow regime, but will eventually
transition to turbulent flow after a critical distance.

❑ The change in characteristics between laminar and turbulent

flow is referred as the transition flow regime.

◼ The image to the right demonstrates

the change in the boundary layer
thickness (δ) and the convection
coefficient (h) when transition from
laminar to turbulent flow

Convection (Introduction) 14
Reynolds number

◼ When determining the boundary layer behavior, the Reynolds

number (Rex), a dimensionless group, is used:

uL inertial forces

ReL = ≡
 viscous forces

◼ Where: ρ is the fluid density (kg/m3), u∞ is the bulk fluid velocity

(m/s), L is the characteristic length of the geometry (m), and μ is
the fluid viscosity (Pa·s)
◼ For a smooth flat plate, the transition from laminar to turbulent flow
begins at the critical distance (xc), where:

u x c
Re x ,c = = 5  105


Convection (Introduction) 15
Reynolds number

◼ The transition region between laminar and turbulent flow generally

occurs between 5·105 ≤ ReL < 3·106 for a smooth flat plate.

◼ We can interpret the Reynolds number as follows:

❑ Inertial forces dominate for large Reynolds number
❑ Viscous forces dominate when the Reynolds number is small

◼ The characteristic length (L) required to calculate the Reynolds

number is selected based on the geometry of interest:
❑ Flat plate: length of the plate
❑ Cylinder: pipe diameter

Convection (Introduction) 16
Flow regime (Example)

◼ A fan that can provide air speeds up to 50 m/s is to be used in a

low-speed wind tunnel with atmospheric air at 25°C.

❑ If we wish to use the wind tunnel to study flat-plate boundary

layer behavior up to Reynolds numbers of Rex = 108, what is
the minimum plate length that should be used?

❑ At what distance from the leading edge would transition occur

if the critical Reynolds number is Rex = 5·105?

Convection (Introduction) 17
Prandtl number

◼ When solving for a set of differential equations to govern the

temperature fields in the boundary layer (which is outside the scope
of this class), another relevant dimensionless group appears:

❑ Prandtl number

   C p momentum diffusivity
Pr = = = ≡
 k f Cp kf thermal diffusivity

◼ Where: ν is the kinematic viscosity (m2/s), α is the thermal

diffusivity (m2/s), kf is the thermal conductivity of the fluid (W/m·K),
and Cp is the specific heat (J/kg·K).

Convection (Introduction) 18
Prandtl number

◼ The Prandtl number relates the velocity and thermal boundary

layers. As result, both boundary layers overlap when Pr = 1.

◼ We can interpret the Prandtl number as follows:

❑ When Pr is large, the thickness of the thermal boundary layer
(δt) is smaller than the velocity boundary layer.

◼ Some typical values are provided below:

Material Prandtl number
Liquid metals 0.004 – 0.03
Gases 0.7 – 1.0
Water 1.0 – 10
Oils 50 – 100 000

Convection (Introduction) 19
Nusselt number

◼ Recalling from an earlier slide, we had the following expression to

determine the convection coefficient based on the fluid temperature
profile at the surface (i.e., y = 0):
kf T
h=− 
Ts − T y y =0

◼ For this type of analysis, it is common to use dimensionless

parameters in the expressions. We will use the following for the
convective rate of heat transfer:

T − Ts
* * y
T = y =
T − Ts L
where L is the characteristic length

Convection (Introduction) 20
Nusselt number

◼ When we incorporate the previous dimensionless parameters, we

obtain the following:

k f T − Ts T * k f T *
h=−  * =
L Ts − T y y * =0
L y * y * =0

◼ From here, we define a dependent dimensionless parameter that is

referred as the Nusselt number:

h L T * Convective heat transfer

Nu =  * ≡
kf y y * =0
Conductive heat transfer

Convection (Introduction) 21
Nusselt number

◼ For a given geometry, the local Nusselt number can be described

using a general correlation of the form:

(
Nu = f x * , ReL , Pr )
◼ Where x* = x/L is the dimensionless length in the x-direction relative
to the characteristic length (L).
◼ Since we are generally interested in the average heat transfer
coefficient (which can be determined by integrating over the entire
surface area of the studied geometry), the functional dependence of
the average Nusselt number will be:
hL
Nu = = f (ReL , Pr )
kf

Convection (Introduction) 22
Nusselt number (Online Example)

◼ Forced air at T∞ = 25°C and V = 10 m/s is used to cool electronic

elements on a circuit board. One such element is a chip, 4 mm by 4
mm, located 120 mm from the leading edge of the board.
◼ Experiments have revealed that flow over the board is disturbed by
the elements and that convection heat transfer is correlated by the
following expression:
❑ Nux = 0.04 Rex0.85 Pr1/3

◼ What is the surface temperature of the chip if it is dissipating

30 mW?

Convection (Introduction) 23