Inl. J. Heor Mos Tramfir Vol. 28. No. 3, pp 553-562, 1985 0017-9310/85$3.00+0.

00
PrmtedinGreatBrltam Pergzimon
PrrssLtd.

A method of analysis for heat pipe heat exchangers

B. J. HUANG and J. T. TSUEI
Department of Mechanical Engineering, National Taiwan University, Taipei, Taiwan 107

(Received 17 January and in revised/brm 7 August 1984)

Abstract -A method of analysis for the thermal performance of heat pipe heat exchangers based on the
conductance mode! wasdeveloped in the present study. In the analysis the specific heat conductance of the heat
pipe was obtained from a performance test of a single heat pipe described in the present paper and the well-
known universal correlations were used to calculate the convective heat transfer coefficients in tube banks. A
computer program based on the finite difference equations of the model was then developed to calculate the
thermal performance of heat pipe heat exchanger. The analysis was finally validated by an experiment and
shown to be applicable in engineering applications.

INTRODUCTION thermosyphons were not exactly identical with those

described by the empirical correlations and the fouling
HEAT EXCHANGERS made of heat pipes have now
inside the thermosyphons was not taken into account in
become one of the most effective and economic devices
Lee’s model. Second, there was insufficient quality
for the recovery of waste heat energy. Although the
control in making the heat pipes or thermosyphons
thermal performance characteristics of a single heat
so that the thermal performance of each one was
pipe have been extensively studied and clearly
different.
understood during the past twenty years, the study of
To develop a method for the prediction of overall
the prediction of the overall performance of heat
thermal performance of a heat pipe heat exchanger, a
exchangers using a bank of heat pipes as the heat
modified method of analysis is therefore proposed in
transfer elements appears to be very limited. A typical
the present study. In this method, the thermal
theoretical study but without experimental verifi-
performance of a single heat pipe or thermosyphon is
cations on the performance calculation of heat pipe
experimentally, instead of theoretically, determined by
heat exchanger was recently carried out by Amode and
a testing method. With the testing results and the
Feldman [ 11. In this study, the basic theories for heat
universal empirical correlation of convective heat
pipes and conventional heat exchangers were employed
transfer for tube banks, the overall thermal perform-
to perform a pure theoretical calculation.
ance of the heat exchanger can be calculated by use of
More recently, Lee and Bedrossian [2] developed a
the proposed method of analysis.
simple analytical model to study the characteristics of
heat exchangers using heat pipes or thermosyphons. In
this model, the thermal resistances inside the heat pipes
PERFORMANCE TEST OF A
or thermosyphons were theoretically calculated by the
SINGLE HEAT PIPE
well-knowncorrelations for the boiling and condensing
processes. However, the two constants in the empirical Heat pipes without fins and using water as the
correlation used to calculate the convective heat working fluid were carefully made in the laboratory for
transfer coefficients outside the heat pipes or the present study. The specifications of the heat pipes
thermosyphons had to be determined by a curve fitting are listed in Table 1. Since the present experiment was
to the experimental results ofan actual heat exchanger. carried out in a short time period (within a month) as
It is well known that the universal empirical correlation compared to commercial operation (usually for several
for convective heat transfer in tube banks is of the form years) and the major concern of this study was on the
thermal performance analysis, the deterioration in
Nu = C Re” Prli3 (1) thermal peiformance due to the generation of non-
where the universal constants, C and n, depend on the condensible gases caused by the use of a carbon steel-
geometric arrangement of the tube bank and have been water combination was minimal and could be
universally tabulated according to a great deal of neglected.
experimental results [S]. For the thermosyphon heat The thermal performance of a heat pipe was
exchanger used by Lee and Bedrossian, C and n should measured by the testing equipment as shown in Fig. 1.
be0.415 and0.581, respectively. Thecurvefitting by the As the heat pipe was inserted into the water jacket, it
experiment of a thermosyphon heat exchanger in [2], was automatically divided into the evaporation and the
however, gave 0.009 and 1.050 for C and II. Such a large condensation sections with equal lengths but with
discrepancy could be attributed to two facts. First, the negligible adiabatic section. This design was adopted in
actual boiling and condensing processes in the order to meet thegeneral requirement in some practical

mm 28:sc 553
554 8. J. HUANG and J. T. TSIJEI

NOMENCLATURE

total area of heat pipe heat exchanger [m’] Q total energy transfer rate of heat exchanger
surface area of condensing section of a heat [WI
pipe Cm’1 4P
energy transfer rate of a heat pipe [W]
surface area of evaporating section of a qpj energy transfer rate in the jth row of heat
heat pipe [m2] pipes [W J
specific heat of cold air flow 4 overall heat transfer resistance in the jth
[KJ kg- ' "C- '1 section PC W-‘]
specific heat of hot air flow Re Reynolds number in tube banks, defined in
[KJ kg-’ “C-‘1 equation (11)
characteristic diameter of heat pipe heat T temperatures c”C]
exchanger [m] U characteristic velocity in tube banks
heat pipe O.D. [m] [m s- ‘1.
thermal conductivity of air [W m-l “C-‘1
Greek symbols
convective heat transfer coefficient
p density of air [kg m-“1
CWm -2 gc-‘l
p viscosity of air [N s m-‘1,
mass flowrate of cold air [kg s- ‘1
M, mass flowrate of hot air [kg s- ‘1 Subscripts
sj number of heat pipes injth row C condenser side ; cold air
Nu Nusselt number in tube banks, defined in h evaporator side; hot air
equation (11) w outside wall surface of heat pipe
Pr Prandtl number j section notation of heat exchanger.

applications for which the adiabatic sections ofthe heat can be applied to commercial heat pipes, it was not
pipes in the heat pipe heat exchanger had to be as small attempted in the present study to measure the
as possible. The condensing section of the heat pipe was temperatures or other properties in the interior of the
then sealed by use of a flange and o-rings in the water heat pipes. Instead, only temperatures over the outside
jacket through which a cooling water was allowed to surface of the heat pipe were measured using eleven T-
pass. In the evaporating section of the heat pipe, a type thermocouples which were fixed equally spaced
2.5 kW electric heating wire was uniformly wrapped onto the pipe surface and read by an Omega 2176A
over the pipe surface as the heating element. The rate of recorder to within f O.YC. Temperature measure-
heating was controlled by a power transformer. To ments were also made at the inlet and outlet of the water
eliminate the heat losses, the tested heat pipe including jacket. To determine the rate of energy removed in the
the water jacket was carefully insulated by 6 cm thick water jacket, the mass flowrate ofthe cooling water was
calcium silicate. Therefore, the overall energy transfer also measured by a rotameter which was calibrated to
rate through the heat pipe can be measured either by within rfi 5%.
measuring the rate of energy added by the electric Figure 2 shows the temperature distributions vs
heater or by measuring the rate of energy removed by energy transfer rates obtained from one of the test runs
the water jacket. These two measurements also for a performance test of a single heat pipe. At this stage,
provided a check for the insulation condition during it is very important to derive a method to correlate the
the experiment. It was observed in the present testing results. As a first-order approximation, it is
experiments that the two values obtained from the assumed that the longitudinal heat conduction in pipe
above two measurements all coincided within iS% wall and wick material is negligible. Thus, there are Eve
deviations. thermal resistances for the transfer of energy from the
As the present method of testing for the thermal hot to the cold side of the heat pipe as shown in Fig. 3,
performance ofasingle heat pipe is non-destructive and where R,., and R,., are the thermal resistances of the
pipe wall at the evaporating and the condensing
Table I. Specifications of heat pipes sections, respectively; R,,,and R,,, are the thermal
resistances in the wick materials at the evaporating and
Effective length 610 mm
Outside diameter 33.7 mm the condensing sections, respectively and R, is the
Wall thickness 1.6mm overall thermal resistance in the vapor core region
Wall materiak carbon steel including interphase condensing and evaporating
Wick structure wrapped screen (I 5 layers) resistances.
Wick material IOO-mesh bronze
For heat pipes operating below the sonic limit, the
Working fluid distilled water (I 15 g)
thermal resistance in the vapor core, R,,can be ignored
 A method of analysis for heat pipe heat exchangers 555

rHEA7 PIPE

TRAN

PUMP UNIT

FIG. 1. Testing equipment of a single heat pipe.

in comparison with the others. Therefore it can be pipe, q,,,can be expressed as

theoretically shown in accordance with the heat pipe
theory [4] that the vapor core temperature, T,, can be up = (U&,,(T,,- T,,) (3)
expressed in terms of the lengths and the mean surface
temperatures of the condensing and the evaporating where (UA), is the specific heat conductance from the
sections surface of the evaporating section to that of the
condensing section. Therefore, the specific conduc-
tance of a heat pipe can be experimentally determined
according to the definition shown in equation (3).
where T,, and Twh are the mean temperatures on the Since the temperaturedistributions on thesurfaces of
outside surface of the condensing and the evaporating condensing and evaporating sections are generally very
sections, respectively, and L, and L, are the lengths of uniform, the mean temperatures on the outside surfaces
condensing and evaporating sections. If L, and L, are of the condensing and evaporating sections, T,, and
set equal as in the present experiment, the vapor core Twhrcanbe experimentally determined by averaging the
temperature is just the arithmetic mean of Twhand T,,. thermocouple readings on them. However the readings
Furthermore, the overall energy transfer rate in a heat for thepositionsat which theaxial wall heat conduction

I , t I 8 t f

t
400 -

360-
320-

280-
240-
200-

160”
120-

80s
40-
01
0 0.1 0.2 0.3
I

04
1

Q5 06 0.7
1

0.8 0.9
1
1.0

POSJTJON, DIMENSIONLESS (X/L)

Frc. 2. Wall temperature distributions vs energy transfer rates.

556 B. J. HUANCand J. T. TSUEI

HEAT OUT
In the present study, five out of 32 heat pipes made in
tttttt t
/PIPE WALL the laboratory were randomly selected to be separately
-SCREEN
tested according to the above testing method. It can be
seen from Fig. 4 that the specific conductance of each
heat pipe is approximately independent of the vapor
core temperature as can be expected from the heat pipe
/ti ttttt theory. Therefore, combining all the data, the mean
1 (Twh) heat conductance for this batch of heat pipes was
determined as 3.36 W “C- ’ with a standard deviation of
+ 0.59 W “C- I. The standard deviation will apparently
decrease with improving quality control. When the
quality control is excellent, as in commercial mass
production, this sampling test can be omitted and a
single performance test for a heat pipe of the same kind
TWh TWC
using the same procedure can directly give the heat
FIG.3. Thermal resistances network of a heat pipe. conductance.
It also can be seen that the heat transfer behavior of
the heat pipes made in the present study is not good.
toward the condensing section is not negligible cannot This is attributed to the fact that no other means which
be included. Therefore, care should be taken in can reduce the thermal resistance of the wick material
averaging the thermocouple readings on the pipe were employed in the present heat pipes. In addition, a
surface to preclude these points. tight contact between the wrapped bronze screen and
For the heat pipes made in the present study, the the pipe wall was produced by the elastic force induced
largest deviation of the surface temperature in the merely by the screen itself when carefully pressing the
evaporating section, precluding the point nearest to the wrapped screen into the pipe. Since the major purpose
pipe mid-point, was within + 12°C for temperatures ofthe present study is to develop a new method for the
around 360°C and even much less for temperatures thermal performance analysis of heat pipe heat
below this, as can be seen from Fig. 2. Thus the averaged exchanger, the study on the fabricating techniques
surface temperatures could give reasonably good which can produce highest efficiency heat pipes was
approximations. skipped over.
In addition, poor quality control in making the heat
pipes can lead to small deviations in heat conductance
THERMAL ANALYSIS OF
between each heat pipe. This necessitates a certain
HEAT PIPE HEAT EXCHANGER
amount oftesting of samples to determine a mean value
for the heat conductance which can statistically A finite difference method based on the conductance
represent the averaged thermal performance of the model is employed in the present study to analyze the
batch of heat pipes to be used to make a heat pipe heat overall heat transfer of the heat pipe heat exchanger.
exchanger. Assume that the heat pipe heat exchanger can be

10 I, I1 I I I I I I1 I I1 1 q
TEST SAMPLE NO. 1 - A
9 2 -m
a- 3 -v
;_
2 I & -0
5 -A
9
o,u 6-
UI
? 0.59 watt Ic
l-0
qi
A A A
Ya2 f+-A A
2 3 3.36 V
x &)A V-&+G m
k -.v A
n 6
om 0m0 n
z
% 2-

VAPOR CORE TEMPERATURE T, ;C

FIG.4. Thermal performance test results of heat pipes.

 A method of analysis for heat pipe heat exchangers 557

divided into N sections along the flow direction and the flows across each section can be written as
each section consists of a row of heat pipes of the same
Tc,j- Tc,j- I = n,(lIRJ(Gj- TJMcCc (7)
kind, as shown in Fig. 5. Neglecting the heat transfer
across the partition and the heat loss to the ambient, the x,j- q,j- 1 = -n,(lfRj)(Rj- T,)fM,C, (8)
total energy transfer in the jth section can be expressed
where M, and M, are the mass flowrates of the hot and
as
the cold flow, respectively, both are taken as positive for
4pj = n,llIRj)(Thj- Tj) (4) parallel-flow design and M, is taken as negative for
counter-flow design. To solve the temperature
where nj is the number of heat pipes in thejth section, Thj
distributions, the iteration method with relaxation
and TCjare the mean fluid temperatures of the hot and
factor is employed [S] with the follbwing equations
the cold flows, respectively, in the jth section, which is
which are derived from equations (7) and (8)
defined as
1 K,
xj=(Th,j+T,,j-,)/2 and T;j=(T,,j+T,,j-,)/2. T”,j=- T;,j_ I - -
l+K,j l+K,j
Rj in equation (4) represents the overall heat transfer
~[T~,j_1-~,j-C,j-1I (9)
resistance from the hot flow to the cold flow for a heat
pipe in the jth section and can be expressed by the 1 K.
following equation, ignoring fouling resistance in the T;,j = -T:,j-,-~
l+K,j l+K,j
pipe wall
CC’,j-l-TE,j-‘G,j-11 (10)
Rj = (lIh,A,)j+(lI~~),+(llh,A,)j (5)
where the superscripts, n and o, represent the new and
where h, and h, are the convective heat transfer the old value, respectively, during the iteration and
coefficients outside the heat pipe wall in the hot and the
cold flow, respectively; A, and A, are the outside K,j = nj/2M,C,,Rj and Kcj = nj/2M,C,Rj
surface areas of a heat pipe in the evaporating and the
The iteration procedure is as shown in Fig. 6. To
condensing sections. Therefore the overall energy
calculate the convective heat transfer coefficients in
transfer rate delivered by the whole heat pipe heat
tube banks, h, and h,, the universal correlation
exchanger can be calculated by an integration over the
obtained by Whitaker [6] is used
N sections
Nu .= f(0.5Re1”+ 0.2Re2/3)Pri/3(p/p,) (11)
Q = 2 qpj (6)
j= 1 where NM and Re are defined as
It is necessary to calculate the temperature
distributions in the hot and cold flows before Nu = (hDdk,)c,/( 1 -E,) (12)
calculating the overall energy transfer rate using
and
equations (4) through (6). By applying the energy
balance to the jth section, the temperature changes of Re = (~D,~/&d(l - s,). (13)

TOP VIEW

j;l

j-l j -1
“p
I II
I HOT - ;
1
- FLOW - 1
-I
I
-,
I
7
--I
COLD - )
FLOW - ,
-A,*
I I
i<,,>
_

j-th ROW 4
nj /he Ac nj/(UA)p nj/h,Ac
T hl - 44.w -7 CJ
Twhj Twcj

FIG. 5. Schematic diagram of heat pipe heat exchanger.

558 B. J. HUANG and J. T. TSUEI

INPUT desqn data on the number of rows in the heat pipe heat exchanger
(UA), ,N , n, zL, L, > c71.
A computer program based on the above equations
D, fluld properties ,etc.
and the iteration process as shown in Fig. 6 was
developed in the present study to calculate the thermal
performance of a heat pipe heat exchanger.

EXPERIMENTS

An experiment for a heat pipe heat exchanger was

SET temperature dlstribut 10”s T,,, and T,,, performed in the present investigation to verify the
I above analysis. The schematic diagram of the

I experimental setup is as shown in Fig. 7. The

specifications of the heat pipe exchanger are listed in
DETERMINE new temperatures’ Th,, ,T,,, , Table 2.
The heat pipe heat exchanger was designed to be
TM, Jwc, by eqns (L) through (11) operated in counter-flow conditions with air as the
energy exchange fluids in the present study, To provide
1 a hot air flow with temperatures up to 300°C for the
I
CHECK convergences of flow temperatures experiment, a 30 kW electric heater was installed in the
NO
rectangular hot air duct. The air flowrates in both hot
Th,, and T,,,
and cold air ducts were measured in two lOO-mm
YES
circular portions of the air ducts in which Pitot tubes

CALCULATE total energy transfer rate Q

Table 2. Specifications of heat pipe heat exchanger

Number of rows, N 4
FIG. 6. Flow chart of the iteration process in thermal analysis. Number of heat pipes in each row, n, 8
Total number of heat pipes 32
Condenser length of heat pipes, L, 305 mm
Evaporator length of heat pipes, L, 305 mm
D,is the characteristic diameter defined as 1.5D (D is the Pipe arrangement staggered
pipe O.D.) for tube banks and E, is the void fraction of Pipe pitch : longitudinal 44.1 mm
the heat pipe heat exchanger, u is the characteristic flow transverse 38.7 mm
Cross-section : for hot flow 305 x 380 mm’
velocity which is usually taken as the maximum
for cold flow 305 x 380 mm*
velocity in the tube bank and f is a parameter depending

FLOWRATE MEASURING
DEVICE

STRAIGHTENER TO AMBIENl

HEATER

r ^“‘“‘TE MEASURING

FIG. 7. Testing facilities of heat pipe heat exchanger

 A method of analysis for heat pipe heat exchangers 559

Table 3. Comparison ofexperiment with theory for total energy transfer rate of the heat pipe heat exchanger

Energy transfer rate, Q

Hot air Cold air (W

Test Flowrate Inlet temp., L$O Flowrate Inlet temp., &

run No. (SCMM) (“C) (SCMM) (“C) Experiment Theory

12.9 244.1 13.2 25.1 6290 5920

12.8 240.0 13.0 25.2 6050 5770
12.6 238.4 13.2 24.6 5870 5750
12.2 276.2 12.7 24.8 6750 6520
9.10 297.1 10.4 28.8 5710 6020
9.10 295.7 10.3 28.7 5640 5990
12.3 274.4 7.20 32.0 5660 5480
8 11.8 296.6 6.80 34.7 5720 5710
9 11.5 261.6 7.00 31.7 4720 5080
10 11.3 298.5 7.20 33.0 5700 5820
11 8.20 299.0 7.10 31.3 4880 5320
12 8.20 300.7 7.10 31.4 4940 5350

and precision pressure gauges were used to measure the RESULTS AND DISCUSSION
velocity distribution at given cross sections. The
velocity was measured at 10 points along the radial Some test runs using the above heat pipe heat
direction in the present experiments to obtain the exchanger equipment were carried out for various air
velocity distribution. By integrating the velocity flow rates and temperatures in the present experiment.
distribution then multiplying by the air density at the Theoretical calculations based on the analytical model
same cross-section, the air mass flowrate in each duct described previously were also carried out for
was determined. The experimental uncertainties were comparisons. The results are listed in Table 3. It can be
estimated to be within f 5%. seen that the errors caused by the analysis in the total
Temperature measurements were made by T-type energy transfer rates are all within + 10% which is
thermocouples at the inlet and the oulet of the hot and acceptable in engineering applications. It is also shown
the cold air flows and all over the heat pipe heat from Table 4 that the measured temperature
exchanger. The mean outside wall temperatures of the distributions ofthe air flows in the heat exchanger are in
condensing and the evaporating sections of the heat fairly good agreement with the analytical results.
pipes at thejth row, TWcjand Twhj, were measured by 16 Larger errors in the predictions of the pipe wall
thermocouples. To do this, two pipes were selected at temperatures occur in the present experiment, as can be
each row and four thermocouples were separately fixed seen from Table 5. This probably resulted from the
at the middle points of the condensing and the contact resistances between the pipe wall and the
evaporating sections of the two heat pipes. The mean temperature probes which were introduced during the
wall temperature of the condensing or evaporating installation of the probes.
section at the jth row, Twcj or Twhj,then was obtained by In fact, the convective heat transfer in tube banks
averaging the two temperatures on the two heat pipes. increases along with the penetrating of the air how into
To measure the air temperatures between two the tube banks due to the transition of the flow pattern
adjacent rows ofheat pipes in the hot or cold channel of from laminar to turbulent [8]. Therefore, the use of the
the heat exchanger, Th,j or K,j, three thermocouples mean convective heat transfer coefficients in tube banks
were installed at each measuring cross-section and the in the present analysis could lead to the small errors in
average readings gave the bulk air temperatures. There temperature predictions as shown in Tables 4 and 5.
were five measuring cross-sections in each channel and From the experiment, we verified that the method of
30 thermocouples were used in this measurement. analysis for the thermal performance of heat pipe heat
Besides, two temperature measurements were also exchanger proposed in the present study is validated.
made at the velocity measuring ducts to obtain the air Although the present experiment was performed using
densities for determining the air mass flow rates. heat pipes without fins for the sake of simplicity, the
To process the 38 thermocouple readings and the present method of analysis still can be applied to heat
related calculations in the present experiment, a Kaye pipes with fins as long as the convective heat transfer
Digistrip II data acquisition and microcomputer coefficients for finned-tube banks are used.
system was used. The maximum uncertainties in Finally, it should be noted that, to ensure good
measuring the total rates of energy transfer of the heat accuracy, the present method of analysis is not valid
pipe heat exchanger were estimated to be within + 5%. when the operating conditions of any heat pipe in the
560 ES.J. HUANG and J. T. TSUEI

Table 4. Air temperature distributions

Air temp. in hot flow, Th., Air temp. in cold how, T,,,
Test (“CJ (‘C)
run No. j=O 1 2 3 4 j=O 1 2 3 4
~.
Experiment 244.1 236.0 223.2 213.8 200.7 49.7 42.2 35.7 29.8 25.1
I 25.1
Theory 244.1 233.5 223.1 213.0 203.2 48.3 42.2 36.4 30.7
Experiment 239.5 232.1 219.6 210.5 197.9 49.1 42.2 36.1 30.0 25.2
2
Theory 239.5 229.1 219.1 209.3 199.7 48.1 42.1 36.3 30.7 25.2
Experiment 238.4 231.1 218.9 210.0 197.5 47.5 40.8 35.1 29.2 24.6
3
Theory 238.4 228.0 211.8 207.9 198.2 47. i 41.2 35.5 30.0 24.6
Experiment 216.2 266.8 250.4 239.2 223.6 52.2 44.0 36.8 30.0 24.8
4
Theory 216.2 262.8 249.8 237.3 225.2 51.4 44.4 37.7 31.1 24.8
Experiment 297. I 288.7 267.0 254.1 235.2 57.3 47.7 39.5 33.2 28.8
5
Theory 297.1 279.6 262.9 246.9 231.6 59.0 50.9 43.2 35.9 28.8
Experiment 295.7 281.7 266.3 253.8 234.7 57.1 47.9 39.6 33.4 28.7
6
Theory 295.7 278.4 261.8 245.9 230.8 59.0 50.9 43.2 35.8 28.7
Experiment 274.4 266.9 253.1 244.4 230.7 73.0 61.3 47.5 37.5 32.0
7
Theory 274.4 263.6 253.0 242.5 232.0 71.8 61.7 51.7 41.8 32.0
Experiment 296.6 292.1 275.0 265.4 248.9 79.2 67.0 51.4 40.4 34.7
8
Theory 296.6 284.5 272.5 260.6 248.9 79.2 67.9 56.7 45.6 34.7
Experiment 261.6 258.1 244.6 237.2 223.8 67.2 57.3 45.2 36.4 31.7
9
Theory 261.6 251.2 241.0 230.9 220.8 70.0 60.3 50.7 41.2 31.7
Experiment 298.5 293.4 275.7 266.0 248.9 74.9 63.2 48.6 38.3 33.0
10
Theory 298.5 285.5 212.8 260.2 247.8 75.8 64.9 54.1 43.5 33.0
~x~riment 299.0 291.3 270.8 258.8 240.2 67.2 56.7 44.3 35.7 31.3
11
Theory 299.0 282.1 265.8 250.0 134.8 70.5 60.1 50.2 40.6 31.3
Experiment 300.7 292.7 272.0 259.6 241.0 61.7 57.1 44.6 36.0 31.4
12
Theory 300.7 283.6 267.2 251.3 235.9 10.7 60.4 so.4 40.7 31.4
Average error 0.0 6.2 2.5 4.8 1.2 -0.7 - 1.6 - 3.5 -3.1 0
-~ -_.- -

heat exchanger are beyond the testing range in the a wider range of operating conditions, as near the heat
performance test of a single heat pipe as described pipe operating limits as possible.
previously. Therefore, checks at any instant during the The present method of analysis can be applied to
analysis are always required and it is important to simulate the thermal performance of the heat pipe heat
perform the performance test for a single heat pipe over exchanger used in the experiment beycnd the testing

Table 5. J?ipe wall tem~rature distributions

Wall temp. in hot air side, Twhj Wall temp. in cold air side, Twcj
Test i”Cl (“C)
run No. j=l 2 3 4 j= 1 2 3 4
-- - ___~ -- ..-..._~-
Experiment 169.2 160.3 144.6 140.5 100.2 92.6 83.2 80.0
Theory 165.6 156.6 147.9 139.4 108.0 100.7 93.6 86.7
Experiment 161.2 159.3 143.8 140.0 97.8 90.8 83.0 80.0
Theory 162.9 154.1 145.6 137.4 106.8 99.6 92.6 85.8
Experiment 173.8 160.6 144.4 140.8 95.2 88.0 79.6 78.9
Theory 161.2 152.5 144.0 135.7 105.3 98.2 91.2 84.4
Experiment 194.Q 1go.4 160.6 156.8 110.4 99.7 88.1 87.0
Theory 183.5 172.6 162.0 151.7 119.5 110.7 102.2 94.0
.5 Experiment 208.3 201.6 176.3 170.3 113.8 101.8 88.0 83.0
Theory 190.9 177.3 i 64.3 151.9 130.7 119.9 109.6 99.8
6 Experiment 207.8 201.4 176.0 170.2 113.2 101.0 87.3 82.4
Theory 190.4 176.8 164.0 151.6 130.6 119.8 109.6 99.7
7 Experiment 208.1 198.8 179.4 176.6 139.5 130.9 117.0 116.7
Theory 199.6 189.2 178.8 168.6 147.8 137.9 128.0 118.3
8 Experiment 223.5 216.9 193.4 190.8 150.4 141.8 126.0 125.8
Theory 215.5 203.8 192.2 180.7 161.3 150.2 139.2 128.4
9 Experiment 202.0 196.2 176.2 174.4 126.6 118.4 108.3 108.4
Theory 190.0 180.0 170.0 160.2 141.9 132.4 122.9 113.6
10 Experiment 223.0 216.7 192.8 189.8 145.2 134.9 120.3 120.2
Theory 213.1 201.0 189.0 171.2 157.6 146.3 135.3 124.4
11 Experiment 214.7 208.4 183.2 178.8 132.6 121.5 107.0 105.3
Theory 200.2 186.0 172.3 159.1 147.9 135.6 123.8 112.3
12 Experiment 215.8 209.2 183.8 179.4 133.6 122.4 107.6 105.8
Theory 201.2 186.9 173.1 159.8 148.5 136.2 124.3 112.8
Average error 11.1 14.4 ,4.3 11.3 - 12.3 -- 12.0 - 14.7 -7.2
 A method of analysis for heat pipe heat exchangers 561

VOLUMETRIC AIR FLOW RATE > SCMM

FIG. 8. Total energy transfer rate at various inlet temperatures of hot air.

conditions. Figure 8 shows the total energy transfer defined as in equations (12) and (13) and the ratio, Y,is
rates at various inlet temperatures of hot air, T&,, and defined as I”= Re,JRe,. The overall heat transfer
equal vofumetric air flow rates, V, = V,, in the hot and coeffkient, U, is defined as
the cold flows. For different combinations of Aowrates,
U=Q/A(LMTD) (14)
the overall heat transfer coefficient of the heat pipe heat
exchanger, U, can be determined in terms of the where (LMTD) is the log mean air temperature
Reynolds number of the cold air flow, I$, and the ratio difference of the heat exchanger, A is the total heat
oftheReynoldsnumbersofthetwoairflows,r,ascan be transfer area of the heat exchanger. It can be seen from
seen in Fig. 9. The Reynolds numbers, Re, and Re,,, are Figs 8 and 9 that the energy transfer rate as well as the

1 I111111 I ,
IO2 103 lOL lo5

REYNOLDS NUMBER OF COLD AIR FLOW, Re,

FIG. 9. Overall heat transfer coefficients of heat pipe heat exchanger at various operating conditions.
562 B. J. HUANG and J. T. TSUEI

overall heat transfer coefficient increases with Acknowledgement-The present study was supported by the
increasing flowrate but tends to reach a constant. This Energy Research Laboratory at the Industrial Technology
Research Institute of the Republic of China.
is due to the fact that the convective heat transfer
resistances in tube banks decreases with increasing flow
REFERENCES
rate but the specific heat conductance of the heat pipes
still remains nearly constant over a wide range of 1. J. 0. Amode and K. T. Feldman, Preliminary analysis of
operating temperatures. heat pipe heat exchangers for heat recovery, ASME Paper
No. 75-WA/HT-36 (1976).
2. Y. Lee and A. Bedrossian, The characteristics of heat
exchangers using heat pipes or thermosyphons, Znt. J. Heat
Mass Transfer 21. 221-229 (1978).
CONCLUSIONS J. P. Holman, Heat Transfe;, p. 222. McGraw-Hill, New
York (1976).
A method of analysis based on the conductance
S. W. Chi, Heat Pipe Theory and Practice. McGraw-Hill,
model was developed in the present study. In the New York (1976).
analysis, the specific heat conductance of heat pipe was S. V. Patankar, Numerical Heat Transfer and Fluid Flow.
obtained from a thermal performance test for a single Hemisphere, New York (1975).
S. Whitaker, Forced convection heat transfer correlations
heat pipe and the universal correlations for the
for flow in pipes, past flat plates, single cylinders, single
convective heat transfer coefficients in tube banks were spheres, and for flow in packed beds and tube bundles, J.
used. A computer program using the finite difference AlChE l&361-371 (1972).
equations of the model was also developed to calculate I. W. M. Kays, Compact Heat Exchanger, 2nd ed., p. 128.
the thermal performance of heat pipe heat exchanger McGraw-Hill, New York (1964).
8. W. M. Kays, A. L. London and R. K. Lo, Heat-transfer and
and the present method of analysis has been validated
friction characteristics for gas flow normal to tube banks-
experimentally and shown to be applicable in use of a transient test technique. Trans. ASME 76,389-396
engineering applications. (1954).

UNE METHODE D’ANALYSE DES ECHANGEURS CALODUCS

R&u&-On dtveloppe une m6thode d’analyse des performances thermiques des Bchangeurs<aloducs ba&.e
sur le modtle de conductance. Dans l’analyse, la conductance thermique specifique du caloduc est obtenue B
partir d’essais d’un caloduc unique d&it dans cet article et les formules universelles classiques sont utilisies
pour calculer les coefficients de convection dans un arrangement de tubes. Un programme de calcul bast sur les
Equations aux diffkrences finies du modile est d&velopp& pour calculer la performance thermique de
l’tchangeur-caloduc. L’analyse est finalement valid&e par une exp&imentation et elle est montr6 applicable
aux problZmes de l’ingknierie.

EIN VERFAHREN ZUR UNTERSUCHUNG VON WARMEROHR-WARMETAUSCHERN

ZusammenfassungpEs wurde ein Verfahren zur Untersuchung des thermischen Verhaltens von WLrmerohr-
Wlrmetauschern entwickelt, das auf einem Leitungs-Model1 basiert. Die spezifische Wgrmeleitfiihigkeit des
Wlrmerohreswurdemit HilfeeinesVersuchsamEinzelrohrermittelt.Zur Berechnungder Wlrmelbergangs-
Koeffizienten bei der Konvektion in Rohrbhndeln wurden die wohlbekannten universellen Korrelationen
verwendet. Damit wurde nach dem Finite-Differenzen-Verfahren ein Computer-Programm zur Berechnung
des thermischen Verhaltens von WBrmerohr-Wlrmetauschern entwickelt. Diese theoretische Untersuchung
wurde schliefilich experimentell iiberpriift und erwies sich fiir ingenieurmal3ige Anwendungen als geeignet.

MET04 AHAJIM3A TEflJlOO6MEHHMKA HA TEIlnOBbIX TPYGAX

AwoTauwa-B pd6OTe pass&%-rOcHoBaHHbIi? Ha Monenll npoBons~ocTH MeTOn aHaJIH3a TennoBoro

perllMa Ten,IOO6MeHHHKa Ha Ten,IOBbIX Tpy6ax. 3+$eKTHBHaB TennO”pOBOflHOCTb TenJIOBOii Tpy6bt
“O,IygeHa “a OCHOBaHAH3KCnepHMeHTa,IbHbIX lrCC,IeAOBaHHiiOTnenbHOii Ten,IOBOfi Tpy6bI, OnllCaHHOfi
B pa6oTe, a ana pacqeTa K03@+iuHeHTa KoHBeKTIiBHoro TenJIOO6MeHa B nyqKax ~py6 ncnonb308anHcb
Xopomo H3BeCrHbIe yHHBepCa,IbHbIe ypaBHeHHB. Paspa6oTaHa npOrpaMMa C’IeTa, OCHOBaHHall “a
ypaBHeHAnx B KOHegHbIx pa3HOCTBX, &“n On~fle,IeH&III TennOBOfi XapaKTepHCTHKA TenJlOO6MeHHHKa
Ha TennoBbIx rpy6ax. B pe3ynbTaTe aHann 060CHOBbIBaeTCR naHHbIMH 3KcnepeMeHTa. nOKa3aHbI
BOSMO~HOCTATeXHBYeCKOrOnpIiMeHeHHa.