International Journal of Heat and Mass Transfer 70 (2014) 811–818

Contents lists available at ScienceDirect

International Journal of Heat and Mass Transfer

journal homepage: www.elsevier.com/locate/ijhmt

Heat transfer of laminar oscillating flow in finned heat exchanger of

pulse tube refrigerator
K. Tang, J. Yu, T. Jin ⇑, Y.P. Wang, W.T. Tang, Z.H. Gan
Institute of Refrigeration and Cryogenics, Zhejiang University, Hangzhou 310027, PR China

a r t i c l e i n f o a b s t r a c t

Article history: In order to characterize the heat transfer of heat exchangers employed by pulse tube refrigerators, an
Received 1 September 2012 experimental apparatus was built to investigate the heat transfer performance of a water-cooled finned
Received in revised form 6 June 2013 heat exchanger operating in laminar oscillating flow. The test results summarized into the Nusselt
Accepted 23 November 2013
number with respect to the maximum Reynolds number and the Valensi number were presented. The
Available online 18 December 2013
increases in the maximum Reynolds number and the Valensi number both lead to a rise in the Nusselt
number. The comparisons of the experimental results and the available typical correlations were
Keywords:
conducted and discussed. A new correlation of the Nusselt number to the maximum Reynolds number
Oscillating flow
Heat transfer
and the Valensi number was proposed, with which the results have a maximum deviation of 6.3%
Heat exchanger compared with the experimental values.
Nusselt number Ó 2013 Elsevier Ltd. All rights reserved.
Pulse tube refrigerator

1. Introduction carried out the measurement of the complex Nusselt number. A

The heat transfer in an oscillating flow is attractive for many
heat engines and refrigerators, such as Stirling engines, thermoa-
coustic engines and pulse tube refrigerators. Much effort has been
made for better understanding the heat transfer characteristics in
the oscillating flow, and then for better designing the heat
exchangers employed in the abovementioned heat engines and
refrigerators. In contrast with a unidirectional flow, the cyclic
reversion of fluid velocity direction of the oscillating flow, together
with the influence of oscillating pressure on the fluid temperature,
brings a big challenge for characterizing the oscillating-flow heat
transfer.
A review of the related literatures shows that two typical strat-
egies have been used for characterizing the heat transfer of the
oscillating flow, of which one emphasized the cyclically varying
temperature profiles in the fluid and the oscillating heat flux at
the wall surface [1–8]. Complex Nusselt numbers were used to cor-
relate the oscillating fluid-to-wall temperature differences and the
oscillating heat flux. Nika et al. [1], Liu and Garrett [2], and Chen
et al. [3,4] formulated the complex Nusselt number based on the
linear thermoacoustic theory [9,10]. Kornhauser and Smith [6,7]
⇑ Corresponding author. Address: Institute of Refrigeration and Cryogenics,
Zhejiang University, 38 Zheda Road, Hangzhou 310027, China. Tel./fax: +86 571
87953233.
E-mail addresses: ktang@zju.edu.cn (K. Tang), 9.yj9@163.com (J. Yu), jintao@z-
ju.edu.cn (T. Jin), wyp_mail@163.com (Y.P. Wang), tangwentao1014@126.com
(W.T. Tang), gan_zhihua@zju.edu.cn (Z.H. Gan).
dimensionless heat flux, similar to the complex Nusselt number,
was proposed in the experimental investigation by Bouvier et al.
[8]. The complex Nusselt number can reflect the variation of the
Nusselt number within one oscillation cycle, and can also indicate
the phase relation of the oscillating fluid-to-wall temperature
differences and the oscillating heat flux, which are important
characteristics of the oscillating flow.
The other strategy focused on the space-cycle averaged heat
transfer characteristics in a channel or a whole heat exchanger,
leading to a space-cycle averaged Nusselt number. Piccolo and
Pistone [11], and Meng et al. [12] studied the space-cycle averaged
heat transfer characteristics by means of numerical simulations.
Cooper et al. [13] performed an experimental investigation of the
convective heat transfer from the heated floor of a rectangular duct
subjected to a low frequency, large tidal displacement oscillating
flow (0.1–0.6 Hz, peak-to-peak displacement amplitudes of 3.5–
22.86 m). Zhao and Cheng [14] studied the heat transfer of the
oscillating air in a long copper circular tube, heated by the uniform
heat flux at outer surface. They correlated the Nusselt number to a
dimensionless oscillation amplitude of fluid and a dimensionless
frequency. Brewster et al. [15] measured the heat transfer coeffi-
cient of a heat exchanger with copper parallel plates mounted onto
a copper block, in a thermoacoustic engine system with an external
driving piston. Leong and Jin [16] investigated the heat transfer in
the oscillating flow through a parallel-plate channel filled with alu-
minum foam. Nsofor et al. [17] observed the heat transfer at the
finned-tube type heat exchanger of a thermoacoustic refrigeration
system, and summarized their experimental data with a Nusselt

0017-9310/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.ijheatmasstransfer.2013.11.067
812 K. Tang et al. / International Journal of Heat and Mass Transfer 70 (2014) 811–818

Nomenclature

A1 area of the base surface Va Valensi number

A2 surface area of the fin
Aflow flow area of the heat exchanger Greek symbols
Ao dimensionless oscillation amplitude d thickness
c specific heat dm viscous penetration depth
C acoustic compliance dj thermal penetration depth
d Diameter c specific heat ratio
dh hydraulic diameter g efficiency
f frequency m kinematic viscosity
h convective heat transfer coefficient q density
H height x angular frequency
k thermal conductivity n correction factor for heat balance
l length
n number of the fins Subscripts
Nu Nusselt number
c cool fluid side
p pressure cor correction
pA pressure amplitude Cu copper
Pr Prandtl number
f working fluid
Q heat flow fin fin
qV volume flow rate of cooling water h warm fluid side
Remax maximum Reynolds number i the ith fin
Rerms root mean square Reynolds number
res reservoir
Rex kinetic Reynolds number w wall
t temperature/time water cooling water
umax velocity amplitude wi inner surface (base surface)
Umax volume velocity amplitude
wo outer surface
V volume

number correlation to Reynolds number. Wakeland and Keolian exchanger. As schematically shown in Fig. 1, the experimental
[18] introduced a concept of ‘‘effectiveness’’ to analyze the heat apparatus consists of a linear compressor, a tested heat exchanger,
exchanger’s performance in an oscillating flow. The space-cycle a needle valve and a reservoir.
averaged Nusselt number is consistent with the conventional The pressure-wave generator is a linear compressor of type
understanding of the Nusselt number, and can be directly used 2S132W, fabricated by QDrive for driving a Stirling-type pulse tube
for designing the heat exchangers working in an oscillating flow. refrigerator. The mean electricity input power of the compressor is
Additionally, the visualization techniques of temperature field 500 W, powered by an AC power supply of type PCR2000M fabri-
[19–22] have been developed and applied in the study of the heat cated by Kikusui Electronics Corp., with the operating parameters
transfer in an oscillating flow. The variation in the temperature of AC 1–270 V, 20 A, 2000 VA, and 40–500 Hz. The jacket of the
field during one cycle could be measured to summarize the compressor is cooled by water to prevent the overheating,
space-cycle heat transfer characteristics. especially for the case with large working current. The maximum
Following the second strategy mentioned above, this study
focused on the heat transfer of a finned heat exchanger usually ap-
plied as the ambient-temperature heat exchanger in a pulse tube
refrigerator. An experimental apparatus was built according to
the operating conditions of the pulse tube refrigerators, which will
be presented in next section in details. Based on the measured
data, the space-cycle averaged Nusselt numbers (shortened as
‘‘Nusselt number’’ for simplicity in the following context) were
calculated and plotted with respect to the maximum Reynolds
number and the Valensi number, which are the important and
widely used similarity criterions for the oscillating flow
[3,4,8,11,14,16–18]. Upon comparing our results with some typical
reported works, a new expression was proposed to correlate the
Nusselt number to the maximum Reynolds number and the
Valensi number.

2. Experimental apparatus and measurement

2.1. Experimental apparatus

The ambient-temperature heat exchangers in a pulse tube

refrigerator cool the working gas at its hot ends. Thus, our experi-
mental apparatus is designed for the test of a water-cooled heat Fig. 1. Schematic of the experimental apparatus.
 K. Tang et al. / International Journal of Heat and Mass Transfer 70 (2014) 811–818
Fig. 2. Schematic of the finned heat exchanger.
working current is restricted below 10 A in our experiments to en-
sure safe operation.
Fig. 2 presents the detailed structure of the tested copper heat
exchanger. The cooling water enters the heat exchanger from the
top, and is then divided into two streams, which flow through
the channels surrounding the heat exchanger, converge at the bot-
tom and flow out of the heat exchanger. The parallel fins, fabri-
cated by electrical discharge machining, construct the channels
for the oscillating working gas. The fins have a thickness of
0.75 mm and a uniform length of 20 mm, and their heights range
from 2.6 mm to 13.6 mm. The fin-to-fin gap is 0.75 mm. The tested
heat exchanger, as well as the connecting tubes for the working gas
and the cooling water, is insulated by polystyrene foam to reduce
the influence of the heat transfer with the surroundings.
A combination of needle valve and reservoir, usually for mea-
suring the acoustic power output from a thermoacoustic engine
[23,24], is adopted to adjust and measure the velocity amplitude
in the heat exchanger. The tests of the heat transfer under different
maximum Reynolds numbers are conducted by adjusting the
opening of the needle valve. The velocity amplitude in the heat ex-
changer can be calculated from the measured pressure amplitude
in the reservoir.
2.2. Measurement
Nine thermometers, three pressure sensors and one flowmeter
are employed in the experiments, whose arrangement can be
found in Figs. 1 and 2.
The temperatures of the cooling water flowing into and out of
the heat exchanger are measured by two sheathed PT100 platinum
resistance thermometers of T1 and T2, which are immerged into
the cooling water through two Swagelok tee-fittings, respectively.
The temperatures of the working gas at the so-called inlet and out-
let of the heat exchanger, as well as the temperature of working
gas in the reservoir, are measured by three PT100 platinum resis-
tance thermometers of T3, T4 and T5 with 2 mm in diameter and
7 mm in length. Two cone chambers are adjacent to the edges of
the heat exchanger to connect the heat exchanger and the connect-
ing tubes, and the working gas from all of the flow passages
constructed by the fins is mixed in the cone chambers. The ther-
mometers T3 and T4 are located inside these two cone chambers
to measure the temperature of the mixed gas, which is regarded
as the mean temperature of the working gas flowing into and out
813
of the heat exchanger. The constant current supply and the voltage
signals of the thermometers (T3, T4 and T5) are transmitted be-
tween the atmospheric surroundings and the high pressure work-
ing gas through the electric connectors. Additionally, the K-type
armored thermocouples (T6, T7, T8 and T9) of 0.5 mm in diameter
are inserted into the flanges of the heat exchanger to measure the
wall temperature. The distance from the tips of thermocouples to
the fin base is 4 mm. The symmetrical arrangement of these ther-
mocouples allows us to calculate the fin-base temperature with
the average of T6, T7, T8 and T9. The temperature differences
among all these nine thermometers are corrected to ±0.1 K by cal-
ibrating them simultaneously in a variable-temperature water
bath, since the present measurement is quite sensitive to the
accuracy of temperature differences instead of the absolute tem-
perature accuracy.
The pressure measurement employs two UNIK-5000-type pres-
sure sensors, P1 and P2, fabricated by DRUCK Ltd., with the range
of 0–5 MPa and the accuracy of 0.2% FS (10 kPa). Since the pressure
amplitude in the reservoir is the key parameter for calculating the
velocity amplitude in the tested heat exchanger, a dynamic pres-
sure sensor of model 106B manufactured by PCB PIEZOTRONICS
Inc., with an accuracy of 0.57 kPa, is used to measure the dynamic
pressure in the reservoir (i.e. P3 in Fig. 1). In addition, the flow rate
of the cooling water is measured by a mini turbine flowmeter with
an accuracy of 0.005 l/min.
All electric signals of abovementioned sensors are transmitted
into the computer via CompactDAQ devices of National Instru-
ments and 34970A Data Acquisition Unit of Agilent.
2.3. Calculation of maximum Reynolds number, Valensi number and
Nusselt number
The maximum Reynolds number Remax is expressed as
umax dh
Remax ¼ ; ð1Þ
m
where umax is the velocity amplitude of the oscillating working gas
in the channel of the tested heat exchanger, dh is the hydraulic
diameter of the channel, and m is the kinematic viscosity of the
working gas. For the parallel-plate channels constructed from the
fins, dh is twice the fin-to-fin gap. umax can be calculated by
U max pA3 xC res 2pfpA3 V res
umax ¼ ¼ ¼ ; ð2Þ
Aflow Aflow Aflow cpm2
where Umax is the volume velocity amplitude of the oscillating
working gas, Aflow is the flow area in the heat exchanger, pA3 is
the pressure amplitude in the reservoir measured by pressure sen-
sor P3, x is the angular frequency, Cres is the acoustic compliance of
the reservoir [23], f is the frequency, Vres is the volume of the reser-
voir, c is the specific heat ratio of the working fluid, pm2 is the mean
operating pressure in the reservoir measured by pressure sensor P2.
The Valensi number Va is formulated as
 2
xd2h 2dh
2
dh
Va ¼ ¼ pffiffiffiffiffiffiffiffiffiffiffiffi ¼2 ; ð3Þ
m ð 2m=xÞ
2 dm
which is the ratio of the time scale of viscous penetration to the
oscillation period, also twice the squared ratio of the hydraulic
diameter to the viscous penetration depth dm.
The Nusselt number is defined as
hdh
Nu ¼ ; ð4Þ
k
where h is the convective heat transfer coefficient of the oscillating
flow, k is the thermal conductivity of the working gas.
The convective heat transfer coefficient h can be expressed as
814 K. Tang et al. / International Journal of Heat and Mass Transfer 70 (2014) 811–818

Qh 2.4. Heat balance analysis

h¼ Pn ; ð5Þ
ðA1 þ i¼1 A2;i gfin;i Þðtf  twi Þ
where Qh is the heat flow transferred from the oscillating flow, A1 is
the area of the base surface, A2,i and gfin,i are the surface area and
the fin efficiency of the ith fin, n is the quantity of the fins, tf is
the temperature of the oscillating flow, twi is the temperature of
the base surface.
The heat flow Qh can be calculated from the cooling water side
by
Q h ¼ nQ c ¼ nqV qwater cwater ðt 2  t 1 Þ; ð6Þ
where qV, qwater and cwater are the volume flow rate, the density and
the specific heat of the cooling water, t1 and t2 are the inlet and the
outlet temperature of the cooling water, respectively. qV is mea-
sured by the flowmeter F, while t1 and t2 are measured by the ther-
mometers of T1 and T2, as shown in Fig. 1. qwater and cwater are
determined according to the atmosphere pressure and the average
temperature of cooling water flowing into and out of the heat ex-
changer, i.e., (t1 + t2)/2. n is a factor to treat the deviation of Qh
and Qc due to the heat loss to the surroundings, in despite of the
thermal insulation employed for the tested heat exchanger and
the adjacent connecting tubes. n can be determined by an additional
experiment for the heat balance analysis, which will be discussed in
next section.
The temperature of the oscillating flow tf is considered as the
average of the temperatures of oscillating flows at the ends of
the heat exchanger, measured by the thermometers T3 and T4.
t3 þ t4
tf ¼ : ð7Þ
2
Taking into account the thermal conduction inside the copper,
the temperature of the base surface twi is calculated by
Qh þ Qc 1 dwo
t wi ¼ two þ ln ; ð8Þ
2 2pkCu l dwi
where kCu is the thermal conductivity of copper, l is the fin length,
dwo and two are the diameter and the temperature of the so-called
outer surface,
As above mentioned, the heat flow Qh transferred by the
oscillating flow is not exactly equal to the heat flow Qc transferred
by the cooling water, due to the heat loss to the surroundings.
Additional experiments have been performed to analyze the heat
balance of our experimental apparatus. The oscillating working
gas is replaced by warm water to obtain the amount of heat trans-
fer in the finned passages by measuring the volume flow rate, the
inlet and outlet temperature of the warm water. The cooling
water side keeps the same as the preceding experimental appara-
tus. The same thermal insulation is employed for the tested heat
exchanger and the connecting tubes. According to the test condi-
tion of the oscillating flow, the inlet temperature of the warm
water is controlled in the range of 300–330 K by a thermostat,
the cooling water is supplied by another thermostat set at
293 K, and the ambient temperature is kept at 298 ± 1 K by an
air conditioner.
The relative deviation of Qh and Qc versus the average temper-
ature th of the inlet and outlet temperature of the warm water is
presented in Fig. 3. (Qh  Qc)/Qh rises with an increase in th, and
for th = 325 K, (Qh  Qc)/Qh is roughly 5%. The rising rate has a
decreasing trend. We fit (Qh  Qc)/Qh and th into the following
function with a coefficient of determination of R2 = 0.977,
Qh  Qc
¼ 5:43  105 t 2h þ 3:556  102 t h  5:77: ð14Þ
Qh
Consequently, the factor n can be calculated by
Qh 1
n¼ ¼
Q c 1  ðQ h  Q c Þ=Q h
1
¼ : ð15Þ
1 þ 5:43  105 t 2h  3:556  102 th þ 5:77
For the oscillating flow, the temperature tf (see Eq. (7)) of the oscil-
lating flow is considered as th. All the experimental results pre-
sented in next section are obtained with Qh according to Eqs. (6)
and (15).
3. Results and discussion

dwo ¼ dwi þ 2dw ; ð9Þ 3.1. Results

t6 þ t7 þ t8 þ t9 Since this work focuses on the finned heat exchanger applied in

t wo ¼ : ð10Þ Stirling-type pulse tube refrigerators, the test conditions are set
4
according to the operation of a typical Stirling-type pulse tube
dwi is the diameter of the base surface, dw is the distance from the
tips of the thermocouples (T6, T7, T8 and T9) to the base of the fins.
t6, t7, t8 and t9 are measured by the thermocouples of T6, T7, T8 and 0.06
T9, as shown in Fig. 2. As above mentioned, the symmetrical
arrangement of the four thermocouples allowed us to consider
the average of t6, t7, t8 and t9 as two. 0.05
For the rectangular fins of the tested heat exchanger, the fin effi-
ciency can be formulated as follows [25]
(Qh-Qc)/Qh

0.04
tanhðmHcor Þ
gfin ¼ ; ð11Þ
mHcor
0.03
sffiffiffiffiffiffiffiffiffiffiffiffiffiffi -5 2 -2
y=-5.43x10 x +3.556x10 x-5.77
2h
m¼ ; ð12Þ 0.02 2
R =0.977
kCu dfin

dfin 0.01
Hcor ¼ H þ ; ð13Þ 300 305 310 315 320 325 330
2
th (K)
where dfin and H are the thickness and the height of the fin,
respectively. Fig. 3. (Qh  Qc)/Qh as a function of th.
 K. Tang et al. / International Journal of Heat and Mass Transfer 70 (2014) 811–818 815

9 9
8 8
7 7 3
2
6 3 6 2
5 5
1 1
Nu

Nu
4 4
3 3
2 5 2 5 2.5 MPa, Va=250
1 2.5 MPa, Va=150 1 3.0 MPa, Va=250
4 4
3.0 MPa, Va=150 3.5 MPa, Va=250
0 0
0 200 400 600 800 1000 1200 1400 0 200 400 600 800 1000 1200 1400
Remax Remax

Fig. 4. Nusselt number Nu as a function of the maximum Reynolds number Remax Fig. 6. Nusselt number Nu as a function of the maximum Reynolds number Remax
with the Valensi number Va = 150. with the Valensi number Va = 250.

9 9
8 8
3
7 7
3
6 6 2
2
5 5
1
Nu
Nu

4 1 4
3 3
2 5 2 5
1 2.5 MPa, Va=200 1 3.0 MPa, Va=300
4 3.0 MPa, Va=200 4 3.5 MPa, Va=300
0 0
0 200 400 600 800 1000 1200 1400 0 200 400 600 800 1000 1200 1400
Remax Remax

Fig. 5. Nusselt number Nu as a function of the maximum Reynolds number Remax Fig. 7. Nusselt number Nu as a function of the maximum Reynolds number Remax
with the Valensi number Va = 200. with the Valensi number Va = 300.

number with the increasing maximum Reynolds number, and the

refrigerator. The working gas is helium with three mean operating Nusselt number also increases with a rise in the Valensi number.
pressures, i.e., 2.5 MPa, 3.0 MPa and 3.5 MPa. The operating fre- Since the tested Nusselt number is a function of the tempera-
quency varies in the range of 40–110 Hz, leading to the Valensi tures (t1, t2, t3, t4, t6, t7, t8 and t9) and the cooling water flow rate
number Va from 150 to 350 in the tested heat exchanger. The (qV), the uncertainties of the tested Nusselt number denoted by
velocity amplitude of the helium gas in the channel between the the error bars in Figs. 4–8 are estimated by Eq. (16), considering
fins ranges from 0.5–3.8 m/s with the corresponding maximum the propagation of errors. The errors of each measured tempera-
Reynolds number Remax varying from 200 to 1200. All the oscillat- ture and flow rate include the systematic and random errors, as
ing flow situations in our experiments are laminar flowpffiffiffi according presented in Eq. (17). The systematic errors depend on the sensors
to
pffiffiffi thepffiffiffiffiffiffiffiffiffiffiffiffi
critical dimensionless
pffiffiffiffiffiffi value (bcri ¼ 2umax dm =m ¼ and the corresponding data acquisition devices. The random errors
2umax 2m=x=m ¼ 2Remax = Va) of 400–780, as reported in are represented by the standard deviation, as formulated by Eq.
literatures [26–28]. The pressure ratio (the maximum divided by (18). The maximum relative error of the tested Nusselt numbers
the minimum of the oscillating pressure) of P1 (see Fig. 1) is kept is 6.6%.
at 1.2, which is a typical value for a Stirling-type pulse tube vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
u 9 
refrigerator. uX @Nu2
Figs. 4–8 show the experimental results of Nusselt number for rNu ¼t r2xi ; ð16Þ
i¼1
@xi
Va = 150, 200, 250, 300 and 350, respectively. Roughly identical
Nusselt numbers are obtained at the same maximum Reynolds qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
number and Valensi number for the tests with different mean rxi ¼ r2sxi þ r2rxi ; ð17Þ
operating pressures, which implies the rationality in choosing the
maximum Reynolds number and Valensi number as the indepen- sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Pn 2
dent variables to summarize the heat transfer performance of the j¼1 ðxi;j  xi Þ
rrxi ¼ ð18Þ
oscillating flow. There is a slight increasing trend of the Nusselt n1
816 K. Tang et al. / International Journal of Heat and Mass Transfer 70 (2014) 811–818

9 diameter of the tested tube.) and the kinetic Reynolds number

2
Rex (Rex ¼ xdh =m, i.e., Valensi number Va in this work.), as Eq.
8 3
(23). The dimension of their test tube is 60.5 cm in length with
7 the inside and outside diameters of 1.35 cm and 1.57 cm, respec-
2 tively. The Rex ranges from 23 to 464 with the Ao varying from
6
8.5 to 34.9 in their experiments.
5
1 Nu ¼ 0:02A0:85 Re0:58
x : ð23Þ
Nu

o
4
3 With Ao = 2Remax/Va and Rex = Va, Eq. (23) can be rewritten as

2 5 Nu ¼ 0:036Re0:85
max Va
0:27
: ð24Þ
1 Nsofor et al. [17] investigated the heat transfer of a finned heat
4 3.5 MPa, Va=350
exchanger in a thermoacoustic refrigerator under the conditions of
0
0 200 400 600 800 1000 1200 1400 the mean operating pressure from 0.3 MPa to 0.8 MPa and the fre-
quency from 300 Hz to 450 Hz with helium as its working fluid.
Remax
The parallel-plate channels were constructed from the copper fins
Fig. 8. Nusselt number Nu as a function of the maximum Reynolds number Remax
of 0.15 mm in thickness with the fin-to-fin gaps of 0.95 mm. Based
with the Valensi number Va = 350. on the 24 tested Nusselt numbers, a correlation was obtained as,

Nu ¼ 0:61Re0:31
rms Pr
0:11
: ð25Þ
3.2. Comparison between our experimental results and the reported pffiffiffi
For the root mean square Reynolds number Rerms ¼ Remax = 2,
correlations
Eq. (25) can be expressed as

The calculation methods of the Nusselt number for the oscillat-
ing flow, reported in the existing literatures, were developed
mainly via three approaches, i.e., citing the available steady-flow
heat transfer correlations [10,11,29], developing simple physical
models [30] and conducting experimental tests of the oscillating
flow [14,16,17].
Typical approaches based on the steady-flow heat transfer cor-
relations are so-called ‘‘Time-Average Steady-Flow Equivalent’’
(TASFE) approximation [29] and ‘‘Root Mean Square Reynolds
Number’’ (RMSRe) model [10]. The former takes the time average
of the known steady-flow Nusselt number correlation over one cy-
cle to obtain the correlation for the oscillating flow. The latter sub-
stitutes the Reynolds number figuring in the known steady-flow
correlation with the root mean square Reynolds number calculated
with the oscillating fluid velocity. Piccolo and Pistone [11] pre-
sented two Nusselt number correlations, as given in Eqs. (19)
and (20), based on the Hausen correlation [25] for laminar flow
in the tube entrance region for TASFE and RMSRe models,
respectively,
Z p=x (
x 0:0668ðdh =LÞPrRemax sin xt
Nu ¼ 3:66 þ
p 0 1 þ 0:04½ðdh =LÞPrRemax sin xt
2=3
Nu ¼ 0:548Re0:31 0:11
max Pr : ð26Þ
The calculation results of Eqs. (19), (20), (22), (24), and (26) are
also presented in Figs. 4–8, with the corresponding labels of 1–5,
respectively.
The correlations of TASFE and RMSRe models based on the stea-
dy flow can reflect the rise in the Nusselt number with the increas-
ing maximum Reynolds number. However, omitting the influence
of Valensi number leads to an increasing deviation with the rising
Valensi number. The higher tested Nusselt numbers than the re-
sults of TASFE and RMSRe models may imply the heat transfer
enhancement by the oscillating flow, compared with the steady
flow.
In contrast with TASFE and RMSRe models, the correlation
based on Swift’s simple boundary layer model can reflect the im-
pact of Valensi number due to the function of thermal penetration
depth and frequency. A rise in Valensi number derived from the in-
crease in the frequency means a reduction in the boundary layer
thickness, which enhances the heat transfer indicated by the
increasing Nusselt number. However, the influence of maximum
)
Reynolds number is ignored so that the deviations of Nusselt num-
dt; ð19Þ ber with the variation of maximum Reynolds number are visible.

The results of the correlation by Zhao and Cheng, Eq. (24), can
pffiffiffi also indicate the rise in Nusselt number with the increasing max-
0:0668ðdh =LÞPrðRemax = 2Þ imum Reynolds number. The negative exponent of the Valensi
Nu ¼ 3:66 þ pffiffiffi 2=3 : ð20Þ
1 þ 0:04½ðdh =LÞPrðRemax = 2Þ Number leads to a slight decrease in Nusselt number with an in-
crease in Valensi Number. This variation is opposite to our test re-
Swift [30] suggested an approximate expression of the sults. The correlation by Nsofor et al. shows better agreement to
convective heat transfer coefficient as our experimental results than Eq. (24), however, the impact of
1 k the Valensi number is not yet considered.
h  pffiffiffi ; ð21Þ Among the abovementioned correlations, Swift’s simple bound-
2 dj
ary layer model, i.e., Eq. (22), can achieve a relatively better agree-
based on a boundary layer conduction heat transfer model. k and dj ment to our experimental results with a maximum error of 21.5%,
are the thermal conductivity and the thermal penetration depth of however, it is not yet satisfactory. For application convenience in
the working fluid, respectively. From Eqs. (4) and (21), the Nusselt the heat exchanger design, we adopt a typical expression as Eq.
number can be expressed as (27) to fit our experimental Nusselt number,

hdh 1 dh 1 dh 1 pffiffiffiffiffiffiffiffiffiffiffi m
Nu ¼ bRemax Van : ð27Þ
Nu ¼ ¼ pffiffiffi ¼ pffiffiffi pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ¼ VaPr: ð22Þ
k 2 jd 2 2m=ðxPrÞ 2
We obtain m = 0.0876 from the linear fitting of log(Nu) and
Zhao and Cheng [14] correlated the Nusselt number to the log(Remax) with a coefficient of determination of R2 = 0.946, based
dimensionless oscillation amplitude of fluid Ao (Ao = xmax/d, where on the data for Va = 250, as shown in Fig. 9. Then, all 84 data points
xmax is the peak-to-peak displacement amplitude, d is the inside are linearly fitted for logðNu=Re0:0876
max Þ and log(Va) with a coefficient
 K. Tang et al. / International Journal of Heat and Mass Transfer 70 (2014) 811–818 817

0.90 4. Conclusions
2.5 MPa, Va=250
3.0 MPa, Va=250 An experimental apparatus was built to investigate the heat
0.85 3.5 MPa, Va=250 transfer of a finned heat exchanger operating in the laminar oscil-
lating flow, aiming for the application of the ambient-temperature
heat exchangers of pulse tube refrigerators. The experiments
focused on the Nusselt number Nu with respect to the maximum
log(Nu)

0.80 Reynolds number Remax and the Valensi number Va. The test
y=0.0876x+0.6 results indicate that for Remax = 200  1200 and Va = 150  350,
2
R =0.946 Nu ranges in 5.0–8.3, and that the increases in Remax and Va can
0.75 both lead to a rise in Nu. The comparison between our experimen-
tal results and the typical correlations reported in the literatures
shows that the boundary layer conduction heat transfer model,
0.70 Eq. (22), achieves a relatively better agreement with our experi-
2.2 2.4 2.6 2.8 3.0 3.2 mental results (with a maximum deviation of 21.5%). A new
log(Remax) correlation of Eq. (27), including the effects of both Remax and
Va, was proposed and achieved a maximum deviation of 6.3%
Fig. 9. log(Nu) as a function of log(Remax) with Va = 250. compared to the experimental results, which means the feasibility
in aiding the design of the heat exchangers employed in pulse
tube refrigerators.
0.70 More working fluids (such as nitrogen, argon and carbon diox-
2.5 MPa ide), as well as different operating pressure ratio, are expected to
3.0 MPa be tested in the future work to verify the applicability of the above
0.65 3.5 MPa correlation, including the effect of Prandtl number on the heat
transfer performance. The investigation of the impact of heat
)
0.0876

0.60 exchanger geometry on the heat transfer performance in an oscil-

log(Nu / Remax

lating flow is also under consideration.

0.55
Acknowledgement

0.50 y=0.405x-0.367 The project is financially supported by the National Natural Sci-
2
R =0.964 ence Foundation of China (51176167 and 50890182).

0.45
2.1 2.2 2.3 2.4 2.5 2.6 References
log(Va)
[1] P. Nika, Y. Bailly, F. Guermeur, Thermoacoustics and related oscillatory heat
Fig. 10. logðNu=Re0:0876 and fluid flows in micro heat exchangers, Int. J. Heat Mass Transfer 48 (18)
max Þ as a function of log(Va).
(2005) 3773–3792.
[2] J. Liu, S.L. Garrett, Relationship between Nusselt number and the
thermoviscous (Rott) functions, J. Acoust. Soc. Am. 119 (3) (2006) 1457–
8.5 1462.
[3] Y. Chen, E. Luo, W. Dai, Heat transfer characteristics of oscillating flow
2.5 MPa
8.0 regenerator filled with circular tubes or parallel plates, Cryogenics 47 (1)
3.0 MPa
+10% (2007) 40–48.
7.5 3.5 MPa [4] Y. Chen, E. Luo, W. Dai, Heat transfer characteristics of oscillating flow
regenerators in cryogenic temperature range below 20 K, Cryogenics 49 (7)
(2009) 313–319.
7.0 [5] Y. Wang, Y.L. He, Q. Li, G.H. Tang, Numerical simulations of gas resonant
-10%
Nucal

oscillations in a closed tube using lattice Boltzmann method, Int. J. Heat Mass
6.5 Transfer 51 (11–12) (2008) 3082–3090.
[6] A.A. Kornhauser, J.L. Smith, Heat transfer with oscillating pressure and
6.0 oscillating flow, in: W.D. Jackson (Ed.), Proceedings of the 24th Intersociety
Energy Conversion Engineering Conference, IEEE, New York, 1989, pp. 2347–
5.5 2353.
[7] A.A. Kornhauser, J.L. Smith, Application of a complex Nusselt number to heat
5.0 transfer during compression and expansion, J. Heat Transfer Trans. ASME 116
(3) (1994) 536–542.
[8] P. Bouvier, P. Stouffs, J.P. Bardon, Experimental study of heat transfer in
4.5
4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 oscillating flow, Int. J. Heat Mass Transfer 48 (12) (2005) 2473–2482.
[9] G.W. Swift, Thermoacoustic engines, J. Acoust. Soc. Am. 84 (4) (1988) 1145–
Nuexp 1180.
[10] G.W. Swift, Thermoacoustics: A Unifying Perspective for Some Engines and
Refrigerators, Acoustical Society of America, American Institute of Physics,
Fig. 11. Comparison of the proposed correlation prediction and the experimental
New York, 227, 2002, pp. 99–103.
data. [11] A. Piccolo, G. Pistone, Estimation of heat transfer coefficients in oscillating
flows: the thermoacoustic case, Int. J. Heat Mass Transfer 49 (9–10) (2006)
1631–1642.
of determination of R2 = 0.964, as shown in Fig. 10. As a result,
[12] F. Meng, M. Wang, Z. Li, Lattice Boltzmann simulations of conjugate heat
b = 0.43 and n = 0.405. transfer in high-frequency oscillating flows, Int. J. Heat Fluid Flow 29 (4)
The calculated Nusselt numbers with Eq. (27) are compared (2008) 1203–1210.
with the experimental data in Fig. 11, which shows that the devi- [13] W.L. Cooper, V.W. Nee, K.T. Yang, An experimental investigation of convective
heat transfer from the heated floor of a rectangular duct to a low frequency,
ations are less than 10%, and the absolute value of the maximum large tidal displacement oscillating flow, Int. J. Heat Mass Transfer 37 (4)
deviation is 6.3%. (1994) 581–592.
818 K. Tang et al. / International Journal of Heat and Mass Transfer 70 (2014) 811–818
[14] T.S. Zhao, P. Cheng, Oscillatory heat transfer in a pipe subjected to a
laminar reciprocating flow, J. Heat Transfer Trans. ASME 118 (3) (1996)
592–597.
[15] J.R. Brewster, R. Raspet, H.E. Bass, Temperature discontinuities between
elements of thermoacoustic devices, J. Acoust. Soc. Am. 102 (6) (1997)
3355–3360.
[16] K.C. Leong, L.W. Jin, An experimental study of heat transfer in oscillating flow
through a channel filled with an aluminum foam, Int. J. Heat Mass Transfer 48
(2) (2005) 243–253.
[17] E.C. Nsofor, S. Celik, X.D. Wang, Experimental study on the heat transfer at the
heat exchanger of the thermoacoustic refrigerating system, Appl. Therm. Eng.
27 (14–15) (2007) 2435–2442.
[18] R.S. Wakeland, R.M. Keolian, Effectiveness of parallel-plate heat exchangers in
thermoacoustic devices, J. Acoust. Soc. Am. 115 (6) (2004) 2873–2886.
[19] C. Herman, E. Kang, Experimental visualization of temperature fields and study
of heat transfer enhancement in oscillatory flow in a grooved channel, Heat
Mass Transfer 37 (1) (2001) 87–99.
[20] C. Herman, M.P. Cetinguel, Experimental visualization of the thermoacoustic
effect, J. Heat Transfer Trans. ASME 131 (8) (2009) 080905.
[21] L. Shi, Z. Yu, A.J. Jaworski, Application of laser-based instrumentation for
measurement of time-resolved temperature and velocity fields in the
thermoacoustic system, Int. J. Therm. Sci. 49 (9) (2010) 1688–1701.
[22] A.J. Jaworski, A. Piccolo, Heat transfer processes in parallel-plate heat
exchangers of thermoacoustic devices – numerical and experimental
approaches, Appl. Therm. Eng. 42 (2012) 145–153.
[23] K. Tang, Z.J. Huang, T. Jin, G.B. Chen, Influence of acoustic pressure amplifier
dimensions on the performance of a standing-wave thermoacoustic system,
Appl. Therm. Eng. 29 (5–6) (2009) 950–956.
[24] M.E.H. Tijani, S. Spoelstra, A high performance thermoacoustic engine, J. Appl.
Phys. 110 (9) (2011) 093519.
[25] F.P. Incropera, D.P. DeWitt, T.L. Bergman, A.S. Lavine, Fundamentals of Heat
and Mass Transfer, sixth ed., Wiley, New York, 2007. pp. 147–155, 512–
513.
[26] P. Merkli, H. Thomann, Transition to turbulence in oscillating pipe flow, J. Fluid
Mech. 68 (3) (1975) 567–575.
[27] M. Hino, M. Sawamoto, S. Takasu, Experiments on transition to turbulence in
an oscillatory pipe-flow, J. Fluid Mech. 75 (2) (1976) 193–207.
[28] T.S. Zhao, P. Cheng, Experimental studies on the onset of turbulence and
frictional losses in an oscillatory turbulent pipe flow, Int. J. Heat Fluid Flow 17
(4) (1996) 356–362.
[29] P.D. Richardson, Effects of sound and vibration on heat transfer, Appl. Mech.
Rev. 20 (3) (1967) 201–217.
[30] G.W. Swift, Analysis and performance of a large thermoacoustic engine, J.
Acoust. Soc. Am. 92 (3) (1992) 1551–1563.