Transient Heat Transfer Lumped Analysis-I

• Let the body temperature be denoted

• Having gotten a feel for the steady state heat transfer, let by T T∞, h
us get a feel for the transient behavior of heat transfer
• The body interacts with the T
• To begin with, let us assume that spatial variation of surroundings with the temperature at
temperature is negligible and the temperature of the body T∞ and heat transfer coefficient h
as a whole changes with time
• First law implies that E& = Q& − W
&
• This implies that the thermal conductivity is very high.
d
• This type of problem is called lumped analysis ⇒ (mcT ) = −hA(T − T∞ )
dt
• We shall evolve a criterion for the validity of this • For constant specific heat
assumption as we go along dT
⇒ mc = − hA (T − T∞ )
• Let us look at the steps in analysis dt

Lumped Analysis-II Lumped Analysis-III

hA hA h A
• Defining θ = T - T∞, we can write − t −
Vρc
t −
ρc V
t
• In general θ = θ0 e mc = θ0e = θ 0e
dθ dθ hA
⇒ mc = −hAθ ⇒ =− θ Let θ = θo at t = 0 • If the object is a cylinder of Radius R and Length L,
dt dt mc
we have
• The solution for the above equation is
A 2πRL + 2πR 2 2(L + R )
hA ⇒ = =
−
mc
t V πR 2 L RL
θ = θ 0e
h 2( L + R )
− t
• If the object is sphere, we have ρc LR
θ = θ0 e
2
4 hA h 4πR h 3
m= πR 3ρ; A = 4πR 2 ⇒ = = • If the cylinder has L>>R , then
3 mc c 4 πR 3ρ ρc R
h 3 3 −
h 2
t
− t
ρc R ρc R
θ = θ0 e θ = θ0 e

Lumped Analysis-III Lumped Analysis-IV

• Thus, the temperature variation is a function of two
• This concept can now be extended to any geometry
non dimensional parameters Biot number and Fourier
• To generalize the result, it is customary to introduce number.
the characteristic length L • We will appreciate these parameters, as we go into
• The logic suggests that V/A is the most obvious more complex cases
choice • We can give a physical interpretation for the Biot
h hL k hL αt
− t − t − number as follows
ρcL k ρcL2 k L2
⇒ θ = θ0e = θ0 e = θ 0e hL L hA conduction Re sis tan ce
Bi = = =
• In the above expression, we have introduced the K KA 1 convection Re sis tan ce
property called thermal diffusivity = k/(ρc) • When Bi is very small, it implies that conduction
• The non-dimensional parameter hL/k and αt/L2 are resistance is very small and hence lumped analysis valid
called the Biot Number and Fourier No respectively • The criterion used is Bi < 0.1

1
 Transients with spatial effects 1-D Transient in a Plate-I
• The governing equation
• If Bi > 0.1 spatial effects become important and so
more complications are involved 1 ∂T ∂ 2 T
= 0 ≤ x ≤ L; 0 ≤ t h, T∞ h, T∞
α ∂t ∂x 2
• Exact analytical solutions can be obtained using x

separation of variables similar to 2-D steady state • Boundary conditions

0 L
analysis T(0,x) = Ti ; [∂T/∂x](t,0) = 0; [-k∂T/∂x](t,L) = h(T(t,L)- T∞)
• Let us look at 1-D transient analysis in a slab
geometry with no heat generation • Let us define θ = T − T∞
• The governing equation for this case is 1 ∂θ ∂ 2 θ
⇒ = 0 ≤ x ≤ L; 0 ≤ t
α ∂t ∂x 2
∂ (T )  ∂ 2T ∂ 2 T ∂ 2 T  1 ∂T ∂ 2 T
ρc = k 2 + 2 + 2  + q ′′′ ⇒ = [∂θ/∂x](t,0) = 0; -k[∂θ/∂x](t,L) =hθ(t,L);
∂t  ∂x ∂y ∂z  α ∂t ∂x 2

θ(0,x) = θ0

1-D Transient in a Plate-II 1-D Transient in a Plate-II

• The solution for θ(t,x) is assumed of the form: • The solution for X and T are;
2
X = C1 cos(λx ) + C 2 sin(λx ) T = C 3e − αλ t
θ(t,x) = X(x)T(t)
• Substituting this in the governing equation, we get
( )
⇒ θ = XT = (C1 cos(λx ) + C 2 sin(λx ) ) C 3e − αλ t
2

2 2
= (C cos( λx ) + C sin(λx ) )(e
1 2 ) − αλ2 t C3 is absorbed
in C1 and C2
1 dT ( t ) d X( x) 1 1 dT 1 d X
⇒ X(x ) = T(t ) =0 ⇒ = • The BC at x = 0; ∂θ/ ∂x = 0, requires a symmetric
α dt dx 2 α T dt X dx 2
solution and hence C2 = 0
• Since LHS is only a function of t and RHS is only
a function of x and yet they be equal, would ( ∂θ
)
⇒ θ = (C1 cos(λx ) ) e − αλ t ⇒ ∂x = (C1λ (− sin(λx ) ) e
2 − αλ2 t
( )
require that both sides be equal to a constant
2 2
BC at x = L ⇒ (− kC1 ( −λ ) sin(λL) ) e (
− αλ t 2
)
= h (C1 cos(λL) ) e − αλ ( 2
t
)
1 1 dT 1 d X
⇒ = = −λ2 ⇒ d X + λ2 X = 0; dT − αλ2 T = 0 h
⇒ (λ sin(λL) ) = (cos(λL) ) ⇒ cot( λL) =
kλ λ L
=
α T dt X dx 2 dx 2 dt k h Bi

1-D Transient in a Plate-III 1-D Transient in a Plate-IV

• The solution for the eigen values can be • The constants have to be determined from
graphically interpreted as shown initial condition
• Thus we will have infinite • This will be again done by orthogonal functions
values of eigen values, but Cot(λL) i • It turns out that
λ L/B
these have to be numerically L
determined. Your book λ1 L λ2 L λ3 L λL ∫ cos(λ n x ) cos(λ m x )dx = 0 for n ≠ m Messy to prove
0
summarizes the first four Will be given as home work
values as a function of Bi in
L
Appendix B3 2 λ n L + sin(λ n L) cos(λ n L)
∫ cos (λ n x )dx = 2λ n
• Thus the solution for θ can be written as 0
∞ 2 L
⇒θ= ∑ C n e −αλ n t cos( λ n x ) Still Cn needs to sin(λ n L)
n =1 be determined
∫ cos(λ n x )dx = λn
0

2
 1-D Transient in a Plate-V 1-D Transient in a Plate-VI
• The initial condition is θ(0,x) = θ0
∞ • Thus the solution for θ can be written as
⇒ θ0 = ∑ C n cos(λ n x ) ∞
n =1 θ sin(λ n L) 2
⇒ = 2∑ e − αλ n t cos( λ n x )
L L ∞ θ0 n =1 λ n L + sin( λ n L) cos( λ n L )
∫ θ 0 cos(λ m x )dx = ∫ ∑ C n cos(λ m x ) cos(λ n x )dx ∞
0 0 n =1 θ sin(λ n L) 2 x
⇒ = 2∑ e − αλ n t cos( λ n L )
• Note that in the RHS, only term n=m will survive θ0 n =1 λ n L + sin( λ n L) cos( λ n L ) L
L L
2 θ  αt x 
∫ θ 0 cos(λ n x )dx = C n ∫ cos (λ n x )dx ⇒ = f  λ n L, 2 , 
0 0 θ0  L L
sin(λ n L) λ L + sin(λ n L) cos( λ n L) • But, λnL is determined by Bi
θ0 = Cn n
λn 2λ n θ  αt x   x
2 sin( λ n L) ∴ = f  Bi, 2 ,  = f  Bi, Fo, 
∴ C n = θ0 θ0  L L  L
λ n L + sin(λ n L) cos( λ n L)

Heisler Charts-I Heisler Charts-II

• The coordinate of the mid-plane is x =0
• The solution for θ was shown to be
θ(0)
= C1e − αλ1 t = C1e −(λ1L ) Fo = f (Fo, λ1L) = f (Fo, Bi )
2 2

∞
⇒
θ sin(λ n L) 2
θ0
⇒ = 2∑ e − αλ n t cos( λ n x )
θ0 n =1 λ n L + sin( λ n L) cos( λ n L ) • Thus, the spatial profile at any time can be written as
• In functional form θ
⇒ = cos(λ1 x ) = f (λ1L, x / L) = f ( Bi, x / L)
θ  αt x   x θ(0)
= f  Bi, 2 ,  = f  Bi, Fo, 
θ0  L L  L • Heisler presented these results in the form of two
• Heisler showed that when Fo > 0.2, just one term is graphs, called Heisler Charts
adequate to describe the solution • For Cartesian system, it is shown in the next slide.
θ 2
sin(λ1L) Similar curves for the cylindrical and spherical one
⇒ = C1e −αλ1 t cos(λ1 x ) Where C1 = 2
θ0 λ1L + sin(λ1L) cos(λ1L) dimensional cases are given in Appendix D of your
book

Heisler Charts-III Criterion for Lumped Analysis

• If the variation of temperature within the slab is less
than 5%, we can call it lumped
θ( x = L) ⇒ λ1L = 0.318 radian
⇒ = cos(λ1L) = 0.95
θ(0)
λ 1L
⇒ Bi = = 0.1
cot(λ1L)
• Thus, Bi should be < 0.1 for lumped analysis to be
valid. This can also be viewed in the chart

3
 2
θ = θ 0 C1e − αλ1 t cos(λ1x )
Energy Storage-I Energy Storage-II
• As energy storage is one of the main application of
∆E  sin(λ1L) 
transient, it is useful to get a method to estimate the =  θ(0) − 1 = f (λ1L, Fo) = f (Bi, Fo)
∆E 0  λ 1L 
total energy stored
• From thermodynamics
L
1L
∆E = 2 ∫ ρAc(T − T0 )dx = 2ρAcL ∫ (θ − θ0 )dx
0 L0
1L − αλ 2 t
= 2ρAcLθ 0
L0
∫ (C1e 1 cos(λ1 x ) − 1)dx
L
∆E 1 2 sin(λ x ) 
=  C1e − αλ1 t 1
− x 
2ρAcLθ 0 L  λ1 0
∆E  2 sin( λ L )   sin(λ1L) 
=  C1e −αλ1 t 1
− 1 =  θ(0) − 1
∆E 0  λ1 L   λ 1L 

1-D Transient in a Semi-Infinite Plate-I 1-D Transient in a Semi-Infinite Plate-II

• Another case of interest in transient heat transfer is • Mathematically, problems that have boundary at
what is called heat transfer in a Semi-infinite plate infinity are often solved by a method called similarity
• It is has many useful applications. In fact, all solution
transient problems start as semi-infinite wall • In this method, a new variable, called similarity
variable is introduced
• The governing equation
• This variable is chosen such that T becomes a
1 ∂T ∂ 2 T
= 0 ≤ x ≤ ∞; 0 ≤ t function of only this variable
α ∂t ∂x 2 h, T∞
• Thus the governing equation will be transformed into
• Boundary conditions an ODE from PDE
T(0,x) = Ti ; T(t,0) = Ts ; T(t,x→∞) = Ti 0 • There are systematic ways by which this can be
derived, but often involves some qualitative
arguments

1-D Transient in a Semi-Infinite Plate-III 1-D Transient in a Semi-Infinite Plate-IV

• In this course we shall give you the form of the • Substituting for the partial derivatives in the heat
variable. Note that it will be a combination of x and t equation, we get
x ∂η 1 −x d 2T 1
η= ∂η x −0.5 1 dT
(4αt )0.5 ⇒ = , = = 2
∂x (4αt )0.5 ∂t (4α )0.5 t 1.5 α d η 2 t ( 4 αt ) 0 .5
dη (4αt )
• Using chain rule d 2T 1 dT − x ( 4αt ) dT
⇒ = =− 2η
∂T dT ∂η dT 1 dη 2 α dη 2 t (4αt ) 0.5 dη
= =
∂x dη ∂x dη (4αt ) 0.5 d 2T dT
• Thus we get an ODE in η = −2η
2
∂ T ∂  ∂T  ∂  dT 1  d 2 T ∂η
= 1 d 2T 1 dη 2 dη
=  =  dη 2 ∂x (4αt ) 0.5 = dη 2 (4αt )
∂x 2 ∂x  ∂x  ∂x  dη (4αt ) 0.5 
• The boundary conditions
T(t,0) = Ts T(η = 0) = Ts
∂T dT ∂η dT x −0.5 dT −x
= = = T(0,x) = Ti ; T(t,x→∞) = Ti T(η = ∞) = Ti
∂t dη ∂t dη (4α) 0.5 ( t )1.5 dη 2t ( 4αt ) 0.5

4
1-D Transient in a Semi-Infinite Plate-V Error Function
η
d 2T dT 2
• The governing equation is = −2η • The integral ∫ e − u du occurs very frequently in physics
dη 2 dη 0
• Though not integrable in explicit form, it has been
• To get the solution, we make the transformation
integrated with series expansion and tables have been
dT d 2 T dT + constructed under what is called Error Function
= T+ ⇒ 2 = x
dη dη dη 2 −u 2
erf ( x ) = ∫e du erf(x) is tabulated in
• The equation in the new variables can be written as π 0 Appendix B of your
∞ book
dT + • It turns out that ∫ e − u du = π
2
dT + ⇒ = −2ηdη
= −2ηT + + 20 x=3 is as good as ∞
dη T
• Hence erf(∞)=1, erf(0) = 0
2 Erf(3) = 0.99998
• Integration gives ln(T + ) = −η 2 + C ⇒ T + = C1e − η • A complimentary error function
η η u is a is also defined as
dT 2 − η2 −u 2
⇒ = C1e −η ⇒ T = C1 ∫ e dη + C 2 = C1 ∫ e du + C 2 dummy
dη 0 0 variable erfc(x) = 1 – erf(x)

Integration 1-D Transient in a Semi-Infinite Plate-VI

∞ ∞ ∞ 2π η
2 2 −( r 2 )
−( x + y )
∫ ∫e π
2
Consider I= ∫ ∫e dxdy = rdrdθ • The solution T = C1 ∫ e −u du + C 2 = C1 erf (η) + C 2
−∞ −∞ 0 0 0 2
∞ 2π
dt 2π  − t ∞ 
• The Boundary condition, T(η=0) = Ts implies
−t
Put r2 = t ⇒I= ∫ ∫e dθ = − e =π π
0 0 2 2  0  Ts = C1 erf (0) + C 2 ⇒ C 2 = Ts
∞ ∞ ∞ 2
−x 2 − y2 2 −x2
∫ e that
I =Note dx ∫ e dy = 4I1 , where I1 = ∫ e dx
• The Boundary condition, T(η=∞) = Ti implies
−∞ −∞ 0
π π π 2
From above 4I12 = π or I1 = Ti = C1 erf (∞ ) + Ts Ti = C1 + Ts ⇒ C1 = (Ti − Ts )
2 2 2 π

Error function can be computed by using the fact that ∴ T = (Ti − Ts )erf (η) + Ts 1 0.99
−x 2 2x 4 x6 x8 T − Ts θ
e = 1− x + − + + ... ⇒ = erf (η) = θ
2! 3! 4! (Ti − Ts ) 0
0 η 1.82

∂T dT 1
=
∂x dη ( 4αt ) 0.5
1-D Transient in a Semi-Infinite Plate-VII
1-D Transient in a Semi-Infinite Plate-VIII
• For η = 1.82, θ=0.99; This implies that η = 1.82
is for all practical purposes is ∞ ∂T dT 1
q ′′ = − k = −k
x ∂x dη ( 4αt ) 0.5
η = 1.82 ⇒ = 1.82 ⇒ x = 3.64(αt )0.5 x =0 η= 0
(4αt )0.5
dT 2 2
• The above numbers can be interpreted in the • We had shown that = C1e − η ⇒ dT = C1 = (Ti − Ts )
dη dη π
following manner η=0

• x > 3.64 (αt)0.5 can be considered as infinitely thick (Ti − Ts ) 2 (Ts − Ti )

• Similarly for t < x2/(13.25 α), the plate can be ∴ q′′ = −k 0.5
= −k
( 4 αt ) π (πα t ) 0.5
considered infinite
• Now we will turn our attention to heat transferred • Solutions are available for other boundary conditions and
are summarized in your book

5
 Application-I Application-II
• Semi-infinite wall finds lots of applications, such as • At the instant they come in contact, thermal
bubble growth, vapor explosion etc. equilibrium would dictate that the contact temperature
• In vapor explosion, analysis liquid metal drops falling Ts would lie in between T1 and T2
in relatively cold water is required • If the contact temperature remains same, thermal
• This can be idealized by two semi-infinite slabs equilibrium demands that the heat flux from one face
brought into intimate contact should be equal to the heat flux from the other face.

• We shall see the gist by trying to calculate what is • The solution on the two slabs can now be worked
known as the contact temperature similar to what we have obtained with T1 and T2
similar to Ti
T2

∴ q1′′ = − k1
(Ts − T1 ) and q ′2′ = − k 21
(Ts − T2 )
T1 (πα1 t ) 0.5 (πα 2 t ) 0.5

Application-III Finite Difference Method - I

• Since heat flux coming out from 1 will be getting into • We found that analytical methods are complex
the other,
⇒ q1′′ = −q ′2′ ⇒ −k1
(Ts − T1 ) = k (Ts − T2 ) • These methods are restrictive and are applicable for
(πα1t ) 0.5 2
(πα 2 t ) 0 .5 simple boundary conditions
• Numerical methods are easy to implement and we can
(T − T ) k
⇒ 1 s = 2
α1 0.5
(k ρ c )
= 2 2 2 0.5
0.5
obtain results quickly
(Ts − T2 ) k1 α 2 0.5 (k1ρ1c1 ) • We will see the gist of the method
• Rearranging, we get
T1 (k1ρ1c1 )0.5 + T2 (k 2ρ 2 c 2 )0.5
TS ==
(k1ρ1c1 )0.5 + (k 2ρ 2 c 2 )0.5

Finite Difference Method - II Finite Difference Method - III

Governing Equation: ∂T ∂ 2T • One of the FDM approximation is
= α
∂t ∂x 2 n n
Physical Domain: ∂T T n +1 − Tin ∂ 2T T n − 2Tin + Tin−1
= i = i +1
1 i-1 i i+1 N ∂t i ∆t 2
∂x i ∆x 2
Computational • This leads to the nodal equation
n+1
Domain: α∆ t n
t n Tin +1 = Tin + (Ti +1 − 2Tin + Tin−1 )
n ∆x 2
T i
• This method is called explicit method, as the
x
values at Tin+1 are readily obtained explicitly, once
the initial and boundary conditions are known

6
 Finite Difference Method - IV Finite Difference Method - V
Implicit Method
• This method suffers from a disadvantage that the
time step ∆t < 0.5α∆t/∆x2 n n
∂T T n +1 − Tin ∂ 2T Tin++11 − 2Tin +1 + Tin−+11
• This is called stability limit. This occurs due to = i 2
=
∂t i ∆t ∂x ∆x 2
explosion of errors above the limit i

• If we need more accurate results, we need more • This leads to the nodal equation
nodes, and this implies small ∆x. This will limt ∆t
to be small and takes more computational time Tin +1 − Tin T n +1 − 2Tin +1 + Tin−+11
= α i +1
• This can be overcome by choosing a different ∆t ∆x 2
method called implicit method
 α ∆t   2α∆t  n +1  α∆t 
Tin++11  − 2  + Tin +1 1 +  + Ti −1  − 2  = Ti
n
 ∆x   ∆x 2   ∆x 

Finite Difference Method - VI

• For the simple case of boundary temperature
known n+1
n
1 2 3 4 5
 1 0 
 α∆t 2α∆t α∆t   T n +1   T n +1 
− 2 1+ −   1 n +1   1 n 
 ∆x ∆x 2 ∆x 2  T2   T3 
 α∆t 2α∆t α ∆t   n +1  =  n 
 T3   T3 
− 1+ −
 ∆x 2 ∆x 2 ∆x 2
  n +1   n 
α∆t 2α∆t α∆t  T4   T4 
 − 1+ − 2
∆x  T5  T5 
n +1 n +1
 ∆x 2 ∆x 2
 0 1 

• The matrix can be inverted in Matlab or any other software