EEN-1020 Heat transfer

Week 2: Fins, 2d Conduction,

Thermal Resistance, and Numerical
Solution in 2d
Prof. Ville Vuorinen
November 3rd-4th 2020
Aalto University, School of Engineering
 In heat transfer course, we have “4 friends”
who typically help us to approach and solve any
problem

1) Energy conservation
2) Fourier’s law
3) Newton’s cooling law
4) Bonus “helpers”:
4.1) heat eqn, conv.-diff. eqn (relatives of friend #1)
4.2) Navier-Stokes eqn (momentum conservation)
4.3) non-dimensional numbers (follow from the other
friends)
 From week1: example solutions of 1d heat
equation with two different boundary conditions
Relevance: HW1 and computer class

●
Diffusive processes are very slow in comparison to convective processes
●
Below, two examples of heat diffusion in iron (profiles taken from different
times)
●
Simulation time is in the order of 0.03-0.1s Color pic: visualization of temperature

Left and right end insulated i.e. dT/dx = 0 Left and right end fixed temperatures

Early time

Late time Late time

Early
 How could temperature distribution look like
in a heated 2d metal fin (assume here: no convection) ?
Model 1: Model 2:
- Cool sides and tip - Insulated sides
- Hot base - Cool tip and hot base
→ T=T(x,y,t) (2d) → T=T(y,t) (1d)
Lecture 2.1 Theory: Fins and thermal resistance

ILO 2: Student can apply Fourier’s law and Newton’s law in fin
theory and thermal resistance context. Further, the student can
analyse 2d heat transfer data in Matlab and formulate an energy
balance for 2d system.
Next, we talk about fins and
fin theory
 Example: cooling an electric circuit by placing
a cooling plate with 3d printed finned microchannels on
top of the circuit

Note:
fins

K.Saari, A.Laitinen, K.Kukko, P.Peltonen, V.Vuorinen, J.Partanen (Int.J.Heat and

Fluid Flow 2020)
 Fins – surface extrusions that increase area
of surface to increase heat transfer
3d printed heat exchangers intended for air cooling (V.Vuorinen, K.Kukko, K.Saari)

Key question:
temperature distribution in a fin ?

Plate fins Pin fins

 Fins – surface extrusions that increase area
of surface to increase heat transfer
3d printed heat exchangers intended for air cooling (V.Vuorinen, K.Kukko, K.Saari)

T=T(z)

Plate fins Pin fins

Local heat flux
q=h(T s −T ∞ )=h(T (z)−T ∞ )
 Observation: air temperature may
vary quite much inside a heat exhanger
CFD simulation of air temperature from a cross-section of heat exchanger under forced convection: P.Peltonen (2017)

c p ṁ Δ T =q

Plate fins Pin fins

https://www.youtube.com/watch?v=AI1voSK4AJc
For engineering estimates, fin theory offers
simplifications for heat transfer prediction:
1) fin = 1d and steady state assumption
2) surface temperature = T(x)
3) heat transfer coefficient constant
4) possibly other assumptions
 Fins
• To enhance heat transfer between solid and fluid phases
• Conduction along the fin, convective heat transport outside the fin
• Temperature distribution inside the fin in crucial role.
• In many circumstances T=T(x) i.e. 1d temperature distribution
• It enables formulation of 1d energy balance i.e. heat equation for a fin
• Such 1d conduction assumption in fin context is called fin theory.

Basic fin types Typical finned-tube heat

exchangers
 Energy balance for a single fin
Assumptions: 1d + steady state
Energy balance: q out −q in −dq c =0

Perimeter = 2L + 2d

Result when dx→ 0:

2
heat equation d T hP
2 = (T −T ∞ )
with energy dx k Ac
loss term

Cross-sectional area = Ld

2 hP
m =
k Ac
Energy balance for a single fin in steady state
(1/2)
Energy arrives into small volume by 1) conduction
dT ( x−dx /2 , t )
q in =kA c
dx

Energy leaves out small volume by 2) conduction

dT ( x +dx / 2 , t )
q out =kA c
dx

(New in comparison to 1st week topics)

Energy leaves from outer surface by 3) convection
dq c =hdA s (T −T ∞ )
 Energy balance for a single fin in steady state
(2/2)
“J/s” on both side:

∂T ( x +Δ x /2 , t ) ∂ T ( x−Δ x / 2 , t )
ρ c p ab Δ x Δ T / Δ t =abk −abk − Ah(T −T ∞ )
∂x ∂x

2
∂T ∂ T Ph
ρcp =k 2− (T −T ∞ )
∂t ∂x ab

2
∂ T Ph
2 = (T −T ∞)
∂x abk
 How would fin material influence temperature profile?
Incropera Ex. 3.9: for very long rods (e.g. H=50-250mm)

H →∞
T ( x)≈T ∞ +(T b−T ∞ ) e−mx

Compare
k copper =398 W / mK
k 224 Al =180 W / mK
k steel =14 W / mK

What kind of temperature profiles result ?

Example temperature profiles assuming
Tb=328K, Tinf=293K
Incropera: 1d Temperature Distribution and Heat
Loss Along a Fin
 And where did the previous table come from?
d2θ
dx 2
−m 2 θ=0 → General solution
θ=C 1 e mx +C 2 e−mx

d2T hP
2
= (T −T ∞ )
dx k A c

hP
θ=(T −T ∞ ) m 2=
k Ac
d2θ
2
−m 2 θ=0
dx

+ boundary conditions
θ( x =0)=T b −T ∞ =θb

dT
hA c (T ( L)−T ∞ )=−kA c ( )
dx x= L → Analytical temperature distribution can be solved
θ( x ) cosh ( m( L−x ))+( h/ mk )sinh (m ( L−x ))
=
θb cosh (mL)+( h/ mk )sinh (mL)
 Fin effectiveness
Fin effectiveness: (heat transfer rate with fin) / (heat transfer rate without fin)

qf
ϵf = Usage of fins typically justified if > 2
hA c ,b θ b
Common assumption (not reality but useful)
→ assume that h is unaffected by 1) spatial position, and 2) presence of fins

Fin effectiveness (for infinitely long fin) reads (Table 3.4):

1 /2
kP
ϵf =
( )
hA c
Heat transfer enhancement if:
→ perimeter to the area increased → prefer thin, closely spaced fins but not too
close to not impede flow (e.g. laminarization/stagnation) between fins
→if k/h is “small” then more need for fins (e.g. natural convection)
→if fluid is gas then more need for fins

Example: automobile radiator (fins on the air flow side, hot water in the inside)
During the last 25 minutes we talk about thermal
resistance and 2d heat equation
 How could temperature distribution look like
in a heated 2d metal fin (assume here: no convection) ?
Model 1: Model 2:
- Cool sides and tip - Insulated sides
- Hot base - Cool tip and hot base
→ T=T(x,y,t) (2d) → T=T(y,t) (1d)
Heat equation in 2d i.e. energy conservation

General form of heat equation Terms opened in 2d

∂T ∂T ∂ ∂ T ∂ ∂T
=∇⋅α ∇ T = α + α
∂t ∂t ∂x ∂ x ∂ y ∂ y
 Heat equation in 2d i.e. energy conservation

General form of heat equation Terms opened in 2d

∂T ∂T ∂ ∂ T ∂ ∂T
=∇⋅α ∇ T = α + α
∂t ∂t ∂x ∂ x ∂ y ∂ y

∂2 T ∂2 T
2
< 0 and 2
<0
∂x ∂y
Energy/temperature escapes
away from this point.

Energy is
conducted from ∂2 T ∂2 T
hot to cold. 2
> 0 and 2
>0
∂x ∂y
Energy/temperature arrives
to this point.
 Question: in the 2d heat conduction
below (metal plate) what does Fourier tell about
max/min heat flux ?
Early time temperature Late time temperature
Zoom to plate
upper left corner
>> T(1:20,1:5)

ans =
269.8795 276.1205 280.3794 283.3261 285.3968
316.1205 309.8795 305.6206 302.6739 300.6032
314.0774 311.9226 309.8685 308.0042 306.3579
313.6455 312.3545 311.0793 309.8467 308.6758
313.4649 312.5351 311.6104 310.7006 309.8141
313.3662 312.6338 311.9038 311.1806 310.4685
313.3043 312.6957 312.0882 311.4845 310.8868
313.2623 312.7377 312.2138 311.6922 311.1742
313.2321 312.7679 312.3041 311.8418 311.3820
313.2095 312.7905 312.3717 311.9540 311.5380
313.1921 312.8079 312.4239 312.0406 311.6586
313.1784 312.8216 312.4651 312.1091 311.7541
313.1673 312.8327 312.4984 312.1645 311.8312
313.1581 312.8419 312.5257 312.2100 311.8948
313.1505 312.8495 312.5486 312.2481 311.9479
313.1440 312.8560 312.5681 312.2805 311.9932
313.1384 312.8616 312.5850 312.3086 312.0325
313.1334 312.8666 312.5999 312.3334 312.0672
313.1289 312.8711 312.6133 312.3557 312.0984
313.1248 312.8752 312.6256 312.3762 312.1270
 >> T(1:20,1:5)
Ghost cell row of the top side BC
ans = Note:
The corner 269.8795 276.1205 280.3794 283.3261 285.3968
cell is 316.1205 309.8795 305.6206 302.6739 300.6032 1) the two
redundant 314.0774 311.9226 309.8685 308.0042 306.3579 values are
different →
313.6455 312.3545 311.0793 309.8467 308.6758 not insulated
313.4649 312.5351 311.6104 310.7006 309.8141 Boundary
313.3662 312.6338 311.9038 311.1806 310.4685
313.3043 312.6957 312.0882 311.4845 310.8868 2) the
Ghost cell average
313.2623 312.7377 312.2138 311.6922 311.1742
column of of the two
313.2321 312.7679 312.3041 311.8418 311.3820 values is
the left
313.2095 312.7905 312.3717 311.9540 311.5380 const.
side BC
313.1921 312.8079 312.4239 312.0406 311.6586 → fixed Ttop
313.1784 312.8216 312.4651 312.1091 311.7541
313.1673 312.8327 312.4984 312.1645 311.8312
313.1581 312.8419 312.5257 312.2100 311.8948
313.1505 312.8495 312.5486 312.2481 311.9479
313.1440 312.8560 312.5681 312.2805 311.9932
313.1384 312.8616 312.5850 312.3086 312.0325
313.1334 312.8666 312.5999 312.3334 312.0672
313.1289 312.8711 312.6133 312.3557 312.0984
313.1248 312.8752 312.6256 312.3762 312.1270
 Temperature profile across a single layered
wall with convection effects on inner and outsides
●
Consider a wall surrounded by warm and cold air
●
Quite easily available: temperatures A and B, heat transfer
coefficients, and thermal conductivities
●
We “want” to know many things, e.g.
1) How many J/s is going through the wall ?
2) What is the temperature in the middle of the wall ?
Convective heat transfer coefficient

T A , hA T B , hB
T ( x) L
T A , hA

T1 T2
 Derivation of heat flux through the wall

Steady state: q 1=q 2=q3 =q

q1 q3
q /(h A A)=(T A −T 1) (1)

q /(k / L)=(T 1 −T 2) (2)

A=area of q /(h B A)=(T 2−T B ) (3)

cross-section
q2 Sum: (1)+(3) and note that T1-T2 appears i.e. (2)

T A−T B +T 2−T 1=q /(h A A )+q /(h B A)

T A −T B
q=
1/( kA / L)+1/(h A A)+1/( hB A) Note: Generalizes to
multilayered materials
with various thicknesses!
Thermal resistance – composite wall
with multiple material layers
Formulate heat flux by Newton’s cooling law:
q=UA Δ T , with Δ T =T A −T B
Overall heat transfer
coefficient

A ΔT 1
q= U=
Σ i 1/(k i / Li )+1/(h A )+ 1/(h B ) R tot A
Thermal resistance

Some benefits of thermal resistance concept:

→ design of thermal insulation (buildings, clothes, combustion)
→ allows to maximize or minimize heat flux
→ allows designs to avoid hot pools of temperature from forming
→ allows to design temperature profiles (e.g. avoid condensation)
Lecture 2.2 Numerical approach: a Matlab
solver for the 2d heat equation and setting up
different boundary condition

ILO 2: Student can apply Fourier’s law and Newton’s law in fin
theory and thermal resistance context. Further, the student can
analyse 2d heat transfer data in Matlab and formulate an energy
balance for 2d system.
 2d Heat Equation

General form of heat equation Terms opened in 2d

∂T ∂T ∂ ∂T ∂ ∂ T
=∇⋅α ∇ T = α + α
∂t ∂t ∂x ∂x ∂ y ∂ y
Time derivative in cell (i, j) at timestep n Second space derivatives at cell (i,j)
n+1 n
∂T
n
T −T 2
∂ T
n
T ni +1, j−2 T ni , j +T ni−1, j
( )
∂ t i, j
≈ i, j
Δt
i, j
x:
( )
∂ x2 i, j
≈
Δ x2
2
∂ T
n
T ni , j +1−2 T ni , j +T ni , j−1
y:
( )
∂ y2 i, j
≈
Δ y2
2d Example Solutions of Temperature

Bottom and top walls are zero gradient Bottom wall is zero temperature gradient
keeping solution 1d i.e. thermally insulated

Conducting
Top & Bottom zero gradient

Insulated
pocket (k=0) Non-
conducting

Zoom: note
that isolines
meet the surface
perpendicular for zero-gradient
 Discretization of a 2d Domain

x=0 x=Lx

Ghost cells
where
y=Ly
boundary
conditions n Ghost cells
given. T i , j +1 where
boundary
conditions
T nghost T ni−1, j T ni , j T ni+1, j given

T ni , j −1
y=0

Δy
Cell center
Δx
 >> T(1:20,1:5)
Ghost cell row of the top side BC
ans = Note:
The corner 269.8795 276.1205 280.3794 283.3261 285.3968
cell is 316.1205 309.8795 305.6206 302.6739 300.6032 1) the two
redundant 314.0774 311.9226 309.8685 308.0042 306.3579 values are
different →
313.6455 312.3545 311.0793 309.8467 308.6758 not insulated
313.4649 312.5351 311.6104 310.7006 309.8141 Boundary
313.3662 312.6338 311.9038 311.1806 310.4685
313.3043 312.6957 312.0882 311.4845 310.8868 2) the
Ghost cell average
313.2623 312.7377 312.2138 311.6922 311.1742
column of of the two
313.2321 312.7679 312.3041 311.8418 311.3820 values is
the left
313.2095 312.7905 312.3717 311.9540 311.5380 const.
side BC
313.1921 312.8079 312.4239 312.0406 311.6586 → fixed Ttop
313.1784 312.8216 312.4651 312.1091 311.7541
313.1673 312.8327 312.4984 312.1645 311.8312
313.1581 312.8419 312.5257 312.2100 311.8948
313.1505 312.8495 312.5486 312.2481 311.9479
313.1440 312.8560 312.5681 312.2805 311.9932
313.1384 312.8616 312.5850 312.3086 312.0325
313.1334 312.8666 312.5999 312.3334 312.0672
313.1289 312.8711 312.6133 312.3557 312.0984
313.1248 312.8752 312.6256 312.3762 312.1270
 Code structure for a 2d heat equation solver
% code for solving 2d heat eqn % the code continues ...

clear * % clears memory T = 293*ones(Ny,Nx);

close all % closes figures CFLmax = 0.1;
nu = (1.6e-5);
Tleft = 293; Tright = 293; % nu = diffusivity
Ttop = 330; Tbot = 330; % [nu] = m^2/s

Nx = 102; Ny = 52; dt = 0.5e-5; % timestep

Lx = 0.05; Ly = 0.025;

% arrays for computing T_xx T_yy SetTBCs;

inx = 2:(Nx-1); iny = 2:(Ny-1);
north = iny+1; for(k=1:1000000)
east = inx+1;
south = iny-1; setTBCs;
west = inx-1; solveTemperature;
visualizeResults;
% (Nx-2) by (Ny-2) internal cells
dx = Lx/(Nx-2); dy = Ly/(Ny-2); end

x=linspace(0,Lx,Nx-2);
y=linspace(0,Ly,Ny-2); Recommendation: create folder
Week2 and type the text in the
[X,Y] = meshgrid(x,y); boxes to a new script e.g. P1.m
 The other (incomplete) .m files could
look like as follows
% setTBCs.m % solveTemperature.m
% Exact form depends on BC’s ComputedT;
T(1,:) = 2*Ttop - T(2,:);
T(iny,inx) = T(iny,inx) + dT;
T(Ny,:) = 2*Tbot - T(Ny-1,:); setTBCs;

T(iny,1)= 2*Tleft – T(iny,2);

T(iny,Nx)= T(iny,Nx-1);

% computedT.m
% Incomplete, please add T_xx

dTdx = ( T(iny,east)-T(iny,west))/(2*dx); % not needed yet

d2Tdy2 = ( T(north,inx)-2*T(iny,inx)+T(south,inx))/(dy^2);
% add calculation of d2Tdx2

dT = dt*( nu*d2Tdx2 + nu*d2Tdy2);

 Update Formula by Explicit Euler Method
αΔ t
CFL= 2
Explicit Euler timestepping for 2d heat equation: Δx
n n n n n n
n+1 n T i +1, j −2 T i , j +T i −1, j T i , j+1 −2 T i , j +T i , j−1
T i,j =T + Δ t α
i, j 2
+Δ t α 2
Δx Δy
Which is equal to the “delta” form:
n n n n n n
n T i +1, j −2 T +T
i, j i−1, j T i , j+ 1 −2 T + T
i, j i , j −1
Δ T =Δ t α
i, j 2
+Δ t α 2
Δx Δy
Where:

Δ T ni , j=T ni ,+1j −T ni , j
 Summary of 2d Scheme

(Known and
1) Set boundary conditions (BC's) to the ghost cells T ni , j hence BC
using T from step n. update possible)

2) Update new temperature at timestep n+1 in the T

n+1 n
=T +Δ t α
n n n
T i +1, j −2 T i , j +T i −1, j
+Δ t α
n n n
T i , j+1 −2 T i , j +T i , j−1
internal cells i,j i, j
Δ x2 Δ y2

T n+1
i,j
3) Update time according to t = t + dt
t n+1 =t n +Δ t

4) Go back to 1)
Extra material 1
Appendix I: Energy balance for a single fin with
non-uniform cross-sectional area

Energy arrives into small volume by 1) conduction

dT
q x =−kA c
dx
Energy leaves from small volume by 2) conduction
dq x
q x+ dx =q x + dx
dx
(New in comparison to 1st week topics)
Energy leaves from surface by 3) convection
dq c =hdA s (T −T ∞ )

Energy balance: q x =q x+dx +dq c

Energy balance (heat equation with source terms):

d 2 T 1 dA c dT 1 h dA s
2
+ − (T −T ∞ )=0
dx A c dx dx A c k dx
Extra material 2: Classroom demo from 2018
 Classroom demo heat exchanger
Acknowledgements:
K.Saari and M.Ahlgren
Fan

12 fins Temperature
Assume 2W heat probe B
escapes from
each B
Temperature
probe A
A
Heating power
P = 24 W
in the base
Assume uniform
heating
 Schematic of a fin which is treated
as 1d object in fin theory

In the class room example:

H=0.10m L=0.05m d=1mm

Base plate
 Estimate order of magnitude for the time
it takes for any heat to reach the fin tip ?
Data for Aluminum + surface temperature measurements

P = 24 W 12 fins → 2W/fin Cp = 900 kJ/kgK

ΔTAB ≈ 2 K k = 237 W/mK α = 10-5 m2/s
A = 0.001m ∙ 0.05 m = 0.00005 m2 δAB ≈ 0.04 m
H = 0.1m
Hint: diffusion time
τ = H2/(4π2α)
 Estimate order of magnitude for the time
it takes for any heat to reach the fin tip ?
Data for Aluminum + surface temperature measurements

P = 24 W 12 fins → 2W/fin Cp = 900 kJ/kgK

ΔTAB ≈ 2 K k = 237 W/mK α = 10-5 m2/s
A = 0.001m ∙ 0.05 m = 0.00005 m2 δAB ≈ 0.04 m
H = 0.1m

Use the diffusion time concept from Week 1:

τ = H2/(4π2α) ~ 0.5 min

 Estimate power (W) conducted through single
fin based on surface temperature measurements?
Data for Aluminum + surface temperature measurements

P = 24 W 12 fins → 2W/fin Cp = 900 kJ/kgK

ΔTAB ≈ 2 K k = 237 W/mK α = 10-5 m2/s
A = 0.001m ∙ 0.05 m = 0.00005 m2 δAB ≈ 0.04 m
H = 0.1m

Hint: Fourier’s law: q = A kΔTAB/δAB

 Estimate power (W) conducted through single
fin based on surface temperature measurements?
Data for Aluminum + surface temperature measurements

P = 24 W 12 fins → 2W/fin Cp = 900 kJ/kgK

ΔTAB ≈ 2 K k = 237 W/mK α = 10-5 m2/s
A = 0.001m ∙ 0.05 m = 0.00005 m2 δAB ≈ 0.04 m
H = 0.1m

Use Fourier’s law from Week 1 + exp. data

Fourier’s law offers a way to estimate heat rate in Watts

along a fin:
q’ = A kΔTAB/δAB = 0.00005m2 ∙ 237 W/mK∙ 2 K/0.04m
≈ 0.59 W < 2W
→ WHY SMALLER POWER ?