Chapter 4 Unsteady state conduction (Text: J. P. Holman, Heat Transfer, 8th ed.

, McGraw Hill, NY)

Consider 1-D unsteady state conduction problem in a slab (Fig. 4-1)
2T q 1 T + = , 0 x L, t > 0 x 2 k t T = T0 in 0 x L, t = 0 T = 0 at x = 0, t > 0 x T k + h(T T ) = 0 at x = L, t > 0 x

There are totally 7 parameters: k , q, , L, T0 , h, and T Nondimensinoalization, T T x let = , = L T0 T T0 T 2 q 1 T + = L2 2 k t qL2 L2 2 T + = 2 k (T0 T ) (T0 T ) t 2 +G = 2 where =

t
L2

= Fourier No. G =

qL2 = dimensionaless heat source k (T0 T )

With initial and boundary conditions

= 1 in 0 1, = 0 = 0 at = 0, > 0 + Bi = 0 at = 0, > 0.
where Bi = hL = Biot No. k

There are only 2 paramters: G and Bi.

The physical meanings of Fo and Bi: t (k / L) L2 rate of heat conduction across L in L3 = = = Fo. L2 cL3 / t rate of heat storage in L3

The larger the Fourier No. is, the deeper is the penetration of heat inot a solid over a given period of time.
Sheu Long Jye Dep. of Mechanical Engineering Chung Hua University [email protected]

Bi = =

hL L / kA conduction resistance = = k 1/ hA convection resistance h heat transfer coefficient at surface of a solid = k/L internal conductance of solid across L

Biot number is the ratio of the heat transfer coefficient to unit conductance of a solid over the characteristic length. For solids in the shape of a slab, long cylinder, or a sphere with no internal heat generation, 1-D, transient temperature distribution within solid may be considered uniform if

Bi =

hLs V < 0.1 with error < 5%, Ls = k A

If we analyze systems which are considered uniform in temperature at any instant during the transient conduction, this type of analysis is called lumped heat capacity system. Consider a hot ball immersed in a cool run of water and assume that lumped heat capacity method might be used. The convection heat loss form the body is evidenced as a decrease in internal energy of the body (Fig. 3-2) dT , T (0) = T0 hA(T T ) = cV dt
t t T T cV = e cV = e , = = time constant. T0 T hA hA

When the internal resistance of the body is significient, the temperature may not be assumed uniform and heat equation for 1-D slab problem is 2T 1 T = x 2 t Consider the heat equation
2T 1 T = , 0 < x < , t > 0 x 2 t subject to IC and BCs T ( x, 0) = Ti , 0 < x < , t = 0 T (0, t ) = T0 , t > 0 Solving by Laplace transform defined as
Sheu Long Jye Dep. of Mechanical Engineering Chung Hua University [email protected]

L[ f ] = f (t )e st dt , L[ f '] = sf f (0) .
0

The Laplace transform of the governing equations gives

dT s T T = 2 dx T T (0, s ) = 0 s T T = i + Ae qx + Be qx , q = s / s The BCs gives T T T T = i + 0 i e s / x s s
s x x 1 e x 1 T = L [T ] = Ti + (T0 Ti )erfc , L [ s ] = erfc 2 t 2 t 2 x 2 where erfc( x) = 1 erf ( x) = 1 e d , (Table A 1)

T Ti x x ) = 1 erf ( ) = erfc( T0 Ti 2 t 2 t T T0 x ) = erf ( Ti T0 2 t T q q " = k = x A =k T0 Ti

x2 4 t

If heat flux boundary condition is applied,

Sheu Long Jye Dep. of Mechanical Engineering Chung Hua University [email protected]

T x

" = q0 x =0

T T = i + Ae s / x s q" T k = 0 x x =0 s s x x2 e t x L1[ ] = 2( )1/ 2 e 4 t xerfc s 2 t s

" " x 2q0 t 1/ 2 4x t q0 x T = L [T ] = Ti + erfc ( ) e k k 2 t If convective boundary condition is applied, the solution is 1
2

T x

= h(T T ) x =0
x =0

T=

Ti + Ae qx s T hT k = hT x x =0 s

x =0

h (T T ) h(T Ti ) Ti k i qx ,T = + A= e s (kq + h) s s(q + h / k ) 1 e qx 1 x ] = erfc L [ s(q + ) 2 t T Ti x = erfc T Ti 2 t k x+ e

h h 2 t k2

1 x + 2t x + t erfc e 2 t

x h t erfc + 2 t k

x T Ti h t , = f 2 t T Ti k

Laplace transform on infinite slab.

Sheu Long Jye Dep. of Mechanical Engineering Chung Hua University [email protected]

 2T 1 T = , L < x < L, t > 0 x 2 t subject to IC and BCs T ( x, 0) = Ti , L < x < L T x k =0

x =0

T = h(T T ) at x = L x

let = T T 2 1 , L < x < L, t > 0 = x 2 t subject to IC and BCs ( x, 0) = i , L < x < L x k =0

x =0

= h at x = L x

By Laplace transform

2 q 2 = i 2 s x = = i
s + A cosh qx + B sinh qx, q = h i cosh qx s (kq sinh qL + h cosh qL) L+ x + erfc 2 t k ( L x )+ + e
h h 2 t k2

i
s

Lx = i 1 erfc 2 t

Lx h t + erfc 2 t k

+ .....

The solutions of 1-D transient heat conduction in infinite slab, cylinder, and sphere are summarized and figured by Heilser chart. (Fig. 4-5 to 4-16)
Multidimensional system Consider an infinite rectangular bar (Fig.)

Sheu Long Jye Dep. of Mechanical Engineering Chung Hua University [email protected]

 2 2 1 + = , L1 < x < L1 , L2 < z < L2 , t > 0 (*) x 2 z 2 t If 1 ( x, t ) satisfies 21 1 1 = in L1 < x < L1 x 2 t and 2 ( x, t ) satisfies 2 2 1 2 = in L2 < z < L2 z 2 t Then = 1 2 satisfies (*) if at the surface the medium subjects to k = h1 at x = L1 , and x k = h2 at z = L2 z

can be obtained as

= 1 2 , where
21 1 1 = , k 1 = h11 at x = L1 2 x t x 2 2 1 2 , k 2 = h2 2 at z = L2 = 2 z t z Heat transfer in multidimensional systems for 2-D system:

q q q q = + 1 q0 1 q0 total q0 1 q0 2 For 3-D q q q q q q q = + 1 + 1 1 q0 1 q0 3 q0 1 q0 2 q0 total q0 1 q0 2 Examples:

Transient Finite Difference Method Consider a 2-D body within which the heat flow is governed by

Sheu Long Jye Dep. of Mechanical Engineering Chung Hua University [email protected]

k(

2T 2T T + 2 ) = c 2 x y t

2T Tm +1, n + Tm 1,n 2Tm ,n , (x) 2 x 2 2T Tm ,n +1 + Tm,n 1 2Tm ,n , (y ) 2 y 2

p +1 p T Tm, n Tm ,n t t If the nodal temperature derivatives are evaluated at time p:

Tmp+1, n + Tmp1, n 2Tmp, n (x) 2 if x = y

Tmp, n +1 + Tmp,n 1 2Tmp,n ( y ) 2

p +1 p 1 Tm ,n Tm ,n t

1 p p p p p Tmp,+ n = Fo(Tm +1, n + Tm 1, n + Tm , n +1 + Tm , n 1 ) + (1 4 Fo)Tm , n

where Fo =

t
(x) 2

If the system is 1-D in x (Fig.)

Tmp +1 = Fo(Tmp+1 + Tmp1 ) + (1 2 Fo)Tmp The above expression are called explicit formulations Stability criterion: For 1-D system: Fo 1/ 2 . For 2-D system: Fo 1/ 4 . Example: (Fig.) x and t are such that
1.1 2 1.1 1.1 (2 + 2) + (1 2i )i2.1 = 2.2 0.21 < 2(break 2nd Law) Tmp +1 = 2 2 Fo = The boundary nodes: Consider the transient energy balance at node (m,n) on a boundary surface, (Fig)

Sheu Long Jye Dep. of Mechanical Engineering Chung Hua University [email protected]

x if x = y

k y

Tmp1,n Tmp,n

p p p p p +1 p x Tm ,n +1 Tm ,n x Tm,n 1 Tm ,n x Tm ,n Tm,n p +k +k + hy (T Tm, n ) = c y 2 2 2 y y t

1 p p p Tmp,+ n = Fo{2 BiT + 2Tm 1, n + Tm , n 1 + Tm , n +1 + [

1 2 Bi 4]Tmp,n } Fo

Fo =

t
(x)
2

, Bi =

hx k
1 2 Bi 2]Tmp } Fo

Corresponding 1-D relation is

Tmp +1 = Fo{2 BiT + 2Tmp1 + [

Stability criterion:

1 For 1-D system: Fo , 1 2 Fo 2 BiFo 0 . 2 1 For 2-D system: Fo , 1 4 Fo 2 BiFo 0 . 4

Implicit formulations:
1 p +1 p +1 Tmp++1, n + Tm 1, n 2Tm , n

(x) 2 if x = y

1 p +1 p +1 Tmp,+ n +1 + Tm , n 1 2Tm , n

(y ) 2

p +1 p 1 Tm ,n Tm,n t

1 p +1 p +1 p +1 p +1 p (1 4 Fo)Tmp,+ n Fo(Tm +1, n + Tm 1, n + Tm , n +1 + Tm , n 1 ) = Tm , n

linear equations system X = A1 B AX = B

Sheu Long Jye Dep. of Mechanical Engineering Chung Hua University [email protected]