1. Trang chủ
  2. » Kỹ Thuật - Công Nghệ

Heat Transfer Handbook part 25 docx

10 242 0
Tài liệu đã được kiểm tra trùng lặp

Đang tải... (xem toàn văn)

THÔNG TIN TÀI LIỆU

Thông tin cơ bản

Định dạng
Số trang 10
Dung lượng 162,89 KB

Các công cụ chuyển đổi và chỉnh sửa cho tài liệu này

Nội dung

They also provided a solution when the initial temperature distribution decays exponentially withx.. Constant Surface Heat Flux and Exponentially Decaying Energy Gener-ation When the sur

Trang 1

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

[231],(71)

Lines: 3169 to 3226

———

-0.79182pt PgVar

———

Normal Page

* PgEnds: Eject

[231],(71)

ln



T s + T

T s − T

T s − T i

T s + T i



+ 2



arctanT

T s − arctan

T i

T s



= 4σA s T s3t

ρV c (3.283)

Equation (3.283) is useful, for example, in designing liquid droplet radiation systems for heat rejection on a permanent space station

Simultaneous Convective–Radiative Cooling In this case, the radiative term,

−σA s (T4−T4

s ) appears on the right-hand side ofeq (3.276) in addition to −hA s (T −

T) An exact solution for this case does not exist except when T= T s = 0 For

this special case the exact solution is

1

3ln



1+σT3

i



/h(T /T i )3

(1 + σT3

i /h)(T /T i )3



= hA ρV c s t (3.284)

Temperature-Dependent Heat Transfer Coefficient For natural convection cooling, the heat transfer coefficient is a function of the temperature difference, and the functional relationship is

whereC and n are constants Using eq (3.285) in (3.276) and solving the resulting

differential equation gives

T − T

T i − T∞ =



1+nh ρV c i A s t

−1/n

(3.286) wheren = 0 and h i = C(T i − T∞)n.

Heat Capacity of the Coolant Pool Ifthe coolant pool has a finite heat capacity, the heat transfer to the coolant causesT∞to increase Denoting the properties ofthe hot body by subscript 1 and the properties ofthe coolant pool by subscript 2, the temperature–time histories as given by Bejan (1993) are

T1(t) = T1(0) − T1(0) − T2(0)

1+ ρ1V1c12V2c2

T2(t) = T2(0) + T1(0) + T2(0)

1+ (ρ2V2c21V1c1) (1 − e −nt ) (3.288)

wheret1(0) and T2(0) are the initial temperatures and

n = hA s ρ1V1c1+ ρ2V2c2

1V1c1)(ρ2V2c2) (3.289)

Trang 2

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

[232], (72)

Lines: 3226 to 3285

———

0.00218pt PgVar

———

Long Page PgEnds: TEX [232], (72)

q0⬙

T t T = T(0, ) = s iat nk⫺⫺⳵T = q

X x=| 0 0 ⫺k⫺⫺⳵ = h TT t

T

X x=| 0 [ a (0, )]

T x( ,0) =T i T x( ,0) =T i

h T, a

T s

Figure 3.37 Semi-infinite solid with (a) specified surface temperature, (b) specified surface heat flux, and (c) surface convection.

As indicated in Fig 3.37, the semi-infinite solid model envisions a solid with one identifiable surface and extending to infinity in all other directions The parabolic partial differential equation describing the one-dimensional transient conduction is

2T

Specified Surface Temperature Ifthe solid is initially at a temperatureT i, and iffor timet > 0 the surface at x = 0 is suddenly subjected to a specified temperature–

time variationf (t), the initial and boundary conditions can be written as

T (0, t) = f (t) = T i + at n/2 (3.291b)

wherea is a constant and n is a positive integer.

Using the Laplace transformation, the solution forT is obtained as

T = T i + aΓ1+n

2



(4t) n/2 i nerfc

x

2√

αt



(3.292)

whereΓ is the gamma function (Section 3.3.2) and i n erfc is the nth repeated integral

ofthe complementary error function (Section 3.3.1) The surface heat fluxq

0 is

q

0 = 2√n−1

αt (n−1)/2 kaΓ



1+n

2

 

i n−1erfc x

2√

αt



x=0

(3.293) Several special cases can be deduced from eqs (3.292) and (3.293)

Trang 3

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

[233],(73)

Lines: 3285 to 3339

———

-1.99896pt PgVar

———

Long Page

* PgEnds: Eject

[233],(73)

Case 1: f (t) = T0 This is the case ofconstant surface temperature which occurs whenn = 0 and a = T0− T i:

T (x, t) − T i

T0− T i = erfc



x

2√

αt



(3.294)

q

0 = k(T0− T i )

Case 2: f (t) = T i +at This is the case ofa linear variation ofsurface temperature

with time which occurs whenn = 2:

T (x, t) = T i + 4ati2erfc



x

2√

αt



(3.296)

q

Case 3: f (t) = T i + at1/2 This is the case ofa parabolic surface temperature

variation with time withn = 1.

T (x, t) = T i + aπt i erfc

 x

2√

αt



(3.298)

q

0 = ka

2

 π

Specified Surface Heat Flux For a constant surface heat fluxq

0, the boundary conditions ofeq (3.291b) is replaced by

− k ∂T (0, t)

and the solution is

T (x, t) = T i+2q0

αt

k i erfc

 x

2√

αt



(3.301)

Surface Convection The surface convection boundary condition

− k ∂T (0, t)

∂x = h [T− T (0, t)] (3.302)

replaces eq (3.291b) and the solution is

T (x, t) − T i

T∞ − T i = erfc



x

2√

αt



− e (hx/k)+(h2αt/k2)erfc

h k

αt + x

2√

αt



(3.303)

Trang 4

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

[234], (74)

Lines: 3339 to 3386

———

5.48827pt PgVar

———

Normal Page PgEnds: TEX [234], (74)

Constant Surface Heat Flux and Nonuniform Initial Temperature Zhuang

et al (1995) considered a nonuniform initial temperature of the form

wherea and b are constants to find a modified version ofeq (3.301) as

T (x, t) = a + bx + 2αt

q

0

k + b



i erfc

 x

2√

αt



(3.305) The surface temperature is obtained by puttingx = 0 in eq (3.305), which gives

T (0, t) = a + 2



αt

π

q

0

k + b



(3.306)

Zhuang et al (1995) found that the predictions from eqs (3.305) and (3.306) matched the experimental data obtained when a layer ofasphalt is heated by a radiant burner, producing a heat flux of41.785 kW/m2 They also provided a solution when the initial temperature distribution decays exponentially withx.

Constant Surface Heat Flux and Exponentially Decaying Energy Gener-ation When the surface of a solid receives energy from a laser source, the effect of this penetration ofenergy into the solid can be modeled by adding an exponentially decaying heat generation term, ˙q0e −ax /k (where a is the surface absorption

coeffi-cient), to the left side of eq (3.290) The solution for this case has been reported by Sahin (1992) and Blackwell (1990):

T = T i + (T0+ T i ) erfc

 x

2√

αt



+ ˙q0

ka2



erfc



x

2√

αt



−1

2e a2αt+axerfc



x

2√

αt + a

αt



−1

2e a2αt−axerfc x

2aαt − a

αt



+ e −ax e (a2αt−1)



(3.307)

Both Sahin (1992) and Blackwell (1990) have also solved this case for a convective boundary condition (h, T∞) atx = 0 Blackwell’s results show that for a given

absorption coefficienta, thermal properties of α and k, initial temperature T i, and surface heat generation ˙q0, the location ofthe maximum temperature moves deeper into the solid as time progresses Ifh is allowed to vary, for a given time the greater

value ofh provides the greater depth at where the maximum temperature occurs.

The fact that the maximum temperature occurs inside the solid provides a possible explanation for the “explosive removal of material” that has been observed to occur when the surface ofa solid is given an intense dose oflaser energy

Trang 5

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

[235],(75)

Lines: 3386 to 3439

———

0.65923pt PgVar

———

Normal Page PgEnds: TEX [235],(75)

Consider one-dimensional transient conduction in a plane wall ofthickness 2L, a long

solid cylinder ofradiusr0, and a solid sphere ofradiusr0, each initially at a uniform temperatureT i At timet = 0, the exposed surface in each geometry is exposed to a

hot convective environment (h, T∞) The single parabolic partial differential equation describing the one-dimensional transient heating ofall three configurations can be written as

1

s n

∂s



s n ∂T

∂s



= 1 α

∂T

wheres = x, n = 0 for a plane wall, s = r, n = 1 for a cylinder, and s = r, n = 2

for a sphere In the case of a plane wall,x is measured from the center plane The

initial and boundary conditions for eq (3.308) are

∂T (0, t)

k ∂T (L or r0, t)

∂s = h [T− T (L or r0, t)] (3.309c) According to Adebiyi (1995), the separation ofvariables method gives the solution forθ as

θ =



n=1

2 Bi

λ2

n+ Bi2+ 2ν · Bi

RνJ−ν(λ n R)

J−ν(λ n ) e−λ

2

whereθ = (T− T )/(T− T i ), R = x/L for a plane wall, R = r/r0 for both cylinder and sphere, Bi= hL/k or hr0/k, τ = αt/L2orτ = αt/r2

0, ν = (1 − n)/2,

and theλnare the eigenvalues given by

λn J−(ν−1)(λ n ) = Bi · J−ν(λ n ) (3.311) The cumulative energy received over the timet is

Q = ρcV (T∞ − T i )



n=1

2(1 + n) Bi2

1− e−λ 2

nτ

λ2

n



λ2

n+ Bi2+ 2ν · Bi (3.312)

Solutions for a plane wall, a long cylinder, and a sphere can be obtained from eqs

(3.310) and (3.312) by puttingn = 0, (ν = 1

2), n = 1(ν = 0), and n = 2(ν = −1

2),

respectively It may be noted that

J −1/2 (z) =



2

πz

1/2

cosz

Trang 6

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

[236], (76)

Lines: 3439 to 3461

———

-2.89299pt PgVar

———

Normal Page

* PgEnds: Eject

[236], (76)

J1/2 (z) =



2

πz

1/2

sinz

J3/2 (z) =



2

πz

1/2

sinz

z − cos z



With these representations, eqs (3.310) and (3.312) reduce to the standard forms appearing in textbooks Graphical representations ofeqs (3.310) and (3.312) are

called Heisler charts.

For some configurations it is possible to construct multidimensional transient conduc-tion soluconduc-tions as the product ofone-dimensional results given in Secconduc-tions 3.8.2 and 3.8.3 Figure 3.38 is an example ofa two-dimensional transient conduction situation

in which the two-dimensional transient temperature distribution in a semi-infinite plane wall is the product ofthe one-dimensional transient temperature distribution in

an infinitely long plane wall and the one-dimensional transient temperature distribu-tion in a semi-infinite solid Several other examples ofproduct soludistribu-tions are given by Bejan (1993)

Explicit Method For two-dimensional transient conduction in Cartesian coordi-nates, the governing partial differential equation is

T x x t( , , )1 2 ⫺TT x t( , )2 ⫺TT x t( , )1 ⫺T

T ,h

T ,h

T ,hT ,hT ,hT ,h

semi-infinite plane wall

infinitely long plane wall

semi-infinite solid

x1

x1

x2

x2

Figure 3.38 Product solution for a two-dimensional transient conduction problem

Trang 7

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

[237],(77)

Lines: 3461 to 3514

———

4.84828pt PgVar

———

Normal Page PgEnds: TEX [237],(77)

2T

∂x2 +∂ ∂y2T2 = 1α∂T ∂t (3.313) which assumes no internal heat generation and constant thermal properties Approx-imating the second-order derivatives inx and y and the first-order derivative in t by

for various nodes (see Fig 3.33) can be expressed using the Fourier modulus, Fo=

α ∆t/(∆x)2, Bi = h ∆x/k, and t = p ∆t.

• Internal node: With Fo ≤ 1

4,

T i,j p+1= FoT1p +1,j + T i−1,j p + T i,j+1 p + T i,j−1 p + (1 − 4Fo)T i,j p (3.314)

• Node at interior corner with convection: With Fo(3 + Bi) ≤ 3

4,

T i,j p+1=2

3Fo



T i+1,j p + 2T i−1,j p + 2T i,j+1 p + T i,j−1 p + 2BiT∞

+1− 4Fo −4

3Bi· FoT i,j p (3.315)

• Node on a plane surface with convection: With Fo(2 + Bi) ≤ 1

2,

T i,j p+1 = Fo2T i−1,j p + T i,j+1 p + T i,j−1 p + 2Bi · T∞

+ (1 − 4Fo − 2Bi · Fo) T i,j p (3.316)

• Node at exterior corner with convection: With Fo(1 + Bi) ≤ 1

4,

T i,j p+1= 2FoT i−1,j p + T i,j−1 p + 2Bi · T



+ (1 − 4Fo − 4Bi · Fo) T i,j p = 0 (3.317)

• Node on a plane surface with uniform heat flux: With Fo ≤ 1

4,

T i,j p+1 = (1 − 4Fo)T i,j p + Fo2T i−1,j p + T i,j+1 p + T i,j−1 p + 2Fo · q∆x

k (3.318)

The choice of∆x and ∆t must satisfy the stability constraints, introducing each

ofthe approximations given by eqs (3.314)–(3.318) to ensure a solution free of numerically induced oscillations Once the approximations have been written for each node on the grid, the numerical computation is begun witht = 0(p = 0), for which

the node temperatures are known from the initial conditions prescribed Because eqs

(3.313)–(3.318) are explicit, node temperatures att = ∆t(p = 1) can be determined

from a knowledge of the node temperatures the preceding time,t = 0(p = 0) This

Trang 8

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

[238], (78)

Lines: 3514 to 3565

———

-5.41692pt PgVar

———

Normal Page PgEnds: TEX [238], (78)

“marching out” in time type ofcomputation permits the transient response ofthe solid to be determined in a straightforward manner However, the computational time necessary to cover the entire transient response is excessive because extremely small values of∆t are needed to meet the stability constraints.

Implicit Method In the implicit method, the second derivatives in x and y are approximated by central differences but with the use oftemperatures at a subsequent

time,p + 1, rather than the current time, p, while the derivative in t is replaced by

a backward difference instead of a forward difference Such approximations lead to

the following equations:

• Internal node:

(1 + 4Fo)T i,j p+1− FoT i+1,j p+1 + T i−1,j p+1 + T i,j+1 p+1 + T i,j−1 p+1= T i,j p (3.319)

• Node at interior corner with convection:

(1 + 4Fo)1+1

3Bi T i,j p+1−2

3Fo



T i+1,j p+1 + 2T i−1,j p+1 + 2T i,j+1 p+1 + T i,j−1 p+1= T i,j p +4

3Fo· Bi · T∞ (3.320)

• Node on a plane surface with convection:

[1+ 2Fo(2 + Bi)] T i,j p+1− Fo2T i−1,j p+1 + T i,j+1 p+1 + T i,j−1 p+1

• Node at exterior corner with convection:

1+ 4Fo(1 + Bi)T i,j p+1− 2FoT i−1,j p+1 + T i,j−1 p+1= T i,j p + 4Bi · Fo · T∞ (3.322)

• Node on a plane surface with uniform heat flux:

(1 + 4Fo)T i,j p+1+ Fo2T i−1,j p+1 + T i,j+1 p+1 + T i,j−1 p+1= T i,j p +2Fo· q∆x

k (3.323)

The implicit method is unconditionally stable and therefore permits the use of higher values of∆t, thereby reducing the computational time However, at each time

t, the implicit method requires that the node equations be solved simultaneously rather than sequentially.

Other Methods Several improvements ofthe explicit and implicit methods have been advocated in the numerical heat transfer literature These include the three-time-level scheme ofDufort and Fankel, the Crank–Nicholson method, and alternating

Trang 9

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

[239],(79)

Lines: 3565 to 3609

———

-5.78989pt PgVar

———

Normal Page

* PgEnds: Eject

[239],(79)

direction explicit methods For a discussion ofthese methods as well as stability analysis, the reader should consult Pletcher et al (1988)

Examples ofperiodic conduction are the penetration ofatmospheric temperature cycles into the ground, heat transfer through the walls of internal combustion engines, and electronic components under cyclic operation The periodicity may appear in the differential equation or in a boundary condition or both The complete solution to

a periodic heat conduction problem consists ofa transient component that decays

to zero with time and a steady oscillatory component that persists It is the steady oscillatory component that is ofprime interest in most engineering applications In this section we present several important solutions

Environment

Revisit the lumped thermal capacity model described in Section 3.8.1 and consider

a scenario in which the convective environmental temperatureT∞ oscillates sinu-soidally, that is,

wherea is the amplitude ofoscillation, ω = 2πf the angular frequency, f the

frequency in hertz, andT ∞,mthe mean temperature ofthe environment

The method ofcomplex combination described by Arpaci (1966), Myers (1998), Poulikakos (1994), and Aziz and Lunardini (1994) gives the steady periodic solution as

θ = √ 1

where

θ = T − T a ∞,m B = ρV c hAω τ =ρV c hAt β = arctan B (3.326)

A comparison ofeq (3.325) with the dimensionless environmental temperature vari-ation (sinBτ) shows that the temperature ofthe body oscillates with the same

fre-quency as that ofthe environment but with a phase lag ofβ As the frequency of

oscillation increases, the phase angleβ = arctan B increases, but the amplitude of

oscillation 1/(1 + B2)1/2decreases.

Consider the semi-infinite solid described in Section 3.8.2 and let the surface temper-ature be ofthe form

Trang 10

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45

[240], (80)

Lines: 3609 to 3664

———

5.8702pt PgVar

———

Normal Page PgEnds: TEX [240], (80)

In this case, eqs (3.291) are still applicable, although the initial condition ofeq

(3.291a) becomes irrelevant for the steady periodic solution, which is

T (x, t) = T i + ae−[(ω/2α)1/2 x] cosωt − ω

1/2

x



(3.328)

Three conclusions can be drawn from this result First, the temperatures at all lo-cations oscillate with the same frequency as the thermal disturbance at the surface

Second, the amplitude ofoscillation decays exponentially withx This makes the

so-lution applicable to the finite thickness plane wall Third, the amplitude ofoscillation decays exponentially with the square root ofthe frequencyω Thus, higher-frequency

disturbances damp out more rapidly than those at lower frequencies This explains why daily oscillations ofambient temperature do not penetrate as deeply into the ground as annual and millenial oscillations The surface heat flux variation follows directly from eq (3.328):

q(0, t) = −k ∂T (0, t) ∂x = ka

α

1/2

cos



ωt −π

4



(3.329) and this shows thatq(0, t) leads T (0, t) by π/4 radians.

In this case, the boundary condition ofeq (3.327) is replaced with

q(0, t) = −k ∂T

∂x (0, t) = q0cosωt (3.330) and the solution takes the form

T (x, t) = T i+q0

k

 α ω

1/2

e−[ω/2α)1/2 x] cos



ωt − ω

1/2

x − π

4



(3.331)

It is interesting to note that the phase angle increases as the depthx increases with the

minimum phase angle ofπ/4 occurring at the surface (x = 0) A practical situation

in which eq (3.331) becomes useful is in predicting the steady temperature variations induced by frictional heating between two reciprocating parts in contact in a machine

This application has been described by Poulikakos (1994)

The surface boundary condition in this case is

Ngày đăng: 05/07/2014, 16:20

TỪ KHÓA LIÊN QUAN