A Short Course on Heat Transfer

Intended as a repetition from previous courses

by
Björn Palm,
Royal Institute of Technology, Stockholm
 Course Contents,
based on Holman’s book Heat Transfer

Chapter 1: Introduction
Chapter 2: Steady-State Conduction - One Dimension
Chapter 3: Steady-State Conduction - Multiple Dimensions
Chapter 4: Unsteady-State Conduction
Chapter 5: Principles of Convection
Chapter 6: Empirical and Practical Relations for Forced-Convection
Heat Transfer
Chapter 7: Natural Convection Systems
Chapter 8: Radiation Heat Transfer
Chapter 9: Condensation and Boiling Heat Transfer
Chapter 10: Heat Exchangers
(Chapter 11: Mass Transfer)

2
 Part 1

Introduction

3
Introduction

What is heat?
Heat is energy transfer caused by temperature difference!

4
 The four laws of thermodynamics:
• Zeroth law:
If two bodies both are in thermal equilibrium with a third
body, they are also in thermal equilibrium with each other,
and they then are said to have the same temperature.

5
• First law: (Energy principle)
Energy cannot be generated or destroyed, only converted to
different forms.

(”Energy consumption” is the transfer of ”prime” energy to

thermal energy in the surrounding)

6
• Second law:
Heat cannot by itself pass from one body to another body
with higher temperature.

(Entropy [disorder] strives to a maximum in a closed system.

Shows the direction of time.)

7
•
8
 Third law:
The entropy of a pure, crystalline material takes its lowest
value at absolute zero temperature, where it is 0.

(There is a lowest limit to the temperature, 0K = -273.15°C.

World record: 0.00000017 K = 170 nanokelvin)

9
Three modes of heat transfer:

• Conduction
Through solid bodies and ”still” fluids hot cold

• Convection
Through moving fluids
(also boiling and condensation)

• Radiation
Between surfaces, through gas or vacuum
hot cold

10
What do we know about conduction?

11
Conduction, thermal conductivity, Fourier´s law:

Fourier’s law:
q = -k⋅A⋅δT/δx
where, q = heat flow (W)
A = area perpendicular to heat flow (m2)
δT/δx = temperature gradient in the direction
of heat flow(°C/m)
k = thermal conductivity (W/(m °C))

Fourier´s law is the defining equation for the thermal

conductivity.

12
Conduction, thermal conductivity, Fourier´s law:

For one-dimensional heat transfer (a plane wall,) with

constant thermal conductivity, Fourier´s law is simplified to

q = k⋅A⋅ΔT/δ
where ΔT = temperature difference (°C)
δ = distance or thickness (m).

13
fig. 1
Temp

δT/δx<0

x
Hot q>0 Cold

14
Fig. 2 Heat transfer through a plane wall

T1
ΔT

T2

15
 Table 1
Material Thermal conductivity (20°), (W/m⋅°C)
Diamond, type IIa 2600
Copper 386
Iron, wrought, 0.5% C 60
Stainless steel, 18/8 16.3
Brick 0.69
Water 0.6
Pine wood, 0.15/0.33 (cross/along fibres)
Cork 0.045
Glass wool 0.038
Mineral wool 0.04
Polyurethane 0.02
Air 0.026
Argon gas 0.018

16
Example to solve:

Conduction:
Calculate the heat flow per square meter (heat flux) through a
mineral wool insulation, 5 cm thick, if the temperatures on the
two surfaces are 30 and 200°C, respectively.

17
Heat transfer by convection, Newton’s law of cooling

Convection is a general term for heat transfer through a

moving fluid.

18
 Fig 3, Different types of convection heat transfer

Forced Boiling
convection

Diffusion,
mass
transfer
Condensation
19
What do we know about convection?

20
All types of convection are governed by Newton’s law of
cooling:

q = h⋅A⋅ΔT
where A = surface area where convection takes place
2
(m )
ΔT = temperature difference (°C)
h = heat transfer coefficient (W/(m2⋅°C))

Newton’s law of cooling is the defining equation for the heat

transfer coefficient h.

21
 Table 2

Type of flow Approximate heat transfer

coefficient (W/(m2⋅°C))
Turbulent flow in tubes,
(diameter ≈ 50 - 25 mm)
Water (0.5 - 5 m/s) 1500 - 20000
Air (1 - 10 m/s) 10 - 50
Laminar flow in tubes, (diameter
≈ 50 - 10 mm)
Water 50 - 250
Air 2 - 15
Air flow past plates (1 - 10 m/s) 10 - 50

22
Type of flow Approximate heat transfer
coefficient (W/(m2⋅°C))
Natural convection
Water 200 - 1000
Air 2 - 10
Condensation
Water 5000 - 15000
Refrigerants 1000 - 5000
Boiling
Water 1000 - 40000
Refrigerants 200 - 5000

23
Example to solve:

Convection:
What is the approximate temperature difference between a hot
plate and the surrounding air if the heat flux from the plate is
800 W/m2 ? Assume that the air is flowing past the surface
with a velocity of 5 m/s giving a heat transfer coefficient of
20 W/(m2K).

24
Radiation

• All bodies send out energy in the form of electromagnetic

radiation.

• The wavelength and intensity is dependent on the

temperature of the surface.

• Radiation may be transferred through vacuum, but also

through air.

25
• fig. 4, Heat transfer by radiation
•

Δ T r = T1 - T 2
T2
T3

T1

26
Heat transfer by radiation between a small body and an
isothermal environment may be calculated by Newton’s law
of cooling, if we define a radiation heat transfer coefficient:

qr = hr⋅A⋅ΔTr
where qr = heat flow due to radiation (W)
hr = radiation heat transfer coefficient
(W/(m2⋅°C))
A = surface area of the small body (m2)
ΔTr = temp. difference between surfaces (°C).

27
• hr is a function of the geometry, the emissivity of the
surfaces and on the temperatures of the surfaces.

• For radiation heat transfer between non-polished surfaces

with temperatures between 0 and 100°C, the radiation heat
transfer coefficient is usually between 4.5 and 12 W/(m2 °C)
(≈5 at room temperature, ≈25 at 200°C)

28
Summing convection and radiation modes of heat transfer

• If the temperature differences for convection and radiation

are equal then the h-values may be added to a total heat
transfer coefficient.

hc + hr = h tot

• If the temperature differences are not the same, the two

modes have to be treated separately.

29
Example to solve:

Convection + radiation:
A cold bottle of beer (+5°C) is placed in a room where the
temperature of the air and of the walls is 25°C. Calculate the
approximate heat flux caused by radiation and by natural
convection

30
Overall heat transfer coefficient

We define the overall heat transfer coefficient by the equation

q = U⋅A⋅ΔTtot
where U = the overall heat transfer coefficient (W/(m ⋅°C))
2
2
A = surface area on either side of the wall (m )
ΔTtot = difference between the fluid temperatures
sufficiently far from the wall.

31
Relation between U, h and k

• U can be related to the h1, h2 in the fluids and to k and δ of

the wall through the temperature differences:

• ΔT on either side of a wall can be written, according to

Newton’s law of cooling as, respectively:
ΔT1 = q/(h1⋅A1)
ΔT2 = q/(h2⋅A2)

• ΔT in the wall is, according to Fourier´s law:

ΔTW = q⋅δ/(k⋅AW)

32
• The total temperature difference is
ΔTTot = q/(U⋅A)

• But
ΔTtot = ΔT1 + ΔT2 + ΔTW

• q (at steady state) must be equal at the inside, the outside

and in the wall, thus
• 1/(U⋅A)= 1/(h1⋅A1)+ δ/(k⋅AW)+ 1/(h2⋅A2)

33
• For the case of a plane wall, the areas are also equal, and the
relation is even simpler.

• If the wall has two or more layers of different materials,

additional terms of [δ/(k⋅AW)] have to be added.

• For a curved surface, the surfaces are not the same. The
overall heat transfer coefficient may be referred to any of the
surfaces, but when specifying the U-value, it must always be
stated to which area it is connected.

34
 Fig. 5

ΔT1

q = U⋅A⋅ΔTtot q
ΔTtot ΔTw
T
k

δ ΔT2

35
Heat transfer resistance

• Thermal resistances, are defined analogous to electric

resistances.

• The thermal correspondence to Ohms law shows how the

thermal resistance has to be defined:

Ohms law: Voltage = Current ⋅ Resistance

Thermal analogy gives: ΔT = q ⋅ Rth

36
• Thermal resistance in the fluid
R = 1/(h⋅A1)

• Thermal resistance in the wall

R = δ/(k⋅A)

• Total thermal resistance

Rtot =1/(U⋅A)

• The total thermal resistance is thus the sum of the resistances

Rtot = R1 + RW + R2

37
Example to solve:

Overall heat transfer coefficient:

The heat flux through a plane wall is 5000 W/m2 and the
overall heat transfer coefficient is 1000 W/(m2⋅°C). Calculate
the temperature difference!

38
 Part 2
More about convection

39
Definition of dimensionless parameters

• A large number of parameters are needed to describe heat

transfer.

• Parameters may be grouped together to form a small number

of dimensionless similarity parameters.

These give simpler and more general equations by which heat

transfer coefficients may be calculated.

40
Reynolds number

The Reynolds number is defined as

Re = u⋅x/ν
where u = velocity of fluid (m/s)
x = characteristic length (m). (For a tube, x = d).
ν = kinematic viscosity of fluid (m /s)
2

• Reynolds number determines whether the fluid flow is

laminar or turbulent.
• Reynolds number determines the ratio of the inertia and
viscous forces in the flow.

41
Reynolds number

Glasrör

Bläckstråle

Laminärt strömningssätt

Turbulent strömningssätt

42
Nusselt number

The Nusselt number is defined as

Nu = h⋅x/k
where h = surface heat transfer coefficient (W/(m ⋅°C))
2

x = characteristic length (m).

k = thermal conductivity (W/(m⋅°C))

• The Nusselt number: a dimensionless temperature gradient

at the surface.
• From Nu, the heat transfer coefficient can easily be
calculated.

43
Example to solve:

Water (k = 0.608 W/(m⋅K)) is flowing through a tube with

inner diameter 15 mm. The Nusselt number is found to be 70.
What is the heat transfer coefficient?

44
Prandtl number

The definition of the Prandtl number is

Pr = ν/α = cp⋅μ/k
where ν = kinematic viscosity of fluid (m /s)
2

α = thermal diffusivity (m2/s)

cp = specific heat (J/kg⋅°C)
μ = dynamic viscosity (N⋅s/m2)
k = thermal conductivity (W/(m⋅°C))

• The Prandtl number is a thermodynamic property of the

fluid.
45
Grashof number

The Grashof number is defined as

Gr = g⋅β⋅ΔT⋅x3/ν2
where g = acceleration of gravity (m/s2)
β = volumetric thermal expansion coeff. (1/°C)
For (ideal) gases, (β = 1/Tabs).
ΔT = temperature difference between surface and
fluid (°C)
x = characteristic length (m).
ν = kinematic viscosity of fluid (m2/s)
• The Grashof number indicates, in free convection, whether
the flow is laminar or turbulent.
• It is the ratio between buoyancy and viscous forces.
46
Graetz number

The Graetz number is defined as

Gz = Re ⋅ Pr ⋅ d/x
where d = (hydraulic) diameter of channel (m)
x = distance from entrance of channel (m)

• Gz is used when calculating heat transfer in laminar tube

flow.
Can be interpreted as a dimensionless, inverted length
(distance from entrance).

47
Rayleigh number

The Rayleigh number is defined as

Ra = Gr ⋅ Pr

The Rayleigh number often appears in equations for free

convection.

48
Laminar and turbulent flow, velocity boundary layer

Fig. 6

49
• The part of the flow where the velocity is influenced by the
surface is called the boundary layer.

• Near the front edge of the plate the thickness of the

boundary layer is thin and it then grows successively thicker.

• As long as the boundary layer is thin, there is no mixing

between layers at different distances from the plate. The
flow is then said to be laminar.

• At some distance from the leading edge, the laminar layer

will become unstable, and eddies will develop, mixing the
different layers. The flow is then becoming turbulent.

50
• Because of the mixing, the difference in velocity between
layers is much smaller in turbulent flow than in laminar, and
the velocity profile thus much flatter.

• Even in the turbulent region there is a laminar sub-layer

closest to the surface. In this sub-layer, the temperature
profile is nearly linear.

• The type of flow, laminar or turbulent, can be determined

from the Reynolds number calculated with the distance from
the leading edge as the characteristic length. Transition will
occur at Re ≈ 5⋅105.

51
Fig. 7

52
• In tube flow, at some distance from the entrance, the
boundary layers from opposite sides will meet. At this point
the flow is fully developed.

• The fully developed flow may be turbulent or laminar.

• The Reynolds number is calculated using the tube diameter

as the characteristic length. Transition takes place at
approximately Re ≈ 2300.

• The heat transfer coefficients are generally higher in

turbulent than in laminar flow.

53
Example to solve:

Water (ν = 0.86⋅10-6 m2/s) flows through a tube with the

diameter 12 mm at a velocity of 2 m/s. Determine if the flow
is laminar or turbulent!

54
Thermal boundary layer

• At heated (or cooled) surfaces, a thermal boundary layer

will form in which the temperature change from the wall
temperature to the temperature of the undisturbed fluid.
• In laminar flow the velocity- and the thermal boundary
layers will be similar, but the thicknesses will not
necessarily be the same.
• The relative thickness of the thermal and velocity boundary
layers is in laminar flow related to the Prandtl number by
δv /δth ≈ Prn where n is a positive exponent (≈1/3).

55
• For gases, the Prandtl number is usually between 0.7 and 1,
and in laminar flow the thermal and velocity boundary layer
thicknesses are thus approximately equal.

• For liquid metals, the Pr<<1 and the thermal boundary layer
(in laminar flow) is considerably thicker than the velocity
boundary layer, while for oils, Pr>>1 and the velocity
boundary layer is the thickest. (For water, Pr range from
13.4 at 0°C to 1.75 at 100°C).

• For turbulent boundary layers, the mixing within the layer

will result in more or less equal thicknesses of the velocity
and turbulent layers.

56
Forced and free convection

• Fluid flow may be caused by

>force (by a fan or a pump or any other means external to
the fluid itself).
>free convection (the movement is caused by temperature
induced density differences in the fluid).

• In forced convection, the type of flow (turbulent or laminar)

is determined from the Reynolds number

• In free convection, the type of flow is determined by the

Grashof number

57
• The dimensionless equations are different for the two cases:

>In forced convection: Nu = f(Re, Pr)

>In free convection: Nu = f(Gr, Pr)

58
 Methods for calculating heat transfer
coefficients (in one phase flow)
Forced convection

59
Turbulent flow in tubes and channels

The heat transfer coefficient is generally referred to the

logarithmic mean temperature difference (LMTD) defined as:

ϑln = [(Th2 - Tc2 ) -(Th1 - Tc1 )] / ln[(Th2 - Tc2 ) / (Th1 - Tc1 )]

ΔT2 − ΔT1
ΔT2
= ln
ΔT1

where ΔT1 and ΔT2 are the temperature differences at each

end of the heated section (inlet and outlet)

60
Turbulent flow in tubes and channels

Fully developed turbulent flow in tubes,

the Dittus - Boelter equation:
Nu = 0.023⋅Re0.8 ⋅ Prn
where n = 0.4 for heating of the fluid
n = 0.3 for cooling of the fluid

Valid in smooth tubes for Re > 10000

0.6 < Pr < 100.
Fluid properties at bulk temperature (≈average of inlet and outlet temperatures).

May, for fluids with low viscosity (μ < 2⋅μH2O ), be used when Re> 2300, that is for the
whole turbulent region.

61
Entrance region, turbulent flow

For the entrance region and for short tubes (10<L/d<400):

Nu = 0.036⋅Re0.8 ⋅ Pr1/3 ⋅ (d/L)0.055

where d = diameter of tube (m)
L = length of tube (m)

Thermal properties at bulk temperature.

62
Non-circular cross-sections, hydraulic diameter

Calculate the diameters as hydraulic diameters, DH , defined

as
DH = 4⋅A/P
where A = cross sectional area of the flow (m2)
P = wetted perimeter (m)
Example:

A = 4 cm ⋅ 8 cm =32 cm2 Cross-

P = 4 + 4 +8 + 8 = 24 cm 4 cm
section
⇒ DH = 4 ⋅ 32 / 24 = 5.33 cm
8 cm
63
Laminar flow in tubes and channels

When the temperature field is fully developed, the Nusselt

number is constant!

The value of the constant depends on the boundary conditions

and on the shape of the cross section

The temperature field is considered to be fully developed for

Gz <10.

64
 Nu number in fully developed laminar flow in tubes and
channels:

Geometry of cross Nu Nu
section (Constant wall temp) (Constant heat flux)
Triangular (equilat.) 2.47 1.89
Square 2.98 3.09
Circular 3.66 4.36
Two infinite plates 7.54 8.24

65
Entrance region in laminar flow

For the entrance region in laminar flow (Gz>10),

(average Nusselt number)

Nu = 1.86⋅Gz1/3 ⋅ (μ / μw)0.14
where μ = the dynamic viscosity of the fluid
at the mean bulk temperature
μw = the dynamic viscosity of the fluid
at the wall temperature

66
Fig. 9 (Fig. 6-5 in Holman)

67
 Flow across a plate, laminar and turbulent flow

Plate heated (or cooled) right from the leading edge,

laminar flow (Rex < 5⋅105 , 0.6<Pr<60, isothermal surface):

Nux = 0.332 ⋅ Rex1/2⋅ Pr1/3 Local Nu number at x.

NuL = 0.664 ⋅ ReL ⋅ Pr

1/2 1/3
Average Nu number for
distance 0 to L from edge.

Thermodynamic properties at film temperature (average of

wall and free stream temperatures).

68
Turbulent flow across plates

For (5⋅105 < Re < 107)

Nux = 0.0296 ⋅Rex0.8 ⋅Pr1/3 Local Nusselt number at x

Average Nu number for both the laminar part of the flow

closest to the edge and the turbulent part (5⋅105 < Re < 107)
NuL = (0.037⋅ReL0.8 - 871) ⋅ Pr1/3

69
Forced convection across single cylinders

Average Nusselt number for the circumference:

Nu = C⋅ Ren ⋅ Pr1/3

Reynolds number (x=d) C n

0.4 - 4 0.989 0.330
4 - 40 0.911 0.385
40 - 4000 0.683 0.466
4000 - 40’000 0.193 0.618
40’000 - 400’000 0.0266 0.805

Fluid properties at film temperature.

70
Forced convection across tube banks

Average Nusselt number:

Nu = C⋅ Ren ⋅ Pr1/3

Constant C and the exponent n are dependent on the ratios of

the tube spacings (normal and perpendicular to flow) and the
tube diameter

(See table 6-4 in Holman).

The Reynolds number should be calculated using the

maximum velocity occurring in the tube bank.

71
Example to solve:

Find the average Nusselt no as air flows past a 0.3 m heated

plate at 0.5 m/s. For the air, ν =15.7 ⋅ 10-6 m2/s, Pr = 0.72.

72
 Free convection
• In free convection the fluid flow is induced by density
differences caused by temperature differences.

• The flow is caused by buoyancy forces, is dependent on

gravitation to appear.

• The thermal and velocity boundary layers in free convection

are quite different from those in forced convection.

73
• The critical Grashof number, where the transition occurs
from laminar to turbulent flow depends on the geometry.

74
Free convection boundary layer
on a vertical flat plate

75
General correlation for free convection

In many different geometries, the average Nusselt number

may be calculated by as
Nu = C⋅(Gr⋅ Pr)m

The constant C and the exponent m depend on the geometry

and on the size of (Gr⋅Pr).

Thermodynamic properties at film temperature.

76
Constant C and exponent m in general corr. for free conv.

Geometry Gr⋅ Pr C m
Vertical plates 104 - 109 0.59 1/4
and cylinders 109 - 1013 0.10 1/3
Horizontal cylinders 104 - 109 0.53 1/4
9 12
10 - 10 0.13 1/3
Upper surface of heated 2⋅104 - 8⋅106 0.54 1/4
plates, or
lower surface of cooled plate 8⋅106 - 1011 0.15 1/3
Lower surface of heated 105 - 1011 0.27 1/4
plates, or
upper surface of cooled plates
77
Free convection - Gr⋅Pr/(ΔT⋅L3) tabulated

• The group

Gr⋅Pr/(ΔT⋅L3) = g⋅β/ν2

is only dependent on material properties and may be found in

the Collection of formulas and tables for different fluids.

78
Free convection - The exponent m

• The exponent m is often 1/3 (turbulent flow) or ¼ (laminar

flow).

• With m = 1/3, h is independent of L as it cancels out of the

equation:

Nu = C⋅(Gr ⋅ Pr)1/3
h⋅L/k = C ⋅[(g⋅β⋅ΔT⋅L3/ν2) ⋅ Pr]1/3
⇒ h independent of L

79
Free convection - Vertical plates, vertical cylinders

• Isothermal vertical plates: L = height of plate.

• Vertical cylinders: treat as vertical plates, if

D/L ≥ 35/Gr1/4

• For Gr < 104, use graphical solution (fig. 12)

80
Free convection, simplified correlations for air

For a given fluid, the dimensionless equation 34 may be

rewritten to give the heat transfer coefficient directly.
Two cases have to be discerned, laminar and turbulent:

For laminar flow:

h = Kl ⋅ (ΔT/H)¼ for 104 < Gr⋅Pr < 108
For turbulent flow
h = Kt ⋅ ΔT1/3 for 108 < Gr⋅Pr < 1012

Where Kl and Kt from Table 6 for air (assuming SI-units).

81
Free convection, simplified correlations for air
Table 3: Constants Kl and Kt for air

Tfilm (Gr⋅Pr)/(ΔT⋅H3) Kl Kt
-50 34.8⋅107 1.57 1.88
0 14.5⋅107 1.49 1.66
50 6.75⋅107 1.41 1.48
100 3.47⋅107 1.35 1.33
7
200 1.18⋅10 1.27 1.14
300 5.1⋅106 1.21 1.01
400 2.54⋅106 1.15 0.91
600 0.85⋅106 1.06 0.76

82
Vertical
Isothermal
Plates,
Nu vs Gr*Pr

83
Free convection from horizontal cylinders

• Constants C and m from table.

• Tube diameter is characteristic length.

84
Free convection from horizontal surfaces

Two cases:
1. Top side of a heated plate and bottom side of a cooled
plate. (Gravitation will force fluid away from surface)
2. Bottom side of a heated surface and top side of a cooled
plate. (A stable layer of fluid will form, decreasing the heat
transfer).

• C and m from table. C is exactly twice as high in case 1.

• Characteristic length: L = A/P

where A is the area and P the perimeter of the surface.
85
Free convection in between vertical plates

Correlation for the average Nusselt number:

−1/ 2
⎡ C1 C2 ⎤
Nuas = ⎢ 2 + 1/ 2 ⎥
⎣ ( Ra s ⋅ s / L) ( Ra s ⋅ s / L) ⎦
where s = distance between plates (m)
L = plate height (m)
Ras = (Gr ⋅ Pr), with s as characteristic length.
Nua,s = average Nusselt no, with s as the
characteristic length.
• ΔT in Gr is the temperature difference between the fin
surface and the undisturbed air.

86
Free convection in between vertical plates

Constants C1 and C2

Boundary condition C1 C2
Symmetric isothermal plates 576 2.87
Symmetric isoflux plates 48 2.51
Isothermal/adiabatic plates 144 2.87
Isoflux/adiabatic plates 24 2.51

87
Free convection in between vertical isothermal plates
7

5
Plate s = 15 mm

s = 10 mm
4
h (W/(m2 K))

s = 8 mm

s = 6 mm

2
s = 5 mm

1
s = 4 mm

0
0 50 100 150 200 250 300
Dt/H
Dt/H
(air @ 60°C)

88
Free convection in between vertical plates, different
boundary conditions, (plate distance 8 mm, air @ 60°C)
7

5
(W/(m2 K))

3
h

2
h, Isothermal
h, Isoflux
1
h, Isothermal/adiabatic
h, Isoflux/adiabatic
0
0 50 100 150 200 250 300
Dt/H

89
Flow in an enclosed space

90
Free convection in enclosed spaces (fig 15)

Define a heat transfer coefficient using the surface

temperatures between which heat is transferred:
h = q/(A⋅(T1 - T2))

Use the distance δ between the surfaces as the characteristic

length

The Nusselt number is expressed as the ratio between an

apparent, or effective thermal conductivity ke and the normal
conductivity k.
Nu = ke / k

91
We get:
q/A = h ⋅ (T1 - T2) = (Nu ⋅ k/δ )⋅ (T1 - T2) = ke /δ ⋅ (T1 - T2)

with Nu ⋅ k = ke

92
For vertical enclosures and constant heat flux

Nu = 0.42 ⋅ (Gr ⋅ Pr)1/4 ⋅ Pr0.012 ⋅ (L/δ)-0.30

Conditions: 104 < Gr ⋅ Pr < 107

1 < Pr < 20000
10 < L/δ < 40

where δ = the distance between the plates (m)

L = the height of the enclosure (m)

93
Vertical enclosures - Larger values of Gr⋅Pr

Nu = 0.046 ⋅ (Gr ⋅ Pr)1/3

Conditions 106 < Gr ⋅ Pr < 109

1 < Pr < 20
1 < L/δ < 40

Use δ, as characteristic length in Nu and Gr,

ΔT = difference between the surfaces.

94
Vertical enclosures - Low Gr ⋅ Pr

At Gr ⋅ Pr < 103 there is no free convection and heat is

transferred by conduction only (Nu = 1).

95
Vertical enclosures, comparison of eqs. 41 and 42.
100
Nu eq38
Nu eq. 41
Nu
Nu eq.
eq3942
Conduction
Nu

10

Nu=1
1
1.E+02 1.E+03 1.E+04 1.E+05 1.E+06 1.E+07 1.E+08 1.E+09 1.E+10
Gr Pr

96
Horizontal enclosures

Two cases:
1.If the top plate is the hotter, a stable situation will occur and
heat is transferred by pure conduction, and thus Nu = 1 (or
ke /k =1).

2.If the bottom plate is the hotter, pure conduction will occur
at Gr < 1700, while at higher values, convection cells will
occur, increasing heat transfer. See table 7-3 in Holman.

97
Combined free and forced convection

To determine which type of convection that is dominant, the

following inequality may be used.

Gr / Re2 > 10

If this criterion is fulfilled, free convection is dominant.

98
Example to solve:

Find the Nusselt no at a heated vertical plate, if the Grashof

no is known to be 1⋅108 and Pr is 1.

99
 Part 3

Radiation

100
Radiation

Thermal radiation.:
”That electromagnetic radiation emitted by a body as a result
of its temperature”.

Thermal radiation is restricted to a limited range of the

electromagnetic spectrum.

101
Radiation

<>

102
103
Blackbody radiation

A blackbody is a perfect radiator.

Three characteristics:

• It absorbs all incident radiation

• It radiates more energy than any real surface at the same
temperature
• The emitted radiation is independent of direction

Also:
Blackbody radiation obey certain simple laws

104
Stefan-Boltzmann’s law

The total power radiated from a blackbody is calculated from

Stefan-Boltzmann´s law:

Eb = σ⋅T4

where Eb = total power radiated per unit area from

a blackbody (W/m2)
σ = 5.669 ⋅ 10 W/(m ⋅K ). (Stefan-Boltzmann
-8 2 4

constant)
T = absolute temperature (K)

105
Planck distribution law

The wavelength distribution of emitted blackbody radiation is

determined from the Planck distribution law (fig. 19):

C1
E λ ,b ( λ , T ) = 5
λ ⋅ [exp( C2 / ( λ ⋅ T )) − 1]

where C1 = 2π⋅h⋅co2 = 3.742⋅108 W⋅μm4 / m2

C2 = (h⋅co / k) =1.439 ⋅ 104 μm K
λ = wavelength (μm)
T = absolute temperature (T)
h = Planck constant
106
k = Boltzmann constant
c0 = speed of light in vacuum

107
Result of increasing temperature on radiation

• Higher intensity
• Shorter wavelength Ù higher frequency

108
 Wien’s displacement law

The wavelength of maximum emissive power is determined

by Wien’s displacement law:

λmax ⋅ T = C3 = 2897.8 μm K

109
Radiation from real surfaces

Real surfaces:
• emit and absorb less than blackbodies
• reflect radiation
• emit and absorb differently depending on angle and
wavelength
• do not obey the simple laws

110
Blackbody and real surface emissions

111
Spectral emissivity and total emissivity

To account for ”real surface”- behavior we introduce the

spectral emissivity, defined by

Eλ(λ,T) = ελ(λ,T) ⋅ Eλ,black (λ,T)

and the total emissivity defined by

E = ε ⋅ Eb = ε ⋅ σ ⋅ T4

(ε = integrated average)

112
Gray diffuse body

To simplify matters it is common to assume the emissivity to

be independent on wavelength and direction. Such a surface is
called a gray diffuse body, for which

Eλ(λ,T) = ελ ⋅ Eλ,black (λ,T)

ελ = constant = ε (0 < ε < 1)

113
Absorptivity, reflectivity, transmittivity

Incident radiation may be absorbed, reflected or transmitted

We define
Absorptivity α: Fraction of incident radiation absorbed
Reflectivity ρ: Fraction of incident radiation reflected
Transmittivity τ: Fraction of incident radiation transmitted
thus α+ρ+τ=1

α
114 τ
Kirchhoff’s identity

The (total) emissivity and the (total) absorptivity of a surface

are equal at equal temperatures (wavelengths):

ε=α

115
Radiation exchange between blackbodies

To calculate radiation exchange we must take into account

• surface areas
• surface geometries
• position in relation to each other
This is done by the shape factor, F12
F12 = fraction of radiation leaving surface 1 intercepted
by surface 2.
2

1
116
Net exchange of radiation, law of reciprocity

Net exchange of radiation between blackbodies:

q1-2 = F12 ⋅ A1 ⋅ (Eb1 -Eb2) = F12 ⋅ A1 ⋅ σ ⋅ (T14 - T24)

Law of reciprocity

F12 ⋅ A1 ⋅ = F21 ⋅ A2

117
Shape factors

Shape factors are often difficult to calculate.

See diagrams and formulas in
Holman, Figs 8.12-8.16 and
CFT pp. 45-47

Important special case:

Small (convex) surface (1) surrounded by other surface (2):
ÖF12 = 1.

118
 Radiation exchange between real surfaces, simple case

For the special case of F12 = 1 the exchange is calculated as

q1-2 = ε1 ⋅ A1 ⋅ (Eb1 -Eb2) = ε1 ⋅ A1 ⋅ σ ⋅ (T14 - T24)

Comparing to Newton’s law of cooling:

q = hr ⋅ A1 ⋅ (T1 -T2)

Öhr = ε1 ⋅ σ ⋅ (T14 - T24) / (T1 - T2) = ε1 ⋅ hr, black

Öhr, black = f(T1, T2), from table!
F12 = 1

A1
119
Table 8 (p.30 in CFT)

120
Total emissivities of selected materials

Material Temp (°C) Emissivity ε

Aluminum, commercial sheet 100 0.09
Copper, polished 100 0.052
Iron, dark-gray surface 100 0.31
Glass, smooth 22 0.94
Snow-white enamel varnish 23 0.906
Black shiny lacquer 24 0.875
Roofing paper 21 0.91
Porcelain, glazed 22 0.92
Red brick 23 0.93
Al-paints 100 0.27-0.67

121
Radiation exchange calculated by resistance networks

Consider the blackbody emissive power as the driving

potential

q1-2 = (Eb1 -Eb2) / Rrad = σ ⋅ (T14 -T24) / Rrad

Consider the radiation resistance as the sum of surface and

space resistances
Rrad = Rsurface1 + Rspace + Rsurface2
J1 J2
Eb1 Eb2
Rsurface1 Rspace Rsurface2
122
Radiation exchange by resistance networks, assumptions

Assume:
• All surfaces are gray,
• All surfaces are uniform in temperature.
• Reflective and emissive properties are constant over the
surfaces.
Define:
• Irradiation, G = total radiation / (unit time, unit area)
• Radiosity, J = total radiation leaving /(unit time, unit area)
(including reflected radiation)
Assume these properties are uniform over each surface.

123
Surface resistance:

Radiosity = emitted radiation + radiation reflected:

J = ε ⋅ Eb + ρ ⋅ G
where ε = emissivity
ρ = reflectivity of surface
Assume surfaces opaque (τ=0)
=> ρ = 1 - α = 1 - ε
(as α = ε).

=> J = ε⋅ Eb + (1 - ε) ⋅ G

or G = (J - ε ⋅ Eb) / (1 - ε)
124
The net energy leaving the surface per unit area:

q / A = J - G = ε⋅ Eb + (1 - ε) ⋅ G - G =
=ε ⋅ Eb - ε ⋅ G = ε ⋅ (Eb -G)

ε⋅A ( Eb − J )
q= ⋅ ( Eb − J ) =
=> 1− ε (1 − ε ) / (ε ⋅ A)

Consider (Eb - J) as the driving potential.

=> Rsurface = (1-ε)/(ε ⋅ A)

125
Space resistance:

Radiation from 1 to 2 (per unit time): J1⋅A1⋅F12 .

Radiation from 2 to 1 (per unit time): J2⋅A2⋅F21.
Net exchange of radiation:
q12 = J1⋅A1⋅F12 - J2⋅A2⋅F21 = (J1 -J2)⋅A1⋅F12 = (J1 -J2)⋅A2⋅F21
(as A1⋅F12 = A2⋅F21 ).

Consider (J1 -J2) as the driving potential,

=> Rspace = 1/(A1⋅F12)

126
Heat exchange due to radiation, two bodies

E b1 − E b 2 E b1 − E b 2 σ ⋅ (T14 − T2 4 )
q= = =
Rrad Rsurface1 + Rspace + Rsurface 2 1 − ε1 1 1 − ε2
+ +
ε1 ⋅ A1 A1 ⋅ F12 ε 2 ⋅ A2

127
Special case, two parallel infinite plates

q σ ⋅ (T14 − T24 )
=
A 1 1
F12 = 1 and A1 = A2 + −1
ε1 ε2

Special case, two concentric cylinders or spheres

q σ ⋅ (T14 − T24 )
=
A1 1 A1 1
F12 = 1 + ⋅ ( − 1)
ε1 A2 ε2
If A1 <<A2 q /A1 = ε1 ⋅ σ ⋅ (T14 - T24)
128
Three body problem

• Calculate all resistances.

• For the nodes J1, J2 and J3, the sum of the energy flows into
each of the nodes must be zero.
Example, node J1
J1 Rspace12J2
E b1 − J1 J 2 − J1 J 3 − J1 Eb Eb
+ + =0
⎛ 1 − ε1 ⎞ A1 ⋅ F12 A1 ⋅ F13 Rsurface Rsurface
⎜ ⎟
⎝ ε1⋅ ⋅ A1 ⎠ Rspace13 Rspace23
• Solve J1 -J3 !
• Calculate the heat exchanges! J3
Rsurface
Eb
129
Radiation shields (q/A)13 (q/A)32

q σ ⋅ (T14 − T24 )
= 1 2 1 3 2
A 1 / ε1 + 1 / ε2 − 1 No shield With shield

(q/A)1-3 = (q/A)3-2 = (q/A)

q σ ⋅ ( T14 − T34 ) σ ⋅ ( T34 − T24 )

= =
A 1 / ε1 + 1 / ε 3 − 1 1 / ε 3 + 1 / ε 2 − 1

Everything is known except the temperature T3.

130
Simplest case, all emissivities equal:

T14 - T34 = T34 - T24 => T34 = ½ ⋅ (T14 + T24)

2 ⋅ σ ⋅ (T − T )
1 4 4
q
= 1 2

=> A 1 / ε3 + 1 / ε2 − 1

Emissivities assumed equal, => heat flux reduced to half

131
Multiple shields of equal emissivities

It may be shown by similar reasoning that the heat flux will

be reduced to

1
(q / A) with shields = ⋅ (q / A) without shields
n +1

where n is the number of shields.

132
Different emissivities

Two plates with equal emissivities ε1

One shield of different emissivity ε2:
ε1 − 1
2
1
(q / A) with shields = ⋅ 1 1 (q / A) without shields
2 ε1 + ε2 − 1
ε2 << ε1 => largest decrease in heat flux

Assume ε1 = 1. =>
1 2 −1 1
(q / A) with shields = ⋅ (q / A) without shields = ⋅ ε2 ⋅ (q / A) without shields
2 1 + ε2 − 1
1
2
thus factor equal to one half times the emissivity of the shield.

133
 Part 4

Conduction,
Boiling, Condensation
and Diffusion

134
Conduction

Introduction

Fourier’s law:
q = -k⋅A⋅δT/δx

For a plane wall, the temperature gradient is constant

q = k⋅A⋅ΔT/δ

135
Conduction in cylindrical shells

When heat is conducted through a wall which is not plane, the

temperature gradient will no longer be constant through the
wall.

q = -k⋅A⋅δT/δx

At any section, q is constant

Assume k constant
⇒ (A⋅δT/δx) = constant.
Thus, if the area A changes with x, as it will for a curved
surface, the temperature gradient must also change.

136
Conduction in cylindrical shells

The simplified version of Fourier´s law may still be used if

we use an average area (or average diameter)

For a cylindrical wall the area is calculated as the logarithmic

mean of the inside and outside surface.

Aln = (A2 - A1 )/ln(A2 /A1) =

π ⋅ L ⋅ dln = π ⋅ L ⋅ (d2 - d1 ) / ln(d2 / d1 )

137
Conduction in cylindrical shells

138
Conduction in fins

Flänsarnas area=Af

Area mellan flänsarna =A2

t t1 tf

ϑmf ϑrot
tf

t2
Fig 11.5 i termo

139
Conduction in fins

Definition of fin efficiency

Because of the lower ΔT, lower heat flux from fin

⇒ The fin surface is less efficient than the base surface,

We define a fin efficiency ηf as the ratio between the actual

heat transferred and the heat which would be transferred if
the whole fin had the base temperature.

ηf ≡ ΔTm, fin / ΔTbase

(Assuming equal h on fin and base)

140
Conduction in fins

For straight fins with constant cross section:

ηf ≡ tanh(m ⋅ L) / (m ⋅ L)

where L = the length of the fin (m)

m = [(h ⋅ P)/(k ⋅ A)]1/2
where P = perimeter of the fin (m)
A = cross section area perpendicular
to heat flow (m2)
For many types of fins, the fin efficiency may be estimated
from the following diagram.

141
142
Overall heat transfer coefficient
For a finned surface, the area is calculated as
Atot = Abase + ηf ⋅ Afin
The U-value of a finned tube is calculated as:
1/(U⋅A) = 1/(h1⋅A1) + δ/(k⋅Aln) + 1/[h2⋅(Abase + ηf⋅Afin )]
where A1 = inside surface area (m2)
Abase = outside surface area in between fins (m2)
Aln = logarithmic mean area of tube (m2)
Afin = fin area (m2)

143
Electric analogy for solving 2-d conduction problems

Make electric model of 2-d conduction problem in:

• Electrolyte tray
• Thin conducting foil
• Resistance mesh

Boundary conditions: Isothermal or adiabatic

Inside of wall, T1

Line of symmetry
Line of symmetry Line of symmetry

Points of equal Rail, T2

potential

144 Outside of wall, T2

Electric analogy models…

Electrolyte tray or foil - continuous models

Resistance mesh: discreet model

Measurements show : Potential in node = mean of the four

surrounding nodes
ÖMeasurements not
necessary!
Calculate! R = 50 Ohm

145
Excel model of wall with rail, horizontal cut
A B C D E F G H I J K L
1 200 200 200 200 200 200 200 200 200 200 200 200
2 188.7026 188.6737 188.5876 188.447 188.2564 188.0241 187.7626 187.49 187.2305 187.013 186.8668 186.8151
3 177.463 177.4044 177.2299 176.9438 176.5547 176.0775 175.5362 174.9668 174.419 173.9545 173.6391 173.5268
4 166.3407 166.2511 165.9838 165.5437 164.941 164.1949 163.3379 162.4221 161.5243 160.747 160.2085 160.0137
5 155.3975 155.2755 154.9105 154.3061 153.4708 152.4232 151.1985 149.8595 148.5091 147.3007 146.434 146.1111
6 144.6982 144.5429 144.0767 143.2993 142.2131 =(D5+E6+C6+D7)/4
140.8286 139.1733 137.3082 135.3518 133.5128 132.1159 131.5626
7 134.3096 134.1213 133.5542 132.6013 131.2538 129.5046 127.3581 124.8484 122.077 119.283 116.9541 115.9075
8 124.2975 124.0787 123.4173 122.298 120.6961 118.578 115.906 112.6502 108.8248 104.5881 100.51 98.15913
9 114.723 114.4787 113.7382 112.4775 110.6545 108.2053 105.0376 101.0218 95.98379 89.73458 82.33865 75.70911
10 105.637 105.375 104.5793 103.2192 101.2393 98.55121 95.01738 90.41565 84.35401 76.02783 63.40091 40
11 97.07489 96.8051 95.98494 94.58083 92.53218 89.74285 86.06506 81.26942 74.98878 66.6218 55.23719 40
12 89.05243 88.78556 87.97451 86.58699 84.56579 81.82295 78.2306 73.60818 67.7099 60.23341 50.92604 40
13 81.56375 81.31023 80.54056 79.22686 77.32104 74.75256 71.42625 67.22281 62.00923 55.67591 48.23357 40
14 74.58213 74.35108
=(A13+B14+B14+A15)/4
73.65065 72.45886 70.73898 68.44003 65.49901 61.84761 57.42829 52.22745 46.33231 40
15 68.06261 67.86134 67.25212 66.21895 64.736 62.76959 60.28217 57.24033 53.62889 49.47328 44.86824 40
16 61.94566 61.77957 61.27757 60.42882 59.21651 57.62016 55.61977 53.20265 50.37367 47.16856 43.66736 40
17 56.1609 56.03372 55.64977 55.00226 54.08106 52.87477 51.3741 49.57686 47.49456 45.15991 42.63266 40
18 50.63054 50.54463 50.28553 49.84939 49.23071 48.42378 47.42502 46.23611 44.86782 43.34387 41.70337 40
19 45.27201 45.22874 45.09834 44.87908 44.5686 44.16462 43.66609 43.07473 42.39673 41.64439 40.83694 40
20 40 40 40 40 40 40 40 40 40 40 40 40

146
Tools/Alternatives/Calculate/Iteration

147
Transient heat transfer - Lumped capacitance method

Use when temperature of the body is approx. uniform.

I.E. if Rth, surface >> Rth, internal Ù h << k / L

We define the Biot number (Bi) as

Bi = h⋅L / k
where h = heat transfer coefficient
k = thermal conductivity of solid
L = characteristic length of solid body
= V/A =volume of body /exterior surface area

148
Transient heat transfer - Lumped capacitance method…

Method can be used when Bi < 0.1.

After a sudden change in ambient temperature:

ΔT = ΔT0 ⋅ exp(-Bi ⋅ Fo)

where ΔT = temperature difference at time τ.
ΔT0= temperature difference at τ =0.
Fo = k⋅τ/(ρ⋅cp⋅L2) = Fourier number (dim. less time)
τ = time after step-change in ambient temperature
k = thermal conductivity of solid body
ρ = density of solid body
cp = specific heat of solid body
149
Transient heat transfer - Lumped capacitance method…

Product of the Biot and Fourier:

h ⋅ (V / A) k ⋅τ h⋅ A
Bi ⋅ Fo = ⋅ 2 = ⋅τ
k ρ ⋅ c p ⋅ (V / A) ρ ⋅ c p ⋅V

Numerator: = 1/Rth, convection

Denominator: Cth Heat capacity of the solid

Rth = 1/(h⋅A) and Cth = m⋅cp = ρ ⋅ V ⋅ cp

=> Bi ⋅ Fo = τ /(R⋅C)th

150
Transient heat transfer - Lumped capacitance method…
τ
ΔT −
Rth ⋅Cth
⋅
=e
=> ΔT0

Compare equation for discharge of a condenser through a

resistor!

Lumped capacitance method handy for quick estimates!

151
Transient heat transfer - Heisler charts

For simple geometries, use diagrams (Heisler charts) where

ΔT/ΔT0 = f (Fo, Bi)

Diagrams are found in:
• Collection of formulas and tables, pp. 42-43,
• book by Holman, fig. 4-7 - 4-13.
Note that the definition of the characteristic length used in Bi
and Fo in the charts are not the same as used above!

Charts valid for Fo > 0.2.

152
Two- or three dimensional problems: multiply the solutions
for the corresponding one-dimensional cases.

Examples:
The temperature in the center of a short cylinder:

(ΔT/ΔT0)center, short cylinder = (T/ΔT0)center, infinite cylinder ⋅

⋅(ΔT/ΔT0)center, infinite wall

Temperature in the center of a cube:

(ΔT/ΔT0)center, cube = [(ΔT/ΔT0)center, infinite wall]3

153
Boiling

Two types of boiling:

pool boiling, (as on the outside of tubes)

and flow boiling (boiling inside tubes)

154
Pool Boiling

The boiling curve

155
Pool Boiling

A simple correlation for nucleate pool boiling:

hnb = C ⋅ 55 ⋅ pr(0.12 - 0.2 ⋅ log10 Rp) ⋅ (-log10 pr)-0.55 ⋅ M-0.5 ⋅ (q/A)0.67

where C = 1 for horizontal, plane surface

C = 1.7 for horizontal copper cylinders
pr = p/pcritical = reduced pressure
Rp = Surface roughness (μm) (about 1 for many
technically smooth surfaces)
M = molecular weight of the fluid (kg/kmol)
q/A = heat flux (W/m2)

156
Flow boiling

Flow regimes in flow boiling

157
Flow boiling

Flow boiling of refrigerants in horizontal tubes:

Complete evaporation: Num = 1.0 ⋅ 10-2 ⋅ (Re2 ⋅ Kf )0.4

Incomplete evaporation: Num = 1.1 ⋅ 10-3 ⋅ Re ⋅ Kf 0.5

where Re = 4 ⋅ m’ /(π⋅d⋅μl)
where m’ = mass flow (kg/s)
d = tube diameter (m)
μl = dynamic viscosity of liquid (Ns/m2)

158
 Kf = Δi/(L ⋅ g)
where Δi = specific enthalpy difference across
tube (J/kg)
L = tube length (m)
g = acceleration of gravity (m/s2)

The equations give average Nusselt number for the tube.

159
Condensation

Two types of condensation:

Film condensation (when the liquid wets the surface)
Drop condensation (opposite)

Drop condensation best but difficult to achieve. Calculate for

film condensation!

Heat transfer resistance in condensation is entirely due to

conduction of heat through the liquid.

160
Condensation

⇒ Heat transfer coefficient may be calculated from the film

thickness as

h = k /δ
where k = thermal resistance of liquid (W/(m⋅°C))
δ = film thickness (m)

Film thickness in laminar flow at a distance x from the top

δ = {4⋅ μ⋅ k ⋅x⋅ ΔT / [g⋅hfg ⋅ρ⋅(ρ - ρv)]}1/4

where μ = dynamic viscosity of liquid (Ns/m2)
161
 x = distance from top of surface (m)
ΔT = temperature difference between surface
and vapour (°C)
g = acceleration of gravity (m/s2)
hfg = heat of vaporisation (J/kg)
ρ = density of liquid (kg/m3)
ρv = density of vapour (kg/m3)

162
Condensation

This gives the local heat transfer coefficient

h = {g⋅hfg ⋅ρ⋅(ρ - ρv)⋅ k3/ [4⋅ μ ⋅x⋅ ΔT ]}1/4

and by integrating, the average heat transfer coefficient

hav =4/3⋅hL =
= 4/3⋅{g⋅hfg ⋅ρ⋅(ρ - ρv)⋅ k / [4⋅ μ ⋅L⋅ ΔT ]} =
3 1/4

= 0.943⋅{g⋅hfg ⋅ρ⋅(ρ - ρv)⋅ k3/ [ μ ⋅L⋅ ΔT ]}1/4

Fluid properties at film temperature.

163
Condensation

Alternative eq if q/A rather than ΔT is known:

hav = 0.924⋅k⋅{g⋅hfg ⋅ρ⋅(ρ - ρv) / [ μ ⋅L⋅ q/A]}1/3

For single horizontal tubes:

exchange L for d,
and the constant 0.943 is changed to 0.725.
(For n horizontal tubes placed on top of each other, the
diameter d should be multiplied by n).

164
For condensation inside horizontal tubes,

The liquid will more or less fill the tube.

It has been suggested that could be used in this case too, but
with the constant changed from 0.943 to 0.555.

As a rule of thumb, condensation heat transfer coefficients of

refrigerants (not NH3) is about 2000 W/(m2⋅K).

165
Condensation

In turbulent flow (Re>1800), plane vertical surfaces

Nu = 0.0030 ⋅ G ⋅ Cv-1/2 for Re> 1800

where Nu = h ⋅ L / k
G = g ⋅ L 3 /ν2
Cv = g⋅ρ⋅L /( k⋅ν ⋅ΔT)
3

Re = 4⋅m’/(P⋅μ)
where m’ = mass flow (kg/s)
P = wetted perimeter (= width for plane
wall, = π⋅d for vertical tube)
166
Diffusion m& d m& d

Luft
air
Våtsurface
Wet yta

på påv′′ på påv′′ på på

Water or frost collects on cold surfaces,

and
wet surfaces dry up:

In these processes water vapor is transferred by diffusion and

convection to or from the surface.

Heat of vaporization is transferred to or from the surface.

167
Diffusion

The heat flow is calculated as:

qd = m´d ⋅ hfg

where m´ = mass of water changing phase per unit time

We define a diffusion heat transfer coefficient hd as

q d = hd ⋅ A ⋅ Δ T
2
where A = wet area (m )
ΔT = temp diff between wet surface and air

168
Diffusion

The ratio between the diffusion heat transfer coefficient hd

and the convection heat transfer coefficient hc may be
calculated by the equation

hd / hc ≈ C ⋅ Δp/ΔT

where C = 1520 above a wet surface and

C = 1750 above a frozen surface

169
 ΔT = T”vw -Tv , temperature difference between
wet (frozen) surface and air (°C)
Δp = p”vw - pv = difference in partial pressure of
water vapour between the surface and
the free air (in bar)
p”vw = saturation pressure of water at the
temperature of the surface
(from steam table).
pv = ϕ ⋅ p”v
ϕ = relative humidity (%)
p”v = saturation pressure of water at the
temperature of the free air (from steam table).

170
Diffusion

Note that hd / hc may be positive or negative, indicating heat

flow in opposite directions.

The ratio may also be attained directly from the T, pv diagram

(fig 28)

To find the heat transfer by diffusion, we first calculate the

convection heat transfer coefficient according to Lesson 2,
then calculate the ratio hd / hc.
Then we can calculate qd and qtot as htot = hc + hd = hc (1 +
hd/hc)

171
 t, på - diagram för
fuktig luft
Diffusion

på på x

t
h
(α d / αkw )frost (α d / αkw )

t
ϕ

h
t
h

172 t
 Part 5

Heat exchangers

173
Heat exchangers

Three main types of heat exchangers:

In recuperative heat exchangers, the heat is transferred from
one fluid to the other through a dividing wall.

In regenerative heat exchangers, heat is transferred via a

moving solid part, (a rotating wheel with a large number of
narrow channels).

In evaporative heat exchangers, hot water is cooled by partly

evaporating the water.

174
Heat exchangers
Regenerativ
värmeväxlare

fyllnings-
material i
en rotor som
långsamt roterar

Evaporativ
värmeväxlare
(kyltorn)

Rekuperativ
kyltorns-
värmeväxlare
fyllning
späd-
vatten luft
kylt vatten
pump för varmt vatten
cirkulation av
vatten genom
kyltornsfyllningen
Fig 29 (Fig. 11.18 i termo)

175
Heat exchangers

Two basic types of recuperative,

Counterflow and parallel flow (fig 30).
Motströms 2 Medströms

1 1

2
t
t1 t1
Δ1
ϑ2 Δ1 ϑ2 = θ
t2 θ ϑ1
Δ2 ϑ1 t2 Δ2

Area Area
0 A 0 A

a) b)
Counterflow Parallel flow

176
Heat exchangers

The heat exchange rate from flow:

q = mh⋅ ch⋅(Th1 -Th2 ) = mc⋅ cc (Tc1 -Tc2 )

where m = mass flow (kg/s)

c = heat capacity (J/(kg⋅°C))
index h and c refers to the hot and cold fluid
index 1 and 2 refers to the two ends.

The product (m⋅c) is called the heat capacity rate and is often
written as C.

177
Heat exchangers

Heat exchange rate from size:

q = U⋅A⋅ϑln

where ϑln = log mean temp difference (or LMTD)

(For parallel flow and counter flow heat exchangers)

ϑln = [(Th2 - Tc2 ) -(Th1 - Tc1 )] / ln[(Th2 - Tc2 ) / (Th1 - Tc1 )]

= (ΔT2 - ΔT1) / ln(ΔT2 / ΔT1)

178
Heat exchangers

A third basic type of flow is crossflow.

The temperatures of both fluids at the outlets are different
from one side to the other of the flow channel.

In most heat exchangers, the flow is neither purely parallel,

counterflow or crossflow, but rather a mixture of these types.

The logarithmic mean temperature difference may still be

used, if corrected by factor F.:

q = U⋅A⋅F⋅ϑln

179
Heat exchangers

Cross flow heat exchanger, temperature profile

1. Δ1

θ Δ2

2.

180
 Heat exchangers

A fourth basic type: Condensers and evaporators:

the temperature of one fluid is constant

The logarithmic mean temperature may be used without any

correction (F=1)

181
Heat exchangers

Temperature effectiveness, NTU-method:

The temperature effectiveness is defined as

ε = actual heat transfer / maximum possible heat transfer

The maximum possible heat transfer would result if the

temperature of the fluid with the lowest heat capacity rate (C
=m⋅c) at the outlet of the heat exchanger reach the inlet
temperature of the other fluid.

182
Heat exchangers

(This fluid is referred to in Holman as the minimum fluid. =

= the fluid with largest temperature change).

The definition gives

ε = Cmin ⋅ΔTmin / Cmin ⋅ΔTmax = ΔTmin / ΔTmax

where ΔTmin = temperature change of minimum fluid.

ΔTmax = the difference between the
inlet temperatures.

183
Heat exchangers

The temperature effectiveness is a function of only two

variables: UA/Cmin and Cmin / Cmax

ε = f(Cmin/Cmax , UA/Cmin )

UA/Cmin is called the number of transfer units (NTU).

See fig. 33.

Equations for f are found in table 10-3 i Holman.

184
Heat exchangers
MOTSTRÖMSVÄRMEVÄXLARE
1,0 η2
7
10
=
η1 3
5
0 ,9

ε
0,8 2 0 ,8

kA
1 ,5 0 ,7
1
Δ1
& 1 = 1,0
W θ
0 ,6
0,6
Δ2 2
0,75
UA/Cmin 0 ,5
A
0,4 0,50 0,4
Δ1 Δ2
0, 3 η1 = ; η2 =
θ θ
0,25 0,2 & = m& ⋅ c
W
0,2 1 1 p1

0,1 & 2 = m& 2 ⋅ c p

W 2
0,05
0 (η2 = η1 ⋅ W
& /W
1
& )
2

0 0,2 0,4 0,6 0,8 1,0

&
W 1
&
W2
Cmin/Cmax fig. 11.25 i Termo

185
 Lesson 6
Methods of enhancing heat
transfer

186
Enhanced heat transfer

What is enhanced heat transfer?

”An enhanced heat transfer surface has a special surface
geometry that provides a higher h⋅A value per unit base
surface area than does a plain surface.”

When should it be applied?

187
When should enhanced heat transfer be applied?

To the side of a heat exchanger where the dominant heat

transfer resistance lies.

It is the overall heat transfer resistance which determines the

temperature difference.

Example:
Heat exchanger with surface areas equal on both sides (1m2).
Total heat transfer resistance (wall is neglected):

1/(UA) = 1/(h1⋅A) + 1/(h2⋅A)

188
Assume h1 = 50 W/(m2⋅⋅°C) (forced flow air)
and h2 = 1000 W/(m2⋅⋅°C) (forced flow water)

The overall heat transfer resistance is

1/(UA)=1/(50⋅1) + 1/(1000⋅1) = 0.02 +0.001 =0.021°C/W

If we double the heat transfer coefficient on the water side

from 1000 to 2000 we would get

1/(UA)=1/(50⋅1)+1/(2000⋅1)=0.02 + 0.0005=0.0205°C/W
No improvement!

189
If we double the heat transfer coefficient on the air side, we
would get

1/(UA)=1/(100⋅1)+1/(1000⋅1) = 0.01+0.001 = 0.011°C/W

A reduction to half of the original total heat transfer

resistance.

190
Other aspects of enhanced heat transfer:

The costs of enhanced surfaces are usually higher than for the
plain surfaces.

The enhanced surfaces are also usually more sensitive to

fouling, which may be important in some applications.

191
What is enhanced heat transfer good for?

Enhanced heat transfer may be used for any of four reasons:

1. To reduce the temperature difference between the fluids at
a given heat exchanger size and capacity.

2. To reduce the size of the heat exchanger at a given capacity

and temperature difference.

3. To increase the capacity of a given surface area.

4. To reduce the pumping power for a given capacity and

temperature difference.

192
 Enhanced heat transfer in gas flow on finned surfaces

Louvered fins
Offset strip fins

Enhancement is achieved by making the fins short in the flow

direction.

In this way the boundary layer is not allowed to grow

193
Single-phase flow inside tubes

Surface with small protrusions, to disturb the boundary layer.

(”dimples”, grooves or minute fins).

Turbulator in the tube, inducing large undulations in the flow.

194
Enhancement of boiling heat transfer

Pool boiling
Enhanced surfaces act by facilitating the nucleation of vapour
bubbles.
Nucleation is facilitated by porous surface structure.

Small amounts of vapour is trapped during boiling.

Evaporation takes place in thin liquid films in the structure.

Enhancement larger for single tubes than for tube bundles.

195
Flow boiling

Nucleation is often suppressed by the convection.

Evaporation takes place at the vapour liquid interface.

A porous coating will induce nucleate boiling and enhance

heat transfer.

Most of the porous surfaces can not be made inside tubes.

Enhancements generally increase the convection in the fluid.

”Micro-fins” (0.2mm), give h-values 2-4 times smooth tube.

196
Star shaped aluminium insert: divides the tube into several
channels. Area is increased,
Hydraulic diameter is decreased,

197
Enhanced heat transfer in condensation

Heat transfer resistance in condensation is due to the

resistance in the liquid film.

Enhanced heat transfer is achieved by reducing the thickness

of this film.

Surface structure where surface tension act to gather the

liquid at certain locations while keeping other areas free of
liquid.

Surfaces often have fins with narrow spacing, surface tension

drags liquid from tips to ”valleys” in between fins.
198
Enhanced heat transfer in condensation inside tubes

Micro-fins have been shown to increase the heat transfer

coefficients by a factor of 2 to 3.

199
Home assignments:

Exercise 1: e Exercise 13 : a, b
Exercise 2: e Exercise 14 : b, d
Exercise 3: c Exercise 15 : a, c
Exercise 4 : d
Exercise 5 : c
Exercise 6 : c
Exercise 7 : a, b
Exercise 8 : c, d
Exercise 9 : a, b
Exercise 10 : d
Exercise 11 : c, d
Exercise 12 : a, b
200