FVM- Transient Heat Transfer

G. Kumaresan, Ph.D.
Institute for Energy Studies
Anna University, Chennai

1
 Transient Conduction
• If the surface temperature of a solid body is suddenly altered, the temperature within
the body begins to change over time.

• Variation of temperature both with position and time makes determination of

temperature distribution under transient condition more complicated.

Types of Transient Problem

1. Systems with negligible internal resistance

2. Systems with negligible surface resistance
3. Systems with finite internal and surface resistance
4. Semi-Infinite solid
 For Unsteady Flows:
One Dimension unsteady (transient) heat conduction ( 1 D transient diffusion)

where ρ density
c cp Specific heat
Unsteady term Diffusion term Source term k Thermal conductivity

The Finite volume integration of equation (1) over a control volume (cv) must be augmented with a
further integration over a finite time step Δt.

Consider the one dimensional control volume,

xWP xPE

3
t t
T
t t
   T   t t

 cv t dVdt 
 c     K  dV  dt     SdV  dt   2 
t t  cv x  x   t cv

 t t T  t t
 T   T   t t

cv  t  c t dt dV  t  KA x e   KA x w dt  t S Vdt   3

L.H.S of equation (3) where S  avg source, strength of source term
 t t T 
cv  t  c t dt dV= c TP  TP  V   4 
V  A.x
x  Width of the control volume
A  Face area of the control volume
V  Volume
S  Average source strength

In equation (4), it is assumed that the temperature at a node is assumed to prevail over the whole control
volume . Subscript ‘o’ refer to temperature at time ‘t’, temperatures at time level t+Δt not superscripted
i.e.

TP temperature at time t at the grid point ‘P’

TP temperature at time t+Δt at the grid point ‘P’

4
 t t

  TE  TP   TP  TW 
t t
 c TP  TP  V    e
AK   AK w    dt   S Vdt
t 
   x PE   xWP  t
t t
IT   TP dt   TP  1    TP t
t

Explicit Semi Implicit Pure Implicit

(Crank
Nicholson)
θ 0 ½ 1
IT TP Δt ½ TP  TP  t 𝑇𝑃 Δ𝑡

For semi-implicit θ=1/2

Where θ is the weighting parameters varying from 0 to 1 [½ 𝑇𝑃 +(1- ½)TP ]Δt
t t [½ 𝑇𝑃 +½ TP ] Δt
t
S Vdt  S V t ½[𝑇𝑃 + TP ] Δt

 SAxt
5
 x  T T T T   T  TP T T 
 c TP  TP  A    AK e E P  AK w P W   1     AK e E  AK w P W   S x
t  xPE xWP   xPE xWP 
x x KT KT KT KT KT KT KT KT
 cTP   cTP   e E  e P  w P  w W  e E  e P  e E  e P
t t xPE xPE xWP xWP xPE xPE xPE xPE
K wTP K wTW KT KT
    w P   w W  S x
xWP xWP xWP xWP
 x  K e  K w  Ke Kw
  c    P
T  
 T  1    T 
E   TW  1    TW 
 t xPE xPE  xPE xWP
E

 x K e  K e K w  K w 
 c      TP  S x
 t x PE x PE xWP xWP 

 x  Ke K w  Ke Kw
  c      PT  
 T  1    T 
E    TW  1    TW 
 t  xPE xPE   xPE xWP
E

 x K K 
   c  1    e  1    w  TP  S x
 t xPE xWP 

6
 In General,
𝑎𝑃 𝑇𝑃 = 𝑎𝑊 𝜃𝑇𝑊 + 1 − 𝜃 𝑇𝑊 ∘ + 𝑎𝐸 𝜃𝑇𝐸 + 1 − 𝜃 𝑇𝐸 ∘ + 𝑎𝑃 ∘ − 1 − 𝜃 𝑎𝑊 − 1 − 𝜃 𝑎𝐸 𝑇𝑃 ∘ + 𝑆ҧ ∆𝑥

aw aE b
𝐾𝑤 𝐾𝑒
S x
𝜕𝑥𝑊𝑃 𝜕𝑥𝑃𝐸
x
where a P    aw  a E    c
t
x
aP    aw  aE   a op where,a op   c
t
Explicit Scheme: When θ =0

Explicit Scheme:
when  0
x Ke Kw  x Ke Kw 
c TP  TE  TW   c    TP  S x
t xPE xWP  t xPE xWP 
aPTP  aW TW  a E TE 
 aP  aW  a E 
 TP b
aPTP  aW TW  a E TE 
 aP  ( aW  a E  S P ) 
 TP  Su
where b  Su  S PTP
x
aP  aP
o
 c & b  Su  S PTP 7
t
 𝑎𝑊 𝑎𝐸
𝐾𝑤 𝐾𝑒
𝜕𝑥𝑊𝑃 𝜕𝑥𝑃𝐸

The coefficient TP may be viewed as the neighbour coefficient connecting the values
at the old time level to those at the new time level.
For the coefficient to be positive we must have aP  aW  aE >0.
For constant ‘k’ and uniform grid spacing, xPE = xWP = Δx,
This condition may be written as
x 2K
c 
t x
x 2
c  t
2K
x 2
t   c
2K
t 2 K
1 where C is called the Courant number. This
x  c
2
condition is referred to as the Courant
C 1 Friedrichs-Lewy or CFL condition. 8
Some points to note:
1. CFD codes will allow you to set the Courant number (which is also referred
to as the CFL number) when using time-stepping. Taking larger time-steps
leads to faster convergence to the steady state, so it is advantageous to set
the Courant number as large as possible within the limits of stability.
2. You may find that a lower Courant number is required during start-up when
changes in the solution are highly non-linear but it can be increased as the
solution progresses. 9
Problems:
1) A thin plate is initially at a uniform temperature of 200°C. At a certain time t the
temperature of the east side of the plate is suddenly reduced to 0°C. The other surface
is insulated. Use the explicit finite volume method in conjugate with a suitable time
step size, calculate the transient temperature of the plate at time (i) t=40 sec (ii) t=80
sec and (ii) t=120 sec . Consider plate thickness L= 2cm, K= 10W/m-k, ρc= 10 x 106
J/m3/K.

Solution
The non dimensional transient heat conduction equation

T  2T
c K
t x 2
The initial condition is
T= 200 and t =0
Boundary Conditions are
(i) 𝜕T/ 𝜕x = 0 at x= 0, t > 0
(ii) T =0 at x = L, t > 0
10
 Discretization
x Ke Kw x Ke Kw
c TP  TE  TW   c TP  TP  TP  S x
t xPE xWP t xPE xWP
x  T  TP   TP  TW 
 c TP  TP 
 Ke  E   K w   S x
t  xPE   xWP 
The above general equation for any node explicit at  = 0
For node (1) (no western node)

Δx
T 1 2
0
x
e
(TP  TP )x TE  TP x

c  Ke 2

t xPE
(TP  TP )x TE  TP
c K
t xPE
11
 x x T T
c TP   c TP  K E  K P
t t x x
c K  x K 
xTP  TE    c   TP
t x  t x 
aPTP  aW TW  aE TE  
 aP  aW  a E 
 TP  b
c
where aP  aP  x
t
K
aE  b  Su  S PTP
x
For node 1
𝑎𝑊 𝑎𝐸 𝑆𝑢
0 𝐾 0
𝛥𝑥

For node 5 (no east node)

4 5

w x
2 12
 c  TP  TP x
t
KB

x
 TB  TP  
Kw
xWP
TP  TW 
2
x
 c  TP  TP   TB  TP   TP  TW 
2K B Kw

t x x
x
 c  TP  TP   TB  TP   TP  TW 
2K K

t x x
x x 2K 2K K K
c TP   c TP  TB  TP  TP  TW
t t x x x x
x K  x K  2K 2K
c TP  TW    c  T  T  TP
t x t x  x x
P B
 
0
aPTP  aW TW  aE TE   a
 P  aW  a 
E  TP 
2K
x
 B  TP
T 
x
where aP  aP   c
t
aW 
K
x
Su 
2K
x
TB  TP 
For node 5

𝑎𝑊 𝑎𝐸 𝑆𝑢
𝐾 0 2𝐾 𝑇𝐵 0− 𝑇𝑃 𝑜
𝛥𝑥 𝛥𝑥

13
 For interior nodes 2,3 and 4
TP  TP K e  TE  TP   K T  TW   S x
c x 
w P

t xPE xWP
TP  TP K  TE  TP  K TP  TW 
c x  
t x x
x x T T TP T
c TP   c TP  K E  K P  K E  K W
t t x x x x
x K K  x K K 
c TP  TW  TE    c    TP
t x x  t x x 
aPTP  aW TW  aE TE  
 aP  aW  aE 
 TP  0

For 2,3 and 4 nodes

𝑎𝑊 𝑎𝐸 𝑆𝑢
𝐾 𝐾 0
𝛥𝑥 𝛥𝑥

14
For all nodes

Node 𝑎𝑊 𝑎𝐸 𝑆𝑢
1 0 𝐾 0
𝛥𝑥

2,3 and 4 𝐾 𝐾 0
𝛥𝑥 𝛥𝑥
0
5 𝐾 0 2𝐾 𝑇𝐵 − 𝑇𝑃 𝑜
𝛥𝑥 𝛥𝑥

15
16
17
18
Comparison of Numerical results with analytical results for time step 2sec

19
Comparison of Numerical results with analytical results for different time step values

20
 Standard Form of discretized Gov. Eqn for unsteady diffusion problem

Explicit Semi Implicit (Crank Pure Implicit

Nicholson)
θ 0 ½ 1

Explicit scheme

21
Standard Form of discretized Gov. Eqn for unsteady diffusion problem

Crank-Nicolson scheme

b  Su  12 S PTP  12 S PTP0

22
Standard Form of discretized Gov. Eqn for unsteady diffusion problem

Implicit scheme

b  Su  S PTP

23
Problems:
1) A thin plate is initially at a uniform temperature of 200°C. At a certain time t=0 the temperature
of the east side of the plate is suddenly reduced to 0°C. The other surface is insulated. Use the
implicit finite volume method in conjugate with a suitable time step size, calculate the transient
temperature of the plate at time (i) t=40 sec (ii) t=80 sec and (ii) t=120 sec . Consider plate
thickness L= 2cm, K= 10W/m-k, ρc= 10*10^6 J/m3/K.

Solution
The non dimensional transient heat conduction equation

T  2T
c K
t x 2
The initial conditions are
T= 200 and t =0
Boundary Conditions are
(i) 𝜕T/ 𝜕x = 0 at x= 0, t > 0
(ii) T =0 at x = L, t > 0
24
Implicit problem cont…

Explicit problem

25