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

Heat Transfer Handbook part 121 doc

10 47 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 193,85 KB

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

Nội dung

Assuming one-dimensional vapor flow, the differential vapor pressure drop can be expressed in terms of the pressure drops due to frictional forces and dynamic pressure, or dP v f · Reµ v

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

[1197],(17)

Lines: 410 to 475

———

2.23228pt PgVar

———

Normal Page

* PgEnds: Eject

[1197],(17)

TABLE 16.2 Expressions for the Effective Capillary Radius r cfor Several Wick Structures

Circular cylinder (artery

or tunnel wicks)

reff = r r = radius ofliquid flow passage

Rectangular groove reff = w w = groove width

Triangular groove reff = w

cosβ w = groove width

β = half-included angle

Wire screens reff = w + d w

2N N = screen mesh number

w = wire spacing

d w= wire diameter

Packed spheres r c = 0.41r s r s= sphere radius

Source: Chi (1976), with permission.

The differential liquid pressure drop in the wick structure assuming one-dimen-sional laminar flow can be expressed as

dP l

whereK represents the wick permeability The wick permeability is related directly

to the porosity ofthe wick structure, which is defined as the ratio ofpore volume to total volume, orε = Vpore/Vtot, and is given by

K = f2ε(r h )2

As the hydraulic radius ofthe porous structure is typically small and the liquid flow velocity is low, the liquid flow can be assumed laminar Thus, the values of(f l· Rel) can be assumed constant and depend only on the flow passage shape, where typical values of the permeability for different wick structures are given in Table 16.3

Assuming one-dimensional vapor flow, the differential vapor pressure drop can

be expressed in terms of the pressure drops due to frictional forces and dynamic pressure, or

dP v

(f · Reµ v ˙m v (x)

2A v r2

h,vρv

2˙m v

A2

vρv

d ˙m v (x)

Recognizing that the total vapor mass flow rate and liquid mass flow rate are equal

at steady-state conditions, the mass flow rate through the system can be expressed in terms ofthe heat transport rate and the latent heat ofvaporization ofthe fluid, or

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

[1198],(18)

Lines: 475 to 475

———

0.8034pt PgVar

———

Normal Page PgEnds: TEX

[1198],(18)

TABLE 16.3 Expressions of Wick Permeability K for Several Wick Structures

8 Open rectangular grooves K = 2

2

h,l

f l· Rel

w s

s = groove pitch

r h,l= w + 2δ2

w = groove width

δ = groove depth

(f l· Rel ) from (a) below

Circular annular wick K = 2r

2

h,l

f l· Rel r h,l = r1− r2

(f l· Rel ) from (b) below

Wrapped screen wick K = d2 3

122 2 d = wire diameter

1.05πNd

4

N = mesh number

37 2 r s= sphere radius

on packing mode)

0

12

16 14

18 16

20 18

22 20

24 22

26 24

r r2 1/

a

r2r1

w

a⬅ ␦— w

Source: Chi (1976), with permission.

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

[1199],(19)

Lines: 475 to 531

———

0.13634pt PgVar

———

Normal Page

* PgEnds: Eject

[1199],(19)

The effective length of the heat pipe,Leff, is used to represent the average distance that the liquid and vapor must travel along the heat pipe

Leff = 1

˙m

0

Q

0

Assuming uniform evaporation and condensation in the evaporator and condenser regions, the mass flow rates in the evaporator and condenser vary linearly and the effective length of a heat pipe becomes

Leff = L e

2 + L a+L c

and the total liquid pressure drop can then be expressed as

0

dP l

KA wρl h fg (16.18)

Assuming that the dynamic pressure drop is fully recovered in the condenser region, the vapor pressure drop will be

0

dP v

(f · Re) vµv QLeff

2A v r2

h,vρv h fg (16.19)

For cases where compressibility effects must be included and the dynamic pressure

is not fully recovered, see Busse (1973) or Ivanovskii et al (1982) Due to low liquid velocities and the small characteristic dimensions ofthe wick structure, the liquid flow is always generally assumed to be laminar However, the vapor flow velocities may be sufficient to correspond to turbulent flow In this case, the heat transfer rate

Q and the Reynolds number Re are related, and the term (f · Re) vmust be evaluated from a friction factor correlation for turbulent flow

To determine the friction factorf v, the vapor flow regime must be evaluated

Expressing the Reynolds number in terms ofthe heat inputQ, the flow regime is

determined from

Rev =A2r v Q

For laminar flow (Re < 2300) in a circular cross section, the term (f · Re) v is a constant

while for turbulent flow(Re > 2300) in a circular cross section the Blasius

correla-tion can be used

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

[1200],(20)

Lines: 531 to 577

———

0.08427pt PgVar

———

Long Page

* PgEnds: Eject

[1200],(20)

f v = 0.038

Values forf ·Re for noncircular cross sections can be obtained from most convective

heat transfer textbooks as a function of the cross-sectional geometry

In the case oflaminar vapor flow, the substitution ofthe individual pressure-drop terms into the capillary limit, eq (16.9), results in an algebraic expression that can be solved directly forQ For turbulent vapor flow, different methods exist

for solving for Q The first method for determining the capillary limit when the

vapor flow is turbulent is that ofan iterative solution This procedure begins with

an initial estimation ofthe capillary limit where the solution first assumes laminar, incompressible vapor flow Using these assumptions, the maximum heat transport capacityQ can be determined by substituting the values for individual pressure drops

and solving for the axial heat transfer Once this value has been obtained, the axial heat transfer can be substituted into expressions for the vapor Reynolds number to determine the accuracy ofthe original assumptions Using this iterative approach, accurate values for the capillary limitation as a function of the operating temperature can be determined where the operating temperature effects the capillary limit due to the temperature dependence ofthe fluid properties

For a more direct solution ofthe capillary limit when turbulent flow is present in the vapor channel, it is possible to substitute the Blausius correlation into the vapor pressure-drop term Then, separating all terms other than heat inputQ(x) and length

x into friction coefficients F vandF l, the capillary limit can be expressed as

(∆P c )max− ∆P− ∆P=

0

(F l + F v )Qdx (16.23) where the liquid frictional coefficientF lis given by

and the vapor frictional coefficientF vis evaluated from the expression

2r2

For the case ofturbulent vapor flow, the vapor friction coefficient was modified by substituting eq (16.22) for(f · Re) v which results in an expression for the vapor

friction coefficient of

F v= 0.038 2

r2

v A vρv h fg



2r v Q

A vµv h fg

(16.26)

Substitution ofthis expression and combining with those discussed previously results

in a general pressure balance relationship for turbulent vapor flow which takes the form

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

[1201],(21)

Lines: 577 to 639

———

-0.09183pt PgVar

———

Long Page PgEnds: TEX

[1201],(21)

(∆P c )max− ∆P− ∆P= 0.019µ v

r2

v A vρv h fg



2r v

A vµ v h fg

0

Q7/4 dx + F l

0

Q dx

(16.27) This expression and a Newton–Raphson method to determine the roots from the re-sulting polynomial equation, the maximum heat transport capacity (i.e., the capillary limit) for a given heat pipe can be determined as a function of the evaporator and condenser lengths and the operating temperature

To solve for the capillary limit without iteration or numerical integration, an esti-mation ofthe capillary limit may be obtained where the friction factor is estimated and assumed constant for the entire operating range Inspection of a traditional Moody (1944) friction factor diagram reveals that beyond a Reynolds number of 105the fric-tion factor becomes constant as the flow enters the fully turbulent region By assuming

a friction factor for Re> 105, the capillary limit results in a quadratic equation for

Q and a much easier solution This method typically produces reasonable results that

tend more on the conservative side

16.2.3 Boiling Limit

At higher heat fluxes, nucleate boiling may occur in the wick structure, which may allow vapor to become trapped in the wick, thus blocking liquid return and resulting in

evaporator dryout This phenomenon, referred to as the boiling limit, differs from the

other limitations discussed previously, as it depends on the radial or circumferential heat flux applied to the evaporator, as opposed to the axial heat flux or total thermal power transported by the heat pipe

Determination ofthe heat flux or boiling limit is based on nucleate boiling theory and is comprised oftwo separate phenomena: (1) bubble formation and (2) subse-quent growth or collapse ofthe bubbles Bubble formation is governed by the size (and number) ofnucleation sites on a solid surface and the temperature difference between the heat pipe wall and the working fluid The temperature difference, or superheat, governs the formation of bubbles and can typically be defined in terms of the maximum heat flux as

Q = keff

wherekeff is the effective thermal conductivity of the liquid–wick combination The critical superheat∆Tcis defined by Marcus (1965) as

h fgρv





(16.29)

whereTsatis the saturation temperature ofthe fluid andr nis the critical nucleation

site radius, which according to Dunn and Reay (1982) ranges from 0.1 to 25.0µm for conventional metallic heat pipe case materials As discussed by Brennan and Kroliczek (1979), this model yields a very conservative estimate ofthe amount of

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

[1202],(22)

Lines: 639 to 654

———

2.83023pt PgVar

———

Normal Page PgEnds: TEX

[1202],(22)

TABLE 16.4 Effective Thermal Conductivity k efor Liquid-Saturated Wick Structures

Wick and liquid in series

k e= ek k l k w

w + k l k e= effectivethermal

conductivity

k l= liquid thermal

conductivity

k w= wick material

thermal conductivity

Wick and liquid in parallel

k e = ek l + k w

Wrapped screen k e= k l[(k l + k w l − k w )]

(k l + k w l − k w )

Packed spheres k e= k l[(2k (2k l + k w l − k w )]

Rectangular grooves k e= (w f k (w + w l k w δ) + wk l (0.185w f k w + δk l )

f )(0.185w f k f + δk l ) w f = groove finthickness

w = wick thickness

δ = groove depth

Source: Chi (1976), with permission.

superheat required for bubble formation and is true even when using the lower bound for the critical nucleation site radius This is attributed to the absence of adsorbed gases on the surface of the nucleation sites caused by the degassing and cleaning procedures used in the preparation and charging ofheat pipes

The growth or collapse ofa given bubble once established on a flat or planar surface is dependent on the liquid temperature and corresponding pressure difference across the liquid–vapor interface caused by the vapor pressure and surface tension ofthe liquid By performing a pressure balance on any given bubble and using the Clausius–Clapeyron equation to relate the temperature and pressure, an expression for the heat flux beyond which bubble growth will occur may be developed (Chi, 1976) and expressed as

Q b= 2πLeffkeffT v

A v h fgρv ln(r i /r v )





(16.30)

wherer i is the inner pipe wall radius andr v is the vapor core radius Relationships

to determine the effective conductivity,keff, ofthe liquid saturated wick are given in Table 16.4

16.2.4 Entrainment Limit

Examination ofthe basic flow conditions in a heat pipe shows that the liquid and vapor flow in opposite directions The interaction between the countercurrent liquid and vapor flow results in viscous shear forces occurring at the liquid–vapor interface,

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

[1203],(23)

Lines: 654 to 689

———

3.7872pt PgVar

———

Normal Page

* PgEnds: Eject

[1203],(23)

which may inhibit liquid return to the evaporator In the most severe cases, waves may form and the interfacial shear forces may become greater than the liquid surface-tension forces, resulting in liquid droplets being picked up or entrained in the vapor flow and carried to the condenser

The majority of previous work has been for thermosyphons or for gravity-assisted heat pipes Ofall the limits for heat pipes, the entrainment limit has produced one ofthe largest amounts ofwork, even though much is debated about when, and if, this limit occurs Busse and Kemme (1980) expressed doubt as to whether entrain-ment actually occurs in a capillary-driven heat pipe because the capillary structure would probably retard the growth ofany surface waves In a majority ofcases stud-ied, the wick structure ofthe heat pipe was flooded (i.e., excess liquid), which al-lowed entrainment to occur Additionally, much ofthe work has been an adaptation

to work conducted in the study ofannular two-phase flow, where the onset ofdroplet formation, the rates of entrainment, and the contribution to momentum transfer by the entrained droplets have been investigated in much detail (Langer and Mayinger, 1979; Hewitt, 1979; Nguyen-Chi and Groll, 1981) It is important to note that for ther-mosyphons, the entrainment and flooding limitation is typically the most important factor limiting heat transport (Faghri, 1995)

The most common approach to estimating the entrainment limit in heat pipies is

to use a Weber number criterion Cotter (1967) presented one ofthe first methods to

determine the entrainment limit This method utilized the Weber number, defined as

the ratio of the viscous shear force to the forces resulting from the surface tension, or

We= 2r h,wρv V h2

By relating the vapor velocity and the heat transport capacity to the axial heat flux as

and assuming that to prevent entrainment ofliquid droplets in the vapor flow, the Weber number must be less than unity, the maximum transport capacity based on entrainment can be written as

v

2r c

(16.33)

wherer c is the capillary radius ofthe wick structure However, this assumption typically results in an overestimation ofthe entrainment limit since the axial critical wavelength may be much greater than the width ofthe capillary structure

In addition to the Weber number criterion, several different onset velocity criteria have been proposed for use with this expression These include one by Busse (1973)

U c=



2πσ

(16.34)

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

[1204],(24)

Lines: 689 to 736

———

0.73114pt PgVar

———

Normal Page PgEnds: TEX

[1204],(24)

and another by Rice and Fulford (1987)

U c=



P v d

(16.35)

These relations can be converted into the heat transport limitation due to entrainment

by combining with the continuity equation, which yields



2πσ

(16.36) or

Q e = A vρvλ



(16.37)

respectively, whered is the wire spacing for screen wicks or the groove width for

grooved wicks However, as mentioned earlier, these criteria may overestimate the entrainment limit, due to problems associated with the characteristic dimensions

Tien and Chung (1979) presented correlations for vertical (gravity-assisted) heat pipes This correlation was applied to data reported by Kemme (1976), who expanded the Weber number criterion suggested by Cotter (1967), to include the balancing force term ofbuoyancy, or

v

A



2πσ

(16.38)

Prenger (1984) developed a correlation for textured wall, gravity-assisted heat pipes A model was presented which included both liquid and vapor inertia terms

It was found that for textured wall heat pipes, the liquid inertia term was dominant because the liquid was partially shielded from the vapor flow This fact allowed the vapor inertia term to be neglected and the model to be reduced to

Dpipe

lσ

(16.39)

which correlated well with previous data taken by Prenger and Kemme (1981) While the model presented by Tien and Chung (1979) found that heat flux limited by entrainment or flooding (as a function of the capillary structure), Prenger (1984) found the entrainment limit to be a function of the depth of the liquid layer or flow channel

A review paper by Peterson and Bage (1991) provides a full description of work

in the entrainment area as well as comparisons ofavailable models An interesting item in this comparison is that the models resulted in values for the heat pipe entrain-ment limit which varied by as much as a factor of 27 and demonstrate the level of unresolved issues in predicting the entrainment limit

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

[1205],(25)

Lines: 736 to 758

———

5.51013pt PgVar

———

Normal Page PgEnds: TEX

[1205],(25)

16.2.5 Viscous Limit

When operating at low temperatures, the available vapor (saturation) pressure in the evaporator region may be very small and be ofthe same magnitude as the required pressure gradient to drive the vapor from the evaporator to the condenser In this case, the total vapor pressure will be balanced by opposing viscous forces in the vapor channel Thus, the total vapor pressure within the vapor region may be insufficient

to sustain an increased flow This low-flow condition in the vapor region is referred

to as the viscous limit As the viscous limit occurs at very low vapor pressures, the

viscous limit is most often observed in longer heat pipes when the working fluid used

is near the melting temperature (or during frozen startup conditions) as the saturation pressure ofthe fluid is low

Busse (1973) provided an analytical investigation ofthe viscous limit The model first assumed an isothermal ideal gas for the vapor and that the vapor pressure at the condenser end was equal to zero, which provides the absolute limit for the condenser pressure Using these assumptions, a one-dimensional model ofthe vapor flow as-suming laminar flow conditions was developed and expressed as

Q v= A v r v2h fgρv P v

16µv Leff

(16.40)

whereP v andρv are the vapor pressure and density at the evaporator end ofthe heat pipe The values predicted by this expression were compared with the results ofprevious experimental investigations and were shown to agree well (Busse, 1973)

For cases where the condenser pressure is not selected to be zero, as could be the case when the viscous limit is reached for many conditions, the following expression is used

Q v= A v r v2h fgρv P v

16µv Leff



1−p P v,c22

v



(16.41)

whereP v,cis the vapor pressure in the condenser Busse (1973) noted that the viscous limit could be reached in many cases whenP v,c /P v ∼ 0.3.

To determine whether the viscous limit should be considered as a possible limiting condition, the vapor pressure gradient relative to the vapor pressure in the evaporator may be evaluated In this case, when the pressure gradient is less than one-tenth of the vapor pressure, or∆P v /P v < 0.1, the viscous limit can be assumed not to be

a factor Although this condition can be used to determine the viscous limit during normal operating conditions, during startup conditions from a cold state, the viscous limit given by Busse (1973) will probably remain the limiting condition As noted earlier, the viscous limit does not represent a failure condition In the case where the heat input exceeds the heat input determined from the viscous limit, this results in the heat pipe operating at a higher temperature with a corresponding increase in the saturation vapor pressure However, this condition typically is associated with the heat pipe transitioning to being sonic limited, as discussed in the following section

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

[1206],(26)

Lines: 758 to 766

———

1.927pt PgVar

———

Normal Page PgEnds: TEX

[1206],(26)

16.2.6 Sonic Limit

The sonic limit is typically experienced in liquid metal heat pipes during startup or low-temperature operation due to the associated very low vapor densities in this con-dition This may result in choked, or sonic, vapor flow For most heat pipes operating

at room temperature or cryogenic temperatures, the sonic limit is typically not a fac-tor, except in the case ofvery small vapor channel diameters With the increased va-por velocities, inertial, or dynamic, pressure effects must be included It is imva-portant

to note that in cases where inertial effect of the vapor flow are significant, the heat pipe may no longer operate in a nearly isothermal case, resulting in a significantly increased temperature gradient along the heat pipe In cases ofheat pipe operation where the inertial effects of the vapor flow must be included, an analogy between heat pipe operation and compressible flow in a converging–diverging nozzle can be made In a converging–diverging nozzle, the mass flow rate is constant and the vapor velocity varies due to the varying cross-sectional area However, in heat pipes, the area

is typically constant and the vapor velocity varies due to mass addition (evaporation) and mass rejection (condensation) along the heat pipe

As in nozzle flow, decreased outlet (back) pressure, or in the case ofheat pipes, condenser temperatures, results in a decrease in the evaporator temperature until the sonic limit is reached Any further increase in the heat rejection rate does not reduce the evaporator temperature or the maximum heat transfer capability but only reduces the condenser temperature due to the existence ofchoked flow Figure 16.12 illustrates the relationship between the vapor temperature along a heat pipe with

Figure 16.12 Temperature as a function of axial position (From Dunn and Reay, 1982, with permission.)

Ngày đăng: 05/07/2014, 16:20