Annals of Nuclear Energy 94 (2016) 213–222

Contents lists available at ScienceDirect

Annals of Nuclear Energy

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

Validation of the finned sodium–air heat exchanger model in MARS-LMR

C.W. Choi ⇑, K.S. Ha
Sodium-cooled Fast Reactor Design Division, Korea Atomic Energy Research Institute (KAERI), 989-111, Daedeok-Daero, Yuseong-Gu, Daejeon 305-353, Republic of Korea

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

Article history: A Prototype Gen-IV Sodium-cooled Fast Reactor (PGSFR), for which the Korea Atomic Energy Research
Received 17 September 2015 Institute (KAERI) has designed a pool type sodium-cooled fast reactor, has a decay heat removal system
Received in revised form 22 February 2016 (DHRS). The DHRS consists of sodium-to-sodium and sodium-to-air heat exchangers and their connecting
Accepted 22 February 2016
pipes. There are four loops, which have two active and two passive type sodium–air heat exchangers, that
are called a finned-tube sodium-to-air heat exchanger (FHX) and a helical-tube sodium-to-air heat
exchanger (AHX), respectively. Recently, Zukauskaus’s air–sodium heat transfer models have been added
Keywords:
in MARS-LMR, which is a safety analysis code for the PGSFR. In this study, to validate the newly added
MARS-LMR
Finned sodium–air heat exchanger
heat transfer models for the FHX, two experiments are selected: one is a performance test for a sodium
Sodium-cooled fast reactor to air heat exchanger (AHX) in the steam generator test facility (SGTF) for the prototype fast breeder reac-
PFBR tor (PFBR), and the other is a start-up test with a dump heat exchanger (DHX) in JOYO. All validation
JOYO results indicate that Zukauskaus’s correlation slightly over-estimates the heat transfer. One possible rea-
sons is a smaller number of rows in the test bundle, which was already mentioned by Zukauskaus and
Karni. The RMS values for the prediction of sodium temperature for the PFBR and the JOYO are 16.25%
and 27.5%, respectively. When a correction factor is applied, their RMS values improve to 3.19% and
20.71%, respectively. In addition, the MARS-LMR’s prediction for the JOYO shows a much better accuracy
with RMS of 7.45% when corrected sodium flow rates are applied.
Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction FHX is a finned serpentine-type air-sodium heat exchanger in the

The Korea Atomic Energy Research Institute (KAERI) has
designed a pool type sodium-cooled fast reactor, which is called
the Prototype Gen-IV Sodium-cooled Fast Reactor (PGSFR) (Kim
et al., 2013). Fig. 1 shows a schematic of the PGSFR. The PGSFR
has a decay heat removal system (DHRS) has independently pas-
sive and active loops to satisfying a system diversity. The decay
heat removal system (DHRS) is a designated safety grade compo-
nent providing a sufficient decay heat removal capability during
abnormal conditions, such as a loss of heat sink (LOHS) accident.
The passive DHRS (PDHRS) relies exclusively on a natural convec-
tion heat transfer, i.e., natural circulation on the sodium side and
natural draft on the air side. And its loop is equipped with one
sodium-to-sodium decay heat exchanger (DHX), natural-draft
sodium-to-air heat exchanger (AHX), and connecting pipes. The
active DHRS (ADHRS) is operated by a EM-pump in a loop-side
and a blower in the air-side. And its loop is integrated with same
components in the PDHRS, except a forced-draft sodium-to-air
heat exchanger (FHX) for an air-side heat exchanger. The AHX is
a helical type air-sodium heat exchanger in the PDHRS, and the
⇑ Corresponding author.
ADHRS. The DHX located in a cold pool, removes heat from a pri-
mary side and transfers heat to the loop side in the DHRS. And heat
in the loop is transferred to ultimate heat sink of ambient air by
capability of the AHX or FHX, which depends on the opening area
of the damper at the inlet region. Under normal operation, a small
amount of heat is removed by the AHX and FHX through a small
open area of the damper. It was designed to prevent the solidifica-
tion of sodium, and to ensure operability when it fully operates
during accident conditions. The current heat removal capacities
are about 0.3 MW and 2.5 MW for normal and accident conditions,
respectively.
The MARS-LMR code has been used for a safety analysis of the
PGSFR. The code is based on the MARS which was developed for
the transient analysis of a light water reactor. The MARS code basi-
cally employs the three-dimensional transient two-fluid model for
the two-phase flow system (Jeong et al., 1999). Also the point
kinetics equation and the heat conduction equation are modeled
to calculate the neutron behavior in the reactor core and the heat
transfer from the heat structure to fluid. This code was modified
to simulate the sodium thermal–hydraulics and neutronic behav-
iors in a transient condition for a liquid metal cooled fast reactor
(Ha and Jeong, 2010). Recently, the heat transfer models for the

http://dx.doi.org/10.1016/j.anucene.2016.02.020
0306-4549/Ó 2016 Elsevier Ltd. All rights reserved.
214 C.W. Choi, K.S. Ha / Annals of Nuclear Energy 94 (2016) 213–222

Nomenclature

Nu Nusselt number ST transversal pitch (m)

Pr fluid Prandtl number SL longitudinal pitch (m)
Prw wall Prandtl number SD diagonal pitch (m)
ReD,max bundle Reynolds number N number of tubes
Re Reynolds number C2 empirical coefficient in Eq. (7)
C1 empirical coefficient in Eq. (1) a empirical exponent in Eq. (7)
Cz correction factor in Fig. 6 b empirical exponent in Eq. (7)
m empirical exponent in Eq. (1) h fin height (m)
V fluid velocity (m/s) s fin pitch (m)
Vmax maximum fluid velocity in a bundle (m/s) Afin/fluid area ratio of fin and fluid
D tube outer diameter (m) d fin thickness (m)
t tube thickness (m) DT temperature difference of inlet and outlet (°C)
q fluid density (m3/kg)
l fluid viscosity (Pa
s)

Fig. 1. Schematics of the PGSFR.

 C.W. Choi, K.S. Ha / Annals of Nuclear Energy 94 (2016) 213–222 215

heat exchangers in the DHRS were added in the MARS-LMR.

Therefore, the validation of a new heat transfer model is essential
to analyze the characteristics of the DHRS during an accident.
A large-sodium thermal test program called a sodium test loop
for a safety simulation and assessment (STELLA) is being pro-
gressed by KAERI. As the first step of the program, the sodium com-
ponent test loop called STELLA-1 has been completed, which is to
be used for demonstrating the thermal–hydraulic performance of
major components such as heat exchangers and a mechanical
sodium pump and providing validation and verification (V&V) data
for their design and safety analysis codes. As a second step of the
program an integral effect test loop, called STELLA-2, will be con-
structed to demonstrate the plant safety and support the design
approval for the prototype reactor (Eoh et al., 2013). The perfor-
mance test for the FHX will be conducted in the STELLA-2 facilities.
Fig. 2 shows the test component of the FHX in the STELLA-2 facil-
ities. Before obtaining the experimental data for the FHX from the
STELLA-2 facilities, the validation test of MARS-LMR code for the
FHX is conducted using existing experimental data.
2. Heat transfer models for a finned heat exchanger
In order to simulate a sodium–air heat exchanger bundle, corre-
lations for each of tube- and shell-sides are necessary. The convec-
tive heat transfer correlation in the tube-side was already added in
MARS-LMR with Aoki’s correlation (Aoki, 1973). A major concern of
this study is a convective heat transfer in the shell-side of a cross-
flow bundle, since Zukauskas’ correlations for shell-side air con-
vective heat transfer are recently added in MARS-LMR
(Zukauskas and Karni, 1989). The tube rows of a bank can be either
in-lined or staggered in the direction of the fluid flow. The config-
uration of a bundle is characterized by the tube diameter, D, and by
the transverse pitch, ST, and longitudinal pitch, SL, measured
between the tube centers, as shown in Fig. 3. The flow around
the tubes in first row of a tube bank is similar to that for a single
SL SL
(a) (b)
ST ST
A1
SD
A1
A2
Fig. 3. Tube arrangement in a bank: (a) in-lined, and (b) staggered.
cylinder in a cross flow. Correspondingly, the heat transfer coeffi-
cient for a tube in the first row is approximately equal to that for
a single tube in a cross flow. For downstream rows, the flow condi-
tions depend strongly on the tube bank arrangement. In-lined
tubes beyond the first row reside in the wakes of upstream tubes,
and for moderate values of SL the heat transfer coefficients associ-
ated with downstream rows are enhanced by mixing of the flow.
Typically, the heat transfer coefficient of a row enhanced with
increasing row number until approximately the fifth row, after that
there is little change in the flow conditions. For large SL, the
influence of reduction of the upstream rows deteriorates the heat
transfer in the downstream rows. For this reason, the operation
of in-lined tube bank with ST/SL < 0.7 is undesirable. For the stag-
gered tube array, the path of the main flow is more tortuous, and
the mixing of the cross-flowing fluid is increased relative to the
in-lined tube arrangement. In general, heat transfer enhancement
is favored by the more tortuous flow of a staggered arrangement,
particularly for small Reynolds numbers (Re < 100).
Zukauskas proposed a wide range of various correlations for a
bundle flow with the following form (Zukauskas and Karni, 1989):

0:25
Pr
Nu ¼ C 1 Rem
D;max Pr0:36 ð1Þ
Prw

where C1 is an empirical coefficient, Prf is the Prandtl number with

bulk temperature, Prw is the Prandtl number with the wall temper-
ature, and m is an exponent of the Reynolds number, which is
defined with the maximum velocity and single tube outer diameter.

qV max D
ReD;max ¼ ð2Þ
l
C1 and m are proposed with the Reynolds number for the arrange-
ment type, as shown in Table 1

Table 1
Constants of Zukauskas’ bundle correlation of Eq. (1).

Configuration ReD,max C1 m
In-lined 1.6–100 0.9 0.4
100–1000 0.52 0.5
1000–2  105 0.27 0.63
2  105–2  106 0.033 0.8
Staggered 1.6–40 1.04 0.4
40–1000 0.71 0.5
y
1000–2  105 0.35(ST/SL)0.2a 0.6
1000–2  105 0.4 0.6
2  105–2  106 0.031 0.8
a
Which is case of ST/SL < 0.2.
Fig. 2. Schematic of the FHX in the STELLA-2 facilities.
216 C.W. Choi, K.S. Ha / Annals of Nuclear Energy 94 (2016) 213–222

For an in-lined arrangement, Vmax occurs at the transverse plane

in Fig. 3, and from the mass conservation requirement for an
s
incompressible fluid.
ST
V max ¼ V ð3Þ
ST  D
For a staggered configuration, the maximum velocity may occur
at either the transverse plane A1 or the diagonal plane A2 in Fig. 3. It
will occur at A2 if the rows are spaced such that SL
2ðSD  DÞ < ST  D ð4Þ
The factor of 2 results from the bifurcation experienced by the h
fluid moving from the A1 to A2 planes. Hence, Vmax occurs at A2 if
"
2 #0:5
ST ST þ D
SD ¼ S2L þ < ð5Þ
2 2

in which case it is given by

ST Fig. 4. Schematic of unit fin between two tubes.
V max ¼ V
ð6Þ
2ðSD  DÞ
For a staggered finned bundle, Zukauskas also proposed correla-
tions, as defined in Eq. (7):

0:2  0:18
0:14
0:25
ST s h Pr
Nu ¼ C 2 ReaD;max Prb ð7Þ
SL D D Prw
where C2 is an empirical coefficient, s is a fin pitch, h is a fin height,
and m and n are exponents for Reynolds and Prandtl numbers,
respectively. Constants for this correlation are summarized in
Table 2. Fig. 4 shows a schematic of a unit fin. Reynolds number,
ReD,max, in Eq. (7) is also based on the maximum fluid velocity
occurring within the tube bank. Therefore, based on the geometrical
effect in Fig. 4, the ratio of the unit fin and fluid areas are defined as
Eq. (8) and used as a weighting factor for an additionally increased
velocity.
2
h
d
Afin=fluid ¼ 1  ð8Þ
ðminðST ; SL ; SD Þ  DÞ
s
Fig. 5. Sodium to air heat exchanger (AHX) in the PFBR (Vinod et al., 2013).

3. Prototype fast breeder reactor (PFBR) experiment

Table 3
Experimental data of the AHX in the SGTF (Nanashima et al., 1979).
3.1. Air-to-sodium heat exchanger (AHX) in the PFBR
Design parameter [unit] Value

A prototype fast breeder reactor (PFBR) also has four loops of Tube ID [mm] 32.9
decay heat removal system called a safety grade decay heat Tube OD, D [mm] 38.1
Tube thickness, t [mm] 2.6
removal system (SGDHRS). The decay heat removal rate is
Transverse pitch, ST [mm] 85 (106)
24 MW, which is approximately 2% of the nominal thermal power Longitudinal pitch, SL [mm] (70)
of the PFBR. The sodium-to-air heat exchanger (AHX) in the PFBR is Fin height, h [mm] 13
quite similar to the FHX in the PGSFR. A heat transfer performance Fin thickness, d [mm] 1.22
test for the AHX is conducted in the steam generator test facility Fin pitch, s [mm] 5.1059
Number of tubes, N [ea] 22
(SGTF) at the Indira Gandhi Centre for Atomic Research (Vinod Number of pass 6
et al., 2013). And they compared the experimental results with Sodium flow [kg/s] 4.64–8.27
the heat transfer model, which used Skupinski and Vuatrey’s Air flow [kg/s] 5.14–11.06
(Skupinski et al., 1965) and Zukauskas’ correlations for the tube- Sodium inlet temperature [°C] 525–537.5
Air inlet temperature [°C] 40
side and shell-side, respectively. They reported that the prediction
by the heat transfer model was agreed well with experimental

results. This result is promising since the resistance to heat transfer

Table 2 offered by the sodium tube-side is extremely low compared to the
Constants of Zukauskas’ finned bundle correlation of Eq. (7). convective heat transfer resistance of air flowing across the finned
ReD,max C2 a b tube. Fig. 5 shows the test heat exchanger used in the test facility.
100–2  104 0.192 0.65 0.36 Table 3 shows the design data of the test heat exchanger. From Ref.
2  104–2  105 0.0507 0.8 0.4 5, eight experimental datasets were obtained and are summarized
2  105–1.4  106 0.0081 0.95 0.4 in Table 4.
 C.W. Choi, K.S. Ha / Annals of Nuclear Energy 94 (2016) 213–222 217

Table 4
Experimental results obtained from the test of the AHX in the PFBR (Vinod et al., 2013).

CASE mNa [kg/s] mair [kg/s] TNa,inlet [°C] TNa,outlet [°C] Tair,inlet [°C] Tair,outlet [°C] qNa [MW] qair [MW]
1 8.17 10.82 531.2 331.4 40.0 238.0 2.26 2.21
2 8.27 10.88 528.8 327.6 40.0 237.0 2.3 2.21
3 8.16 10.85 531.8 331.5 40.0 240.8 2.26 2.25
4 6.66 8.17 531.7 331.8 40.0 259.6 1.79 1.85
5 4.64 5.29 537.5 326.5 40.0 297.4 1.35 1.41
6 4.68 5.14 534.0 330.5 40.0 303.4 1.32 1.40
7 7.75 11.06 525.0 310.0 40.0 239.0 2.3 2.27
8 7.67 7.96 525.0 351.0 40.0 261.0 1.86 1.82

transfer region, the differences of sodium and air outlet tempera-

Fig. 6. Correction factor for heat transfer from (1) staggered and (2) in-lined finned
tube bundle with small number of tubes (Zukauskas and Karni, 1989).
3.2. MARS-LMR results for the AHX
Using information described in Table 3, a finned tube bundle is
modeled for the MARS-LMR input. First of all, the node dependency
test is carried out with the heat exchanger represented by various
number of nodes. Tests were performed with as few as 10 nodes
and as many as 50 nodes. When 50 nodes are applied to the heat
tures from the case of 40 nodes were less than 0.1 °C. Therefore,
the number of nodes of 50 was selected for all validation calcula-
tions. There is a treatment for the input parameter of the heat
transfer area. The heat transfer areas of the tube-side and shell-
side should be the same in the heat structure input of MARS-
LMR. However, the heat transfer area in a finned tube is drastically
larger than that in the in-tube-side. Therefore, tube-side heat
transfer area is identically applied to heat transfer areas in both
tube- and shell-sides. And the multiplication factor for the area
ratio between the finned area and the in-tube area is applied to
the shell-side. Here, the finned area was 7.8-times larger than
the in-tube area. Considering the fin efficiency of 0.84, the multipli-
cation factor was calculated as 6.66 for the AHX.
Zukauskas and Karni reported that a smaller number of tube
rows can deteriorate the heat transfer (Zukauskas and Karni,
1989), and proposed a correction factor, Cz corresponding to the
number of tube rows as shown in Fig. 6. The correction factor
indicates that lower number of the tube rows deteriorates or
enhances a heat transfer. In other words, when the number of tube
rows are less than 4, the correction factor should be applied in the
Zukauskas’ correlation. The test heat exchanger in the SGTF has only
a single row for one pass, as shown in Fig. 5. Thus, an appropriate
correction factor is necessary to evaluate an effective heat transfer

Fig. 7. Comparison of tube and shell-side temperature difference between MARS-LMR calculation and PFBR-AHX experimental results.
218 C.W. Choi, K.S. Ha / Annals of Nuclear Energy 94 (2016) 213–222

Fig. 8. Comparison of heat transfer rate between MARS-LMR calculation and PFBR-AHX experimental Results.

Table 5 coefficient. Figs. 7 and 8 show a temperature difference between the

Deviation of MARS-LMR prediction for the PFBR-AHX experiments.
CASE Parameter
BASE Sodium temperature [°C]
Air temperature [°C]
Heat transfer rate [MW]
Correction Sodium temperature [°C]
0.77 Air temperature [°C]
Heat transfer rate [MW]
inlet and outlet for sodium and air sides and the heat transfer
RMS (%) amount, respectively. The MARS-LMR results indicate that the
sodium-side temperature shows a better prediction than the air-
16.25
7.03
side temperature. Based on Fig. 6, the test cases are recalculated
10.06 with the correction factor of 0.77. The reference case shows a pre-
3.19
diction of the higher heat transfer rate and the corrected results
8.47 show a slightly under-estimated heat transfer rate, as shown in
5.18 Fig. 8. However, the corrected results have more accurate predic-
tion. The heat transfer rates of the reference and the corrected cases

Fig. 9. Cooling system of JOYO (Nanashima et al., 1979).

 C.W. Choi, K.S. Ha / Annals of Nuclear Energy 94 (2016) 213–222 219

Fig. 10. Schematic of the DHX in JOYO (Doi et al., 1980).

Table 6
Experimental data of the DHX in the JOYO (Doi et al., 1980).

Design parameter [unit] Value Table 8

Deviation of MARS-LMR prediction for the JOYO-DHX experiments.
Tube ID [mm] 38.7
Tube OD, D [mm] 42.7 CASE Parameter RMS (%)
Tube thickness, t [mm] 2.0 BASE Sodium temperature [°C] 27.5
Transverse pitch, ST [mm] 86 Air temperature [°C] 6.78
Longitudinal pitch, SL [mm] 74.5 Heat transfer rate [MW] 8.01
Fin height, h [mm] 19
Fin thickness, d [mm] 1.6 Correction Sodium temperature [°C] 20.71
Fin pitch, s [mm] 5.08 0.88 Air temperature [°C] 8.07
Number of tubes, N [ea] 80 Heat transfer rate [MW] 7.26
Number of pass 2 Corrected Na flow Sodium temperature [°C] 7.45
Sodium flow [kg/s] 149.5–157.9 Air temperature [°C] 8.01
Air flow [kg/s] 15.784–154.79 Heat transfer rate [MW] 7.49
Sodium inlet temperature [°C] 567.–537.5
Air inlet temperature [°C] 40

Table 7
Experimental results obtained from the test of the DHX in the JOYO (Nanashima et al., 1979; Doi et al., 1980).

CASE mNa [kg/s] mair [kg/s] TNa,inlet [°C] TNa,outlet [°C] Tair,inlet [°C] Tair,outlet [°C] qNa [MW] qair [MW]
1 154.125 19.888 386.2 356.2 25.93 320.52 5.94 5.99
2 154.401 42.588 399.1 349.9 23.59 282.17 9.75 11.22
3 153.996 56.598 408.0 347.2 23.38 263.25 12.01 13.81
4 154.724 82.843 423.3 343.5 23.9 236.75 15.83 17.90
5 154.985 106.282 432.6 338.4 19.7 217.28 18.71 21.29
6 156.799 51.937 294.7 254.5 28.3 198.57 8.27 8.96
7 157.161 104.478 305.5 245.5 31.41 166.42 12.37 14.27
8 156.966 140.844 315.0 241.2 30.88 153.24 15.18 17.43
9 157.079 154.794 320.8 240.2 25.69 147.69 16.58 19.08
10 150.985 18.601 391.3 366.0 26.52 335.33 4.90 5.87
11 151.180 35.552 406.8 363.6 24.16 303.34 8.37 10.12
12 150.605 48.389 416.2 361.2 25 282.89 10.61 12.71
13 151.498 73.017 432.4 358.5 24.08 253.25 14.33 17.01
14 151.712 88.683 443.1 357.2 19.7 238.04 16.67 19.65
15 157.658 46.259 298.3 263.8 28.47 211.69 7.13 8.59
16 157.947 90.436 313.9 258.3 31.23 178.93 11.49 13.52
17 157.646 115.471 326.0 258.4 30.28 168.78 13.93 16.18
18 157.604 134.433 332.6 256.4 25.53 159.77 15.68 18.25
19 150.119 23.090 390.6 368.8 28.4 311.6 4.20 6.67
20 150.161 20.535 390.5 370.8 25.9 315.9 3.80 6.08
21 150.148 25.315 395.9 369.2 24.4 311.9 5.14 7.43
22 149.902 19.213 397.4 370.9 27 310 5.10 5.55
23 150.730 31.277 389.5 352.3 30.6 283 7.21 8.04
24 149.811 19.291 396.2 369.4 23 309.3 5.15 5.63
25 149.520 34.234 412.0 368.7 20.5 291 8.29 9.43
26 149.839 37.627 412.1 366.1 31.8 287.6 8.83 9.81
27 150.315 48.695 418.7 364.0 29.9 273.3 10.53 12.07
28 149.986 52.344 422.6 364.8 29.8 270.5 11.10 12.82
220 C.W. Choi, K.S. Ha / Annals of Nuclear Energy 94 (2016) 213–222

Fig. 11. Comparison of tube and shell-side temperature difference between MARS-LMR calculation and JOYO-DHX experimental results.

Fig. 12. Comparison of heat transfer rate between MARS-LMR calculation and JOYO-DHX experimental results.

are about 10% and 5%, respectively. Deviations between MARS-LMR
and the experiments are summarized in Table 5.
4. JOYO experiments
4.1. Dump heat exchanger (DHX) in JOYO
Nanashima et al. and Doi et al. conducted a JOYO start-up test
to confirm whether the heat transfer characteristics of the
intermediate heat exchanger (IHX) and dump heat exchanger
(DHX) satisfy the design values (Nanashima et al., 1979; Doi
et al., 1980). Fig. 9 shows a schematic of the JOYO cooling system,
which has two intermediate heat exchangers (IHX) and four dump
heat exchangers (DHX). Thus, two DHXs and one IHX connected to
a single loop, and there are two loops named A and B. The detailed
tube configuration is shown in Fig. 10 (see Table 6). The DHX in
JOYO is a sodium–air staggered finned tube heat exchanger, which
is the same type as the FHX in the PGSFR and the AHX in PFBR. The
 C.W. Choi, K.S. Ha / Annals of Nuclear Energy 94 (2016) 213–222
Fig. 13. Corrected sodium flow rates based on air-side heat transfer rat.
hot sodium flow from one IHX is bifurcated into two DHXs and
heat is transferred through finned tubes. The air flow in each
DHX shell is driven by an individual blower. Therefore, the single
sodium and two air inlet temperatures and flow rates are mea-
sured in each DHX loop. The sodium flow rate in each tube-side
was not measured independently. So, it is assumed that the
tube-side flow rate in each DHX is 50% of the measured total flow
rate. The design specifications of the DHX and test conditions are
described in Table 7. A total of 28 experimental cases are selected
from references Skupinski et al. (1965) and Nanashima et al.
(1979), which are summarized in Table 8.
4.2. MARS-LMR results for the DHX
Using information described in Table 7, a finned tube bundle is
modeled for the MARS-LMR input. Similar to the PFBR case, the 50
nodes in the heat transfer region are used for all validation calcu-
lations. In addition, the multiplication factor for the area ratio
between the finned area and the in-tube area is applied to the
shell-side. Here, the finned area was 11.8-times larger than the
in-tube area. Considering a fin efficiency of 0.83, the multiplication
factor was calculated as 9.86 for the DHX. In the reference case,
which has a multiplication factor of 9.86, the MARS-LMR results
indicate that the air-side temperature shows a better prediction
than the sodium-side temperature, as shown in Fig. 11. Based on
Fig. 6, the test cases were recalculated with a correction factor of
0.88 for two tube rows. The total heat transfer rates for all exper-
imental cases are shown in Fig. 12. RMS deviation of the heat trans-
fer rates for the reference and the correction factor of 0.88 are
8.01% and 7.26%. There is little improvement for the correction fac-
tor. However, the sodium temperature difference still has the lar-
gest deviation. The sodium-side flow rate is measured before
bifurcating to the sodium inlet of the sodium tube-side in each
DHX. For the reference case, the individual sodium flow rate is
assumed to be 50% of the total measured sodium flow rate. To con-
sider the distributed flow rate, the sodium flow rates in tube-side
221
are evaluated with air-side heat transfer rate as shown in Fig. 13.
The air-side heat transfer rate can be evaluated with individually
measured parameters of the inlet and outlet temperatures and
air flow rate. In addition the heat loss to air-side is neglected. Based
on the air-side heat transfer rate, the sodium flow rate were newly
evaluated to conserve the heat transfer rate. Fig. 13 indicates that
the sodium flow rate is around 10–20% higher than the 50% rated
flow. Therefore, all experimental cases are re-calculated with the
corrected flow rate and a correction factor of 0.88. As shown in
Fig. 11, when the newly corrected sodium flow rate is used, RMS
deviation of the sodium temperature difference is improved from
20.71% to 7.45%, however, the RMS values for air-side and total
heat transfer rate are not affected. The deviations for cases for
the reference, correction factor, and new sodium flow rate are
summarized in Table 8.
5. Conclusions
The DHRS in the PGSFR has two kinds of sodium–air heat
exchangers: the one is helical type (AHX) and the other is finned
serpentine type (FHX). The DHRS is an important system for a heat
removal during an accident. For example, the DHRS is an only heat
removal system under a loss of heat sink accident, which has fail-
ure of heat removal through all steam generators. Therefore, its
accuracy of the modeling in MARS-LMR is critical to estimate the
safety evaluation. Recently, the heat transfer correlations for air-
side in a bundle heat exchanger are added in the MARS-LMR code.
In this study, the finned air-sodium heat exchanger is focused
upon. Two experimental datasets for the validation of newly devel-
oped heat transfer models are obtained from the open literature.
One is a performance test of the AHX in the PFBR, and the other
is a start-up test of the DHX in the JOYO.
The heat transfer models for a finned tube bundle in MARS-LMR
are slightly over-estimated compared with the experiments. The
deviation RMS values of the heat transfer rate for the PFBR and
the JOYO are 10.06% and 5.01%, respectively, which does not
222 C.W. Choi, K.S. Ha / Annals of Nuclear Energy 94 (2016) 213–222
consider the correction factor. When the number of tube rows is
less than 5, the heat transfer is deteriorated. Therefore, the correc-
tion factor should be applied to the original correlation. When the
appropriate correction factors are applied, the deviation in the RMS
values of the heat transfer rate for the PFBR and the JOYO are 5.18%
and 7.26%. Obviously, the prediction with the correction factor
shows better results. In addition, there is less information for a
sodium flow rate in the JOYO experiments. A reasonable calcula-
tion of the sodium flow rate based on the air side heat transfer rate
can give a better prediction of the sodium temperature difference.
In this validation study, the experimental uncertainties (e.g. tem-
perature and flow measurements, heat loss, and flow by pass) are
not considered because of no-information of uncertainties. The
experiments for the FHX in the PGSFR have been conducted using
a sodium test loop for a safety simulation and assessment (STELLA-
1 and 2) in KAERI. In the near future, the experimental data
obtained from more appropriate design conditions will be used
to validate the heat transfer models for the FHX in the MARS-LMR.
Acknowledgments
This work was supported by the Nuclear Research & Develop-
ment Program of the National Research Foundation (NRF) grant
funded by the Korean government MSIP (Ministry Science, ICT
and Future Planning).
References
Aoki, S., 1973. Current liquid-metal heat transfer research in Japan. Prog. Heat Mass
Transfer 7.
Doi, M., Endoh, M., Nanashima, T., Maetani, H., Hayami, M., Wada, H., Hirose, T.,
Terata, K., Yamamoto, H., 1980. JOYO 75MWt start-up test report: PT-12 heat
transfer characteristics of IHX and DHX, PNC-SN-941-80-65.
Eoh, J.H., et al., 2013. Design features of a large-scale sodium thermal-hydraulic test
facility: STELLA. In: International Conference on Fast Reactors and Related Fuel
Cycles (FR13), IAEA-CN-199/303, Paris, France, March 4–7, 2013.
Ha, K.S., Jeong, H.Y., 2010. Simulation of the EBR-II loss of flow tests using the MARS
code. Nucl. Technol. 169, 134–142.
Jeong, J.J., Ha, K.S., Chung, B.D., Lee, W.J., 1999. Development of multi-dimensional
thermal-hydraulic system code, MARS 1.3.1. Ann. Nucl. Energy 26.
Kim, Y., Lee, Y.B., Lee, C.B., Chang, J., Choi, C., 2013. Advanced design concept of
sodium-cooled fast reactor and related R&D in Korea. Sci. Technol. Nucl. Instal.
2013 (21).
Nanashima, T., Endoh, M., Enomoto, T., Hirose, T., Yamamoto, H., 1979. JOYO start-
up test report: PT-12 heat transfer characteristics of IHX and DHX, PNC-TN941-
79-128.
Skupinski, E., Tortel, J., Vautrey, L., 1965. Determination of convection coefficients of
sodium potassium alloys in circular tubes. J. Heat Mass Transfer 8–6, 937–940.
Vinod, V., Pathak, S.P., Paunikar, V.D., Suresh Kumar, V.A., Noushad, I.B., Rajan, K.K.,
2013. Experimental evaluation of sodium to air heat exchanger performance.
Ann. Nucl. Energy 58, 6–11.
Zukauskas, A., KARNI, J., 1989. High-performance Single-phase Heat Exchangers.
Hemisphere Publishing Corporation, Ch.13.