International Journal of Heat and Mass Transfer 103 (2016) 457–467

Contents lists available at ScienceDirect

International Journal of Heat and Mass Transfer

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

Experimental investigation on the pressure drop and heat transfer

characteristics of a recuperator with offset strip fins for a micro gas
turbine
Kyu Hyung Do a,b, Byung-Il Choi a,b, Yong-Shik Han a, Taehoon Kim a,b,⇑
a
Korea Institute of Machinery and Materials, Daejeon 34103, Republic of Korea
b
Korea University of Science and Technology, Daejeon 34113, Republic of Korea

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

Article history: In this study, the pressure drop and heat transfer characteristics of a recuperator with offset strip fins for
Received 15 April 2016 a micro gas turbine were experimentally investigated in the high-temperature range. A unit recuperator
Received in revised form 28 June 2016 with offset strip fins was fabricated using furnace-brazing technology in vacuum atmosphere.
Accepted 20 July 2016
Experiments were conducted by varying the mass flow rate and inlet temperature of the hot air stream
in the range of 1.5
4 g/s and 250
500 °C, respectively. Based on the experimental data, the total pressure
drop results measured at high temperature were much larger than those in ambient-temperature condi-
Keywords:
tion, and we showed that the inlet pressure of each air stream should be measured to correctly figure out
Recuperator with offset strip fins
Micro gas turbine
the pressure drop characteristics of the recuperator in the high-temperature range. Furthermore, the
Fabrication of a unit recuperator effectiveness was almost constant regardless of the inlet temperature of the hot air stream, which means
Modified analytical model that the fluid mean temperature variation hardly affected the effectiveness of the recuperator. Two types
Experimental investigation of analytical models were proposed to predict the pressure drop and the effectiveness of the fabricated
recuperator and the model prediction results were also compared with the experimental data. The com-
parison with the experimental data showed that the results from the simple model may lead to incorrect
results for the thermal efficiency of the micro gas turbine because the recuperator effectiveness was over-
estimated and the pressure drop of the recuperator was underestimated. On the other hand, the modified
model proposed in the present work successfully estimated the pressure drop and heat transfer charac-
teristics of the fabricated recuperator with offset strip fins for a micro gas turbine.
Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction
Numerous applications for mobile power sources, such as
robots, small unmanned aerial vehicles, and military applications,
are available, which require energy density not attainable by bat-
teries. They require higher power and energy densities in the range
of 100–600 W/kg and 200–6000 Wh/kg, respectively [1]. Second-
ary batteries can have relatively large power densities; however,
their energy densities hardly reach 250 Wh/kg, which limits their
autonomy. Charging time can also be a problem as well as very
cold external temperature. For these reasons, much effort has been
over the past decade to develop a mobile power generator that
could meet the growing demand for portable electricity [2].
⇑ Corresponding author at: Korea Institute of Machinery and Materials, Daejeon
34103, Republic of Korea. Fax: +82 42 868 7355.
E-mail address: thkim79@kimm.re.kr (T. Kim).
Recently, a micro gas turbine has been focused as one of the
best choices among various energy sources, including both ordi-
nary and emergency power supplies, because it could reach much
a higher energy density than existing batteries, although only a few
percentage of the energy density of common hydrocarbon fuels,
which is approximately 12 kWh/kg [3–6]. Because the specific fuel
consumption of a micro gas turbine is inversely proportional to the
thermal efficiency, we need to improve the thermal efficiency to
achieve compactness and lightness. The easiest way to enhance
thermal efficiency is to increase either the pressure ratio or the tur-
bine inlet temperature (TIT) from the thermodynamic cycle point
of view. However, for the micro gas turbine, both achievable pres-
sure ratio and TIT are significantly lower compared with those of
conventional gas turbines due to the limitations in materials and
manufacturing processes. An alternative method is to shift into
regenerative cycle. The introduction of a recuperator in conven-
tional gas turbines is well known to most often increase the ther-
mal efficiency because it reduces the fuel consumption in the

http://dx.doi.org/10.1016/j.ijheatmasstransfer.2016.07.071
0017-9310/Ó 2016 Elsevier Ltd. All rights reserved.
458 K.H. Do et al. / International Journal of Heat and Mass Transfer 103 (2016) 457–467

Nomenclature

A total heat transfer surface area [m2] T temperature [°C]

AER area enhancement ratio DT temperature difference [°C]
Af extended surface area [m2] tend end plate thickness [m]
Ak total wall cross-sectional area for longitudinal conduc- U uncertainty or overall heat transfer coefficient
tion [m2] [W m
2 °C
1]
Ao minimum free-flow area [m2] W recuperator width [m]
Aw total wall area for heat conduction [m2] Wout outlet channel width [m]
C heat capacity rate
C⁄ heat capacity rate ratio Greek symbols
CD drag coefficient a⁄ aspect ratio of rectangular ducts
Co Kozeny coefficient De effectiveness deterioration factor
cp specific heat capacity [kJ kg
1 °C
1] d fin thickness [m]
Dh hydraulic diameter of subchannel [m] dw separation sheet thickness [m]
D0h hydraulic diameter of bypass channel [m] df filler material sheet thickness [m]
dh hydraulic diameter of offset strip fin array [m] e effectiveness
f friction factor for the longitudinal flow through the off- /trans bypassing channel fraction
set strip fins g fin efficiency
f Re Poiseuille number k longitudinal wall conduction parameter
0
f friction factor for the transverse flow through the offset l viscosity [Pa s]
strip fins with bypass channels q density [kg m
3]
G mass velocity [m2s] ð1=qÞm fluid mean specific volume [m3 kg
1]
H recuperator height [m] r ratio of free flow area to frontal area
h heat transfer coefficient [W m
2 °C
1]
h0 height of offset strip fins [m] Subscripts
j Colburn j factor for the longitudinal flow through the C cold air stream
offset strip fins
cf creeping flow
j0 Colburn j factor for the transverse flow through the off- ch bypassing channel
set strip fins with bypass channels H hot air stream
Kc contraction coefficient
i hot or cold air stream
Ke expansion coefficient in inlet
ks thermal conductivity of an offset strip fin [W m
1 °C
1] m mean
L recuperator length [m] max maximum
ls strip length of an offset strip fin [m]
min minimum
m _ mass flow rate [kg s
1] out outlet
N number of fluid passages RI Region I
NTU number of transfer units RII Region II
Nu Nusselt number
RIII Region III
Pr Prandtl number s solid
p pressure [Pa] tot total
Dp pressure drop [Pa] trans transverse flow
q_ heat transfer rate [W]
w wall
Re Reynolds number wet wetted
s fin spacing [m]

combustor to achieve a certain TIT [7]. In contrast, the presence of
a recuperator leads to additional pressure drops at both cold and
hot sides, resulting in reduced output power of the micro gas tur-
bine. Normally, when the recuperator effectiveness increases by
1%, the overall thermal efficiency of the micro gas turbine would
increase by 0.35%. On the other hand, a 1% decrease in the pressure
loss of the recuperator would improve the thermal efficiency by
0.33% [8]. Therefore, a thermal design of a recuperator with high
effectiveness and low pressure loss is needed to achieve higher
thermal efficiency.
The present work mainly focuses on a recuperator with offset
strip fins for a micro gas turbine among the various types of
recuperators such as the Swiss-roll, microchannel, and primary-
surface recuperators [7,9,10]. Many studies have extensively
investigated the pressure drop and heat transfer characteristics of
rectangular offset strip fin heat exchangers in the past decades. Kays
and London [11] were the first to demonstrate a correlation of offset
strip fins based on experimental results. Manson [12] conducted
experiments using fin geometries in addition to the experiments
of Kays and London. Using a power method, Wieting [13] obtained
a correlation using 22 fin geometries. Mochizuki and Yovanovich
[14] presented more accurate correlations by modifying those of
Wieting [13]. Webb and Joshi [15] suggested correlations in both
laminar and turbulent regimes, but these correlations could not
be applied to the transition regime. Manglik and Bergles [16] pre-
sented a correlation that could be applied to laminar, transition,
and turbulent regimes using experimental data from the literature.
Despite this broad investigative effort, few experimental studies
were conducted on a recuperator with offset strip fins, which can
be applied to micro gas turbines. Furthermore, previous experimen-
tal data for the friction factor f and Colburn j factor were obtained in
low or moderate temperature ranges, which may not be applicable
to the actual operating temperature range of a micro gas turbine.
Therefore, the main objective of the present work is to experi-
mentally investigate the pressure drop and heat transfer character-
istics of a recuperator for a micro gas turbine. For this purpose, a
 K.H. Do et al. / International Journal of Heat and Mass Transfer 103 (2016) 457–467
unit recuperator with offset strip fins is fabricated using furnace
brazing-technology in vacuum atmosphere. Experiments are con-
ducted by varying the mass flow rate and the inlet temperature
of the hot air stream to estimate the pressure drop and heat trans-
fer characteristics of the fabricated recuperator. In addition, an
analytical model is suggested to accurately predict the perfor-
mance of the fabricated recuperator.
2. Experimental investigation
2.1. Fabrication of a unit recuperator with offset strip fins
In this study, a single-pass counterflow recuperator with offset
strip fins for a micro gas turbine was manufactured. Fig. 1(a) shows
the schematic diagram of a micro gas turbine. This micro gas tur-
bine is designed for the energy capacity of 500-W within a 1-L vol-
ume. It consists of a generator, a combustor, a compressor, a
turbine, and a recuperator module. The generator, compressor
and turbine share the axis and are arranged in this order. The gen-
erator is located in front of the compressor. The inlet air passes
through the passage beside the generator and enters the compres-
sor through the recuperator module. The turbine inlet gas comes
from the annular-type combustor and passes through the guide
vane. The guide vane is located in front of the turbine and the
rotating speed of the turbine and compressor in the normal oper-
ating condition is 400,000 rpm. Then combustion gas is discharged
into the atmosphere through the recuperator module.
The recuperator module is composed of six unit recuperators,
which are installed outside the combustor in the circumferential
(a) Layout of a 500-W micro gas turbine
Hot Air Stream
ls
Cold Air Stream
L
(b) Offset strip fins
Fig. 1. Schematic diagram of a recuperator module and offset strip fins.
459
direction at equal intervals. To meet the volume requirement, the
height, width, and length of the unit recuperator should be set as
H = 35.9 mm, W = 12 mm, and L = 81 mm. The unit recuperator
has six basic parts, which consist of the offset strip fin sheets for
the hot and cold streams, separation sheets, side plates for both
the hot and cold streams, and end plates made from commercial
SUS304 stainless steel. The internal structure of the offset strip
fin sheets of the cold and hot streams was determined based on
the design optimization results [17]. Fig. 1(b) shows the schematic
diagram of the offset strip fins. The cold and hot streams enter into
and leave from the opposite sides of the core, providing a counter-
flow arrangement in the main part of the recuperator. To fabricate
the unit recuperator, each basic part was prepared and then
sequentially stacked up, as shown in Fig. 2. Furnace-brazing tech-
nology in vacuum atmosphere was used to connect the plates and
the offset strip fins. To obtain sufficient stable bonding strength,
50-lm-thick filler metal sheets, which are made of BNi-2 brazing
alloy, were inserted between the offset strip fin and separation
sheets. The unit recuperators were brazed inside a high vacuum
brazing furnace equipped with a diffusion pump at vacuum of
more than 5  10
4 Torr. On the basis of the experimental study
presented by Lee et al. [18], the brazing temperature and holding
time were chosen to be 1070 °C and 10 h, respectively. After the
vacuum-brazing process, leak detection test was carried out using
a simple test rig to ensure absolute leak tightness of the fabricated
unit recuperator. The detailed dimensions between the designed
and fabricated unit recuperator are listed in Table 1, which shows
that the height, length, and width of the fabricated unit recupera-
tor were slightly larger than those of the designed one. The main
reasons for the increase in the height of the fabricated unit
δw Plate
h′
δ
s
460 K.H. Do et al. / International Journal of Heat and Mass Transfer 103 (2016) 457–467

optical measuring microscope (BXFM-100, Olympus Co. Ltd.).

Because of the misalignment along the length of the recuperator,
bypass channels without offset strip fins appeared at the end of
End plate
the hot air stream outlet region. On the other hand, the cold air
stream outlet region was fully filled with offset strip fins, and no
Side plate for cold stream bypass channel existed. From the measurement result, the average
Offset strip fin sheet for cold stream
width of the bypass channel Wgap in the hot air stream outlet
region was approximately 4.1 mm.
Separation sheet
Side plate for hot stream
Offset strip fin sheet for hot stream 2.2. Experimental apparatus
Separation sheet
To evaluate the pressure drop and heat transfer characteristics
of the recuperator with offset strip fins for a micro gas turbine,
experiments were conducted by varying the mass flow rate and
the inlet temperature of the hot stream of the fabricated unit recu-
perator. The experimental apparatus is schematically shown in
Fig. 4, which mainly consists of a unit recuperator assembly, an
(a) Basic parts of a unit recuperator air-heating system, and a data acquisition system. To measure
the temperatures and pressures at the inlet and outlet of both
hot and cold air streams, inlet and outlet headers were welded
on the fabricated unit recuperator. Four 1/16-in sheathed K-type
thermocouples were installed inside the headers to measure the
inlet and outlet temperatures of the hot and cold air streams,
Wout W respectively. The pressure drops of the hot and cold air streams
were measured by differential pressure transmitters (Sensor Sys-
hH′ + δ hC′ + δ
tem Technology Co. Ltd., SSD) with a measurement accuracy of
±0.3% FS. The inlet pressure values of each air stream were also
measured using an absolute pressure transducer (Honeywell Inc.,
FP2000) with a measurement accuracy of ± 0.1% FS to calculate
the air density. The mass flow rates of the hot air stream and cold
air stream inlets were controlled and measured using mass flow
controllers (Brooks Instrument, SLA5853S) with a measurement
(b) Fabricated unit recuperator accuracy of ±1.0% FS. Each air stream was heated using a 6-kW
high-temperature threaded inline air heater (OSRAM Sylvania
Fig. 2. Schematic diagram of a unit recuperator with offset strip fins. Inc., F076525), which can generate air temperature of up to
900 °C. The temperature at the inlet was kept constant by the hea-
ter controller (Yokogawa Electric Co., UP55A) at a certain level of
Table 1 temperature.
Detailed dimensions of the unit recuperator. In the present study, the unit recuperator assembly, including
Parameters Designed value [17] Fabricated value
the inlet and outlet headers and the air-heating system with two
inline air heaters, was separately installed in the machined porous
L [mm] 81 83
H [mm] 36.9 46.4
ceramic cages to minimize heat loss by convection and radiation.
W [mm] 12 13.1 Furthermore, a ceramic fiber insulator (Cerakwool) was used to fill
h’H [mm] 0.8 0.8 the gap between the case and the unit recuperator assembly and
h’C [mm] 0.4 0.5 the inline air heaters.
ls [mm] 3.0 3.0
During the experiments, the mass flow rate and the inlet tem-
s [mm] 1.4 1.4
d [mm] 0.1 0.1 perature of the hot air stream of the recuperator were varied in
dw [mm] 0.1 0.2 the range of 1.5
4 g/s and 250
500 °C, respectively, whereas the
df [mm] – 0.0222 inlet temperature of the cold air stream of the recuperator was
tend [mm] 0.6 1.1 fixed as 171 °C. All the temperature, pressure, and mass flow rate
NH 22 22
NC 23 23
data were collected and recorded using a data acquisition system
(Agilent Technologies Inc., 34970A) with a sampling rate of one
data point per second. We assumed that each test case reached
the steady-state condition, which indicates that the temperature
variation was less than 0.1 °C for 2 min. Each test was conducted
recuperator were the limitations in manufacturing processes,
three times.
height of the offset strip fin sheet for the cold stream, increase in
the thickness of the separation sheet and the end plate, and inser-
tion of the filler metal sheets. In addition, the increment in the 3. Analytical approach
length and width of the manufactured recuperator resulted in
slight misalignment. Fig. 5(a) shows a simplistic view of the flow configuration of the
To observe the internal structure of the unit recuperator, the fabricated unit recuperator with offset strip fins. Fig. 5(a) shows
fabricated recuperator specimen was divided into six equal pieces that the hot and cold air streams enter into and leave from the
using a dicing saw machine, as shown in Fig. 3(a). Fig. 3(b) shows opposite sides of the core region, providing a counterflow arrange-
the measurement result of the hot air stream outlet region using an ment in most parts of the recuperator except for the outlet region
 K.H. Do et al. / International Journal of Heat and Mass Transfer 103 (2016) 457–467 461

(a) Fabricated specimen divided into six equal pieces

(b) Measurement result of the hot air stream outlet

Fig. 3. Photographs of the unit recuperator.

in each air stream. This means that the pressure drop and heat The loss coefficients for abrupt contraction, Kc and abrupt
transfer characteristics of the fabricated recuperator can be easily expansion, Ke can be computed from the classic expressions [19]
estimated under the assumption of the counterflow arrangement, as follows:
as shown in Fig. 5(b). In the present work, two types of analytical
K c  0:42ð1
r2 Þ ð2Þ
approaches were used to predict the pressure drop and the effec-
tiveness of the fabricated recuperator. One is a simple analytical 2
model that applies a simplified flow configuration, as shown in K e ¼ ð1
r2 Þ ð3Þ
Fig. 5(b), and the other is a modified analytical model that consid- In the present study, to calculate the friction factor of the longi-
ers the actual flow configuration of the fabricated recuperator, as tudinal flow through the offset strip fins, the most comprehensive
shown in Fig. 5(c). correlations presented by Manglik and Bergles [16] and Muzychka
and Yovanovich [14] were used as follows:
3.1. Simple analytical model

0:1856
0:3053

0:2659
s d d
f ¼ 9:6243Re
0:7422
dh 0
In the simple analytical model, the pressure drop and heat h ls s
transfer characteristics of the recuperator are assumed to be the "
0:920
3:767
0:236 #0:1
s d d
same as those of the counterflow arrangement throughout the  1 þ 7:669  10
8 Re4:429
dh 0 ð4Þ
entire region of the recuperator. Under this assumption, the pres- h ls s
sure drop for both hot and cold air streams can be expressed by
the following equation: 2  
1=2 3 31=3
f ReD ðdh =Dh Þ
"
 # 6
h
Red
þ 1:328 Redh dls 7
G2i 4L 1 qi;in 6 h h
7 ð5Þ
Dpi ¼ ð1
r2i þ K c;i Þ þ f i q
ð1
r2i
K e;i Þ f ¼6   7
2qi;in dh;i i;in qi qi;out 4  
1=5 0
3 5
m
þ 0:074 Redh dls þ h2ldþsd=2
0
s ðh þsÞ
C D
h
ð1Þ
462 K.H. Do et al. / International Journal of Heat and Mass Transfer 103 (2016) 457–467

Fig. 4. Experimental apparatus.

where For a rectangular offset strip fin array, the values of f ReDh and
0 0 NuDh are given by the following results for a fully developed flow
2sh 4sh ls
C D ¼ 0:88; Dh ¼ 0 ; dh ¼ 0 0 ; in rectangular ducts, as presented by Shah and London [20]:
sþh 2ðsls þ h ls þ h dÞ þ sd
Gdh m_ f ReDh ¼ 23:94
30:05a þ 32:37a2
12:08a3 ð9Þ
Redh ¼ ; G¼ :
l Ao
NuDh ¼ 7:45
16:90a þ 22:10a2
9:75a3 ð10Þ
To predict the heat transfer characteristics, the Number of
 0
Transfer Units (NTU) was obtained from the following equation: where 0 < a ¼ h =s < 1 for a rectangular sub-channel.

1 In the case of a counterflow recuperator, the effectiveness is
UA UA 1 1 dw 1 dependent on the NTU and heat capacity rate ratio C⁄, as shown
NTU ¼ ¼ ¼ þ þ ð6Þ
_ p Þmin C min C min ðgo hAÞH ks Aw ðgo hAÞC
ðmc in Eq. (11).
where
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 1
exp½
NTUð1
C  Þ


0  e¼ ð11Þ
tanhðmlÞ Af 2h d h þd 1
C  exp½
NTUð1
C  Þ
go ¼ 1
1
; ml ¼ 1þ
d ;
ml A ks d ls 2
jGcp However, the longitudinal wall conduction effect cannot be
h¼ 2=3
: neglected for a high-effectiveness recuperator with a short flow
Pr length because it significantly reduces the effectiveness of the
The heat transfer coefficients of the hot and cold air streams can recuperator with very steep temperature changes in the flow
be estimated using the Colburn j factor and j correlations for the direction. Therefore, the influence of the longitudinal wall heat
longitudinal flow through the offset strip fins, which were sug- conduction is considered in the present model. The effectiveness
gested by Manglik and Bergles [16] and Muzychka and Yovanovich at C⁄ < 1 has been obtained and correlated by Kroeger [21] as
[14], as follows: follows:


0:1541
0:1499

0:0678
s d d 1
C
j ¼ 0:6522Re
0:5403 e¼1
ð12Þ
dh
h
0
ls s W expðr 1 Þ
C 
"
0:504
0:456

1:055 #0:1
s d d where
 1 þ 5:269  10
5 Red1:340 0 ð7Þ
h
h ls s ð1
C  ÞNTU 1 þ c W
r1 ¼ ; W¼ ;
1þa 1
c W
28 !5 97=10 32=7

<
Nu ðd =D Þ5 0:641ðf ReD Þ
1=3 
2 1=3
= a 1=2 1þc
6 Dh h h
þ h d h 7 W ¼ ;
6 : Red Pr1=3 2=3 D l ; 7 1þa 1=a
c
c2
6 Red h s
7 ð8Þ
j¼6 7 :
h h
6   

1=5 7=2 7
4 5 1
C 1 ks Ak C min
þ 0:037 Redh dls c¼ ; a ¼ C  kNTU; k ¼ ; C ¼ :
h 1 þ C 1 þ a LC C C max
 K.H. Do et al. / International Journal of Heat and Mass Transfer 103 (2016) 457–467 463

Hot air q_ RII ¼ eRII C min ðT H;in;RII
T C;in;RII Þ ð17Þ

Hot air
stream outlet
stream inlet The pressure drop in Region II is expressed by Eq. (16), and the
Cold air heat transfer rate based on Eq. (12) is expressed by Eq. (17).
stream inlet Core region Regions I and III near the flow outlet of each air stream can be
modeled as a heat exchanger with a crossflow arrangement. Fur-
thermore, in actual situation, no hot air stream inlet exists in
Cold air Region I; however, a pseudo inlet for the hot air stream with the
Outlet stream outlet same mass velocity of the outlet is assumed for the sake of conve-
region
nience in analysis. Under this assumption, the pressure drop of the
(a) Flow configuration of the fabricated recuperator hot and cold air streams in Region I can be expressed by the follow-
ing equations, respectively:
Hot air Hot air "
 #
G02 0 4W 1 qH;in
stream outlet stream inlet DpH;RI ¼ H
fH qH;in
ð1
r2H
K e;H Þ ; ð18Þ
2qH;in dh;H qH m qH;out
Core region


G2C 4W out 1
DpC;RI ¼ ð1
r2C þ K c;C Þ þ f C qC;in : ð19Þ
Cold air Cold air 2qC;in dh;C qC m
stream inlet stream outlet
Outlet 0
In Eq. (18), f H denotes the friction factor for the transverse flow
region
through the offset strip fins with bypass channels. The friction fac-
(b) Simplified flow configuration tor correlation for the transverse flow through the offset strip fins
with bypass channels suggested by Muzychka and Yovanovich [22]
Region III was used in this study and is expressed as
Hot air (Crossflow) 0
stream outlet Hot air f ¼ /trans f ch þ ð1
/trans Þf trans ð20Þ
stream inlet
where
Cold air
stream inlet f ReD0h 6=7 6=7 7=6 W gap
Core region f ch ¼ ; f trans ¼ ½ðf cf þ f lam Þ þ f tur  ; /trans ¼ ;
ReD0h W out

Cold air

1=5


Region I Region II 8C o 2ðs þ dÞ 1
stream outlet f cf ¼ ; f tur ¼ 0:074 Redh ;
(Crossflow) (Counterflow) Redh dh AER

(c) Flow configuration of the modified model (

1=2 )


2ðs þ dÞ 1
f lam ¼ 1:328 Redh
Fig. 5. Simplistic view of the flow configuration of the recuperator with offset strip dh AER
fins. (

1=2
)

0
1 2h 1
þ 1:328 Redh 1
;
2 dh AER
Once the effectiveness obtained from Eq. (11) or Eq. (12) is
determined, the outlet temperatures for both the hot and cold air 0sffiffiffiffiffiffiffiffiffiffiffiffiffiffi1
0 2
streams and the total heat transfer rate can be respectively esti- 2W out ðh þ dÞ 10 2@ b A pls
D0h
2
mated as
¼ 0 ; Co ¼ ð1 þ b ÞE ; b¼ :
ðW out þ h þ dÞ p 1þb
2 sþd
C min
T H;out ¼ T H;in
e ðT H;in
T C;in Þ; ð13Þ
Ch E() is the complete elliptic integral of the second kind, and AER is
the area enhancement ratio defined as AER ¼ Awet =Abare . Equation
C min (20) implies that the asymptotic value of the friction factor is the
T C;out ¼ T C;in þ e ðT H;in
T C;in Þ; ð14Þ
Cc friction factor for the transverse flow through the offset strip fins
0
when the width of the bypass channel goes to zero and the f value
q_ tot ¼ eC min ðT H;in
T C;in Þ ð15Þ is the same as that of the fully developed flow in the rectangular
0
duct with width Wout and height h þ d when the value of /trans
3.2. Modified analytical model approaches one.
The effectiveness of a single-pass crossflow heat exchanger with
In the present work, a modified analytical model that considers both fluids unmixed is expressed in the following form [11]:
the actual flow configuration is also developed to more accurately 

1 0:22  0:78
predict the pressure drop and the effectiveness of the fabricated e ¼ 1
exp NTU fexpð
C NTU Þ
1g : ð21Þ
C
recuperator. In the modified analytical model, the recuperator is
divided into three regions, as shown in Fig. 5(c). Region II is mod- Consideration of the longitudinal wall conduction effect of a
eled as a single-pass counterflow recuperator, and the pressure crossflow heat exchanger is more complicated than that of a coun-
drop and the effectiveness in this region are similar to those terflow heat exchanger because the temperature gradients in the
obtained from the simple analytical model. wall between the two fluid flows are generally different. Only the

 numerical results of the ratio of the effectiveness difference to
4f i ðL
2W out ÞG2i 1 the effectiveness without longitudinal wall conduction, namely,
Dpi;RII ¼ ð16Þ
2dh;i qi m De=e were presented in a tabulated form for C⁄ = 1 [23]. These
464 K.H. Do et al. / International Journal of Heat and Mass Transfer 103 (2016) 457–467

results, as a function of NTU; kH ; kC ; and ðgo hAÞC =ðgo hAÞH , were DpC ¼ DpC;RI þ DpC;RII þ DpC;RIII =2; ð32Þ
utilized in this study to consider the longitudinal wall conduction
effect. To calculate the heat transfer coefficient of the hot air _ p ÞH ðT H;in
T C;out;RI Þ þ ðmc
ðmc _ p ÞC ðT C;out;RIII
T C;in Þ
stream in Region I, the Colburn j factor correlation for the trans- q_ tot ¼ ; ð33Þ
2
verse flow through the offset strip fins with bypass channels [22]
was used, which can be expressed by the following equation: q_ tot
e¼ ð34Þ
0
j ¼ /trans jch þ ð1
/trans Þjtrans ð22Þ C min ðT H;in
T C;in Þ

where
NuD0h 2=9 4. Results and discussion
9=2 7=5 7=5 45=14
jch ¼ 1=3
; jtrans ¼ ½ðjcf Þ þ fðjlam Þ þ ðjtur Þ g  ;
ReD0h Pr In this study, experiments were conducted by varying the inlet
 

1=3
4:93 dh ls temperature of the hot air stream and the mass flow rate for both
jcf ¼ 0 ; jtur ¼ 0:191 Redh ; the hot and cold air streams to investigate the pressure drop and
Redh Pr1=3 ðh þ dÞ dh
heat transfer characteristics of the fabricated unit recuperator with
(

1=3 )

0 
1=2
 offset strip fins, especially in the high-temperature range. The total
1 2h ls 1 pressure drop across the recuperator, excluding the losses due to
jlam ¼ 0:664 Redh þ 0:191 Redh 1

2 dh dh AER inlet and outlet headers, was measured for each air stream. The


1=2
 total heat transfer rate and the corresponding effectiveness were
2ðs þ dÞ 1
þ 0:664 Redh : respectively determined from the experimental measurements by
dh AER
the following equations:
In Eqs. (20) and (22), the values of f ReD0h and NuD0h can be _ p DTÞH þ ðmc
ðmc _ p DTÞC
q_ exp ¼ ; ð35Þ
obtained from Eqs. (9) and (10), respectively, when the aspect ratio 2
0
is a ¼ ðh þ dÞ=W out .
Once the effectiveness, including the longitudinal wall conduc- q_ exp
eexp ¼ : ð36Þ
tion effect, is obtained, the outlet temperatures for both the hot C min ðT H;in
T C;in Þ
and cold air streams and the heat transfer rate in Region I can be
Uncertainty analysis was also conducted for the mass flow
respectively evaluated as
rates, pressure drops, total heat transfer rate, and effectiveness.
C min The uncertainties of each parameter can be determined by ASME
T H;out;RI ¼ T H;in;RI
eRI ðT H;in;RI
T C;in;RI Þ; ð23Þ
Ch test uncertainty [24]. The uncertainties for the mass flow rates
and pressure drops obtained by measurements consist of the bias
C min and precision errors. The heat transfer rate and the effectiveness
T C;out;RI ¼ T C;in;RI þ eRI ðT H;in;RI
T C;in;RI Þ; ð24Þ
Cc can be calculated based on the measurement data using Eqs. (35)
and (36), respectively. On the basis of these equations, the uncer-
q_ RI ¼ eRI C min ðT H;in;RI
T C;in;RI Þ ð25Þ tainties of the total heat transfer rate and the effectiveness can
be calculated and expressed as follows:
Using the same method in Region I, the pressure drop for the vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2 ffi
u 2 
hot and cold air streams and the effectiveness in Region III can u @ q_ @ q_
þ @ðT H;in
T C;out Þ U ðT H;in
T C;out Þ
u @ðmc _ p ÞH U ðmc _ p ÞH
be obtained and the corresponding heat transfer rate and outlet U q_ ¼ u
t  2  2 ; ð37Þ
temperatures for each air stream are respectively expressed as þ @ðmc @ q_ _
þ @ðT C;out@ q
T Þ U ðT C;out
T C;in Þ
_ p Þ U ðmc _ p ÞC
follows: C C;in


 ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
s

G2H 4W out 1 2
2
2
DpH;RIII ¼ ð1
r2H þ K c;H Þ þ f H q ; ð26Þ @e @e @e
2qH;in dh;H H;in qH Ue ¼ U q_ þ U Cmin þ U ðT H;in
T C;in Þ :
@ q_
m
@C min @ðT H;in
T C;in Þ
"
 #
G02 4W 1 qH;in ð38Þ
0
DpC;RIII ¼ C
fC qC;in
ð1
r2C
K e;C Þ ; ð27Þ
2qC;in dh;C qC m qH;out
4.1. Pressure drop characteristics of the recuperator
q_ RIII ¼ eRIII C min ðT H;in;RIII
T C;in;RIII Þ; ð28Þ
Fig. 6(a) shows the experimental data of the total pressure drop
C min according to the hot air stream inlet temperature when the mass
T H;out;RIII ¼ T H;in;RIII
eRIII ðT H;in;RIII
T C;in;RIII Þ; ð29Þ
Ch flow rate and cold air stream inlet temperature are 4.0 g/s and
171 °C, respectively. The inlet temperature of the hot air stream
C min was varied from 250 to 500 °C. The black square scatters indicate
T C;out;RIII ¼ T C;in;RIII þ eRIII ðT H;in;RIII
T C;in;RIII Þ: ð30Þ
Cc the total pressure drop results of the cold air stream and the red
solid scatters represent those of the hot air stream. From the graph,
Once the pressure drop for the hot and cold air streams and the we can see that the pressure drop results for both the cold and hot
heat transfer rate in each region are obtained from Eqs. (16)–(19) air streams slightly increase with increasing hot air stream inlet
and Eqs. (25)–(28), the total pressure drop for both the hot and temperature, which is mainly due to the increment in the fluid
cold air streams and the total heat transfer rate from the modified mean temperature, as shown in Fig. 6(a). The total pressure drop
analytical model can be finally calculated. The corresponding effec- is proportional to the fluid mean specific volume ð1=qÞm , as
tiveness is expressed as following equations: expressed in Eq. (1). As the fluid mean temperature of each stream
increases, the corresponding fluid mean specific volume also
DpH ¼ DpH;RI =2 þ DpH;RII þ DpH;RIII ; ð31Þ increases. Therefore, the total pressure drops of both the cold
 K.H. Do et al. / International Journal of Heat and Mass Transfer 103 (2016) 457–467 465

Experimental data (mC = mH = 4.0 g/s, TC,in=171 C)

o
sharply increases, which in turn leads to the reduction in the total
18 Total pressure drop 700 pressure drop difference between the high- and ambient-
Δ PC Δ PH temperature conditions. Therefore, for a large mass flow rate
16
_ C ¼ 4:0 g=s, the difference between the pressure drop under the

Fluid mean temperature (TC or TH ) [ C]

Total pressure drop (Δ pC or Δ pH) [kPa]

Fluid mean temperature m

o
600
14 TC TH high-temperature condition and that under the ambient-
temperature condition is significantly reduced by approximately
12 500
55%. Consequently, the inlet pressure for each air stream should
10 be measured to correctly determine the pressure drop characteris-
400
8
tics of the recuperator correctly in the high-temperature range.

6 300
4.2. Heat transfer characteristics of the recuperator
4
200
Fig. 7(a) shows the experimental data for the effectiveness and
2
heat transfer rate according to the hot air stream inlet tempera-
0 100 ture. The effectiveness is almost constant regardless of the inlet
250 300 350 400 450 500
o
temperature of the hot air stream, which indicates that the fluid
Hot air stream inlet temperature (TH,in ) [ C]
mean temperature variation may not (or may scarcely) affect the
(a) Effect of the hot air stream inlet temperature effectiveness of the recuperator. The experimental data for the
effectiveness with respect to the mass flow rate are shown in
18 Fig. 7(b), which shows that the effectiveness decreases with
Experimental data for total pressure drop increasing mass flow rate. This result can be explained as follows:
16 o o
Hot test (TH,in=500 C, TC,in=171 C)
Total pressure drop (Δ pC or Δ pH) [kPa]

effectiveness is strongly related to NTU, which can be defined as a

14 Δ PC Δ PH function of the overall heat transfer coefficient, total heat transfer
o o
Cold test (TH,in=20 C, TC,in=20 C)
12
Δ PC Δ PH
10
0.80 1600
8

6 0.75 1400

4 0.70 1200

Heat transfer rate ( q ) [W]

2
Effectiveness (ε )

0.65 1000
0
1.0 1.5 2.0 2.5 3.0 3.5 4.0 0.60 800
Mass flow rate (mC or mH) [g/s]
0.55 600
(b) Effect of the mass flow rate
0.50 Experimental Data 400
Fig. 6. Experimental data of the total pressure drop of each air stream.
o
(mC = mH = 4.0 g/s, TC,in=171 C)
0.45 200
εexp qexp
and hot air streams become larger with increasing hot air stream
0.40 0
inlet temperature. 250 300 350 400 450 500
The experimental data of the total pressure drop with respect to Hot air stream inlet temperature (TH,in ) [ C]
o

the mass flow rate are shown in Fig. 6(b) at TH,in = 500 °C and
TC,in = 171 °C. The mass flow rate of both the cold and hot air (a) Effect of the hot air stream inlet temperature
streams was varied from 1.5 to 4.0 g/s. The pressure drop increases
with increasing mass flow rate, which results from the increase in 0.85 1600
the mass velocity for both the hot and cold air streams because,
0.80 1400
generally, the pressure drop is proportional to the square of the
mass velocity. The graph shown in Fig. 6(b) includes the experi-
0.75 1200
Heat transfer rate ( q ) [W]

mental data of the pressure drop under ambient-temperature con-

Effectiveness (ε )

dition. We can see from the graph that the values of the total 0.70 1000
pressure drop measured at high temperature are much larger than
those in ambient-temperature condition. These results primarily 0.65 800
come from the specific volume change. When air is assumed to
0.60 600
be an ideal gas, its specific volume is a function of the fluid temper-
ature as well as the absolute pressure of the fluid. By increasing the
0.55 Experimental Data 400
fluid temperature or decreasing the absolute pressure of the fluid, o o
(TH,in=500 C, TC,in=171 C)
the specific volume increases, and the pressure drop accordingly 0.50 200
εexp qexp
becomes larger, as expressed in Eq. (1). For a small mass flow rate
m_ C ¼ 2:0 g=s, the total pressure drop of the cold air stream 0.45 0
1.5 2.0 2.5 3.0 3.5 4.0
obtained from high-temperature condition is approximately 2.1
Mass flow rate (mC or mH) [g/s]
times larger than that from ambient-temperature condition. As
the mass flow rate increases, the fluid mean temperature differ- (b) Effect of the mass flow rate
ence remains nearly constant, but the absolute pressure difference
between the cold and the hot test results of the cold air stream Fig. 7. Experimental data of the effectiveness and heat transfer rate.
466 K.H. Do et al. / International Journal of Heat and Mass Transfer 103 (2016) 457–467

surface area, and minimum heat capacity rate expressed in Eq. (6). 0.90
With increasing mass flow rate, the increment of the heat capacity
rate is dominant compared with that of the heat transfer coeffi-
0.85
cient even though the overall heat transfer coefficient and heat
capacity rate simultaneously increase. As a result, the effectiveness
declines due to the decrement in the NTU value. 0.80

Effectiveness (ε )
4.3. Comparison between the experimental data and model prediction 0.75
results
0.70
In the present study, two analytical models were developed to o
Effectiveness results (mC = mH = 4.0 g/s, TC,in=171 C)
predict the pressure drop and heat transfer characteristics in the
εexp
high-temperature range and the proposed model results were 0.65
compared with the experimental data. Figs. 8 and 9 show the com- Simple model ( j correlation from [14])
parison of the experimental and model prediction results of the Modified model ( j correlation [16], j ' correlation [22])
0.60
total pressure drop and effectiveness, respectively. In Fig. 8, the
250 300 350 400 450 500
black lines indicate the total pressure drop of the cold air stream, o
and the red lines represent that of the hot air stream. In addition, Hot air stream inlet temperature (TH,in ) [ C]
the dashed lines represent the results obtained from the simple (a) Effectiveness according to the hot air
analytical model based on the friction factor correlation [16] in stream inlet temperature

0.90
24
o
Total pressure drop results (mC = mH = 4.0 g/s, TC,in=171 C)
22
Δ pC,exp Δ pH,exp Δ pC,model Δ pH,model 0.85
Total pressure drop (Δ pC or Δ pH) [kPa]

20
Simple model ( f correlation from [16])
18 Simple model ( f correlation from [14])
0.80
Modified model ( f correlation [16], f ' correlation [22])
Effectiveness (ε )

16

14 Hot air stream

Cold air stream
0.75
12

10

8 0.70 o o
Effectiveness results (TH,in=500 C, TC,in=171 C)
6
εexp
Cold air stream
4 0.65
Simple model ( j correlation from [14])
2 Modified model ( j correlation [16], j ' correlation [22])
Hot air stream
0
0.60
250 300 350 400 450 500
1.5 2.0 2.5 3.0 3.5 4.0
o
Hot air stream inlet temperature (TH,in ) [ C]
Mass flow rate (mC or mH) [g/s]
(a) Effectiveness according to the hot air (b) Effectiveness according to the mass flow rate
stream inlet temperature
Fig. 9. Comparison of the experimental and model prediction results of the
24 effectiveness.
o o
22 Total pressure drop results (TH,in=500 C, TC,in=171 C)
Δ pC,exp Δ pH,exp Δ pC,model Δ pH,model
Total pressure drop (Δ pC or Δ pH) [kPa]

20 Eq. (4), and the dotted lines represent those from the f correlation
Simple model ( f correlation from [16])
18 [14] expressed in Eq. (5). The solid lines indicate the pressure drop
Simple model ( f correlation from [14])
16 Modified model ( f correlation [16], f ' correlation [22]) prediction results evaluated from the modified analytical model
which is applied to the f correlation of the longitudinal flow
14
through the offset strip fins in Eq. (4) in the core region and to
12 0
Hot air stream the f correlation of the transverse flow through the offset strip fins
10 Cold air stream [22] in Eq. (20) in the outlet region of the recuperator, respectively.
8 We find that the simple model prediction using the different fric-
6 tion factor correlations in Eqs. (4) and (5) yield similar results.
Cold air stream
Fig. 8 shows the total pressure drop results obtained from the sim-
4
ple model significantly underestimate those from the experimental
2
Hot air stream
data regardless of the hot air stream inlet temperature as well as
0 the mass flow rate. The maximum relative error is approximately
1.5 2.0 2.5 3.0 3.5 4.0 80%. On the other hand, the total pressure drop results obtained
Mass flow rate (mC or mH) [g/s] from the modified model are in fairly good agreement with those
from the experimental data in all test configurations although
(b) Effectiveness according to the mass flow rate
there is a slight discrepancy between the experimental data and
Fig. 8. Comparison of the experimental and model prediction results of the the results obtained from the modified model, which may come
pressure drop. from the error of the friction factor correlation of the transverse
 K.H. Do et al. / International Journal of Heat and Mass Transfer 103 (2016) 457–467
flow through the offset strip fins developed by Muzychka and
Yovanovich [22] and the effect of the sub-channel shape of the fab-
ricated unit recuperator. As a result, these results show that the
pressure drop near the outlet region in each air stream cannot be
neglected and should be properly considered.
The comparison of the experimental and model prediction
results of the effectiveness is shown in Fig. 9. The black dashed
lines indicate the effectiveness results obtained from the simple
model based on the Colburn j factor correlation [14] in Eq. (8),
and the red solid lines indicate those from the modified model
0
using the j correlation in Eq. (7) in the core region [16] and j cor-
relation in Eq. (22) in the outlet region of the recuperator [22],
respectively. Fig. 9 shows that the effectiveness results obtained
from the simple model largely overestimate those from the exper-
imental data with respect to the hot air stream inlet temperature
and the mass flow rate. On the other hand, excellent agreement
between the modified model prediction and the tested results is
achieved within a maximum error of 2.2%. As a result, the predic-
tion results using the simple model may yield higher thermal effi-
ciency of the micro gas turbine because the thermal efficiency for
micro gas turbine increases by increasing the recuperator effec-
tiveness or decreasing the pressure drop of the recuperator. Conse-
quently, we show that the modified model proposed in the present
work successfully estimates the pressure drop and heat transfer
characteristics of the fabricated recuperator with offset strip fins
for a micro gas turbine.
5. Conclusion
In this study, the pressure drop and heat transfer characteristics
of a recuperator with offset strip fins for a micro gas turbine were
experimentally investigated in the high-temperature range. A unit
recuperator with offset strip fins was fabricated using furnace-
brazing technology in vacuum atmosphere. Experiments were con-
ducted by varying the mass flow rate and the inlet temperature of
the hot air stream in the range of 1.5
4 g/s and 250
500 °C,
respectively. The total pressure drop slightly increased with the
increase in the hot air stream inlet temperature due to the incre-
ment in the fluid mean specific volume. The pressure drop
increased with the increase the mass flow rate, and the values of
the total pressure drop measured at high temperature were much
larger than those under ambient-temperature condition. The
experimental data showed that the inlet pressure of each air
stream should be measured to correctly determine the pressure
drop characteristics of the recuperator in the high-temperature
range. The effectiveness was almost constant regardless of the inlet
temperature of the hot air stream, which means that the fluid
mean temperature variation hardly affected the effectiveness of
the recuperator.
A simple analytical model that applies a simplified flow config-
uration and a modified analytical model that considers the actual
flow configuration of the fabricated recuperator were proposed
to predict the pressure drop and the effectiveness of the fabricated
recuperator. The proposed model prediction results were also com-
pared with the experimental data. The comparison with the exper-
imental data showed that the results from the simple model may
lead to incorrect results of the thermal efficiency of the micro
gas turbine because the recuperator effectiveness was overesti-
mated and the pressure drop of the recuperator was underesti-
mated. On the other hand, good agreement between the
modified model prediction and the tested results was achieved.
Consequently, the modified model suggested in the present work
467
successfully estimated the pressure drop and heat transfer charac-
teristics of the fabricated recuperator with offset strip fins for a
micro gas turbine.
Acknowledgements
This work was supported by the principal research programs of
the Korea Institute of Machinery and Materials (KIMM), Republic of
Korea.
References
[1] T. Nagashima, Lessons learnt from the ultra-micro gas turbine development at
University of Tokyo, Micro Gas Turbines (2005) 14-1–14-58. Educational Notes
RTOEN-AVT-131.
[2] O. Dessornes, S. Landais, R. Valle, A. Fourmaux, S. Burguburu, C. Zwyssig, Z.
Kozanecki, Advances in the Development of a Microturbine Engine, J. Eng. Gas
Turbines Power 136 (2014) 071201.
[3] A.H. Epstein, S.A. Jacobson, J. Protz, L.G. Fréchette, Shirtbutton-sized gas
turbines: the engineering challenges of micro high speed rotating machinery,
Proceedings of the 8th International Symposium on Transport Phenomena and
Dynamics of Rotating Machinery, Honolulu, Hawaii, 2000, pp. 26–30.
[4] S. Kang, S.-J.J. Lee, F.B. Prinz, Size does matter, the pros and cons of
miniaturization, ABB Rev. 2 (2001) 54–62.
[5] K. Isomura, M. Murayama, H. Yamaguchi, N. Ijichi, N. Saji, O. Shiga, S. Tanaka, T.
Genda, M. Hara, M. Esashi, Component Development of Micromachined Gas
Turbine Generators, Technical Digest Power MEMS 2002, International
Workshop on Power MEMS, Tsukuba, Japan, 12–13 November 2002, 2002,
pp. 32–35.
[6] E. Matsuo, H. Yoshiki, T. Nagashima, C. Kato, Development of Ultra Gas
Turbines, Technical Digest Power MEMES 2002, International Workshop on
Power MEMES, Tsukuba, Japan, 12–13 November 2002, 2002, pp. 36–39.
[7] T. Stevens, M. Baelmans, Optimal pressure drop ratio for micro recuperators in
small sized gas turbines, Appl. Therm. Eng. 28 (2008) 2353–2359.
[8] G. Lagerstrom, M. Xie, High performance and cost effective recuperator for
micro-gas turbines, Proceedings of ASME Turbo Expo 2002, 2002. GT-2002-
30402.
[9] H. Shih, Y.C. Huang, Thermal design and model analysis of the Swiss-roll
recuperator for an innovative micro gas turbine, Appl. Therm. Eng. 29 (2009)
1493–1499.
[10] M.L. Ferrari, M. Pascenti, L. Magistri, A.F. Massardo, Micro gas turbine
recuperator: steady-state and transient experimental investigation, J. Eng.
Gas Turbines Power 132 (2010) 022301.
[11] W.M. Kays, A.L. London, Compact Heat Exchangers, 3rd ed., McGraw-Hill, New
York, 1984.
[12] S.V. Manson, Correlations of heat transfer data and of friction data for
interrupted plate fins staggered in successive rows, Washington, DC:NACA
Tech, Note 2237, National Advisory Committee for Aeronautics, 1950.
[13] A.R. Wieting, Empirical correlations for heat transfer and flow friction
characteristics of rectangular offset-fin plate-fin heat exchangers, ASME J.
Heat Transfer 97 (1975) 480–490.
[14] Y.S. Muzychka, M.M. Yovanovich, Modeling the f and j characteristics of the
offset strip fin array, Enhanced Heat Transfer 8 (2001) 261–277.
[15] R.L. Webb, H.M. Joshi, Heat transfer and friction in the offset-strip fin heat
exchangers, Int. J. Heat Mass Transfer 30 (1987) 69–84.
[16] R.M. Manglik, A.E. Bergles, Heat transfer and pressure drop correlations for the
rectangular offset strip fin compact heat exchanger, Int. J. Exp. Therm. Fluid
Sci. 10 (1995) 171–180.
[17] K.H. Do, T.H. Kim, Y.S. Han, B.-I. Choi, M.B. Kim, Comparison of Thermal
Performance of Recuperators with Offset Strip and Plain Fins, Proceedings of
23rd International Symposium on Transport Phenomena Auckland, New
Zealand, 2012.
[18] Y.W. Lee, J.H. Kim, Y.S. Song, C.S. Seok, Mechanical strength and
microstructural characterization of SUS304 brazed joints with BNi-2 filler
metal: influence of brazing temperature and time, Solid State Phenom. 124
(2007) 1673–1676.
[19] F.M. White, Fluid Mechanics, McGraw-Hill, New York, 1987.
[20] R.K. Shah, A.L. London, Laminar Flow Forced Convection in Ducts, Supplement
1 to Advances in Heat Transfer, Academic Press, New York, 1978.
[21] P.G. Kroeger, Performance deterioration in high effectiveness heat exchangers
due to axial heat conduction effects, Adv. Cryogen. Eng. 12 (1967) 363–372.
[22] Y.S. Muzychka, M.M. Yovanovich, Modeling the f and j characteristics transvers
flow through an offset strip fin at low reynolds number, Enhanced Heat
Transfer 8 (2001) 243–259.
[23] R.K. Shah, A.C. Mueller, Heat exchanger basic thermal design methods, in: W.
M. Rohsenow, J.P. Hartnett, E.N. Ganic (Eds.), Handbook of Heat Transfer
Applications, McGraw-Hill, New York, 1985, pp. 4-1–4-77.
[24] ASME, Test uncertainty, ASME PTC 19.1-2005, 2005.