Annals of Nuclear Energy 168 (2022) 108876

Contents lists available at ScienceDirect

Annals of Nuclear Energy

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

Experimental and numerical study on thermal performance of a fluoride

salt-to-air heat exchanger
Yu-Shuang Chen a,b, Jian Tian a,⇑, Hai-Hua Zhu a, Yuan Fu a, Na-Xiu Wang a
a
Shanghai Institute of Applied Physics, Chinese Academy of Sciences, Shanghai 201800, China
b
University of Chinese Academy of Sciences, Beijing 100049, 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 better understand the thermal performance of Fluoride salt-to-Air Heat Exchanger (FAHX), the
Received 29 November 2020 heat transfer process of FAHX with multiple-tube banks staggered arrangement in a Fluoride-salt High
Received in revised form 10 March 2021 Temperature Test Loop (FHTL) is studied experimentally and numerically. Results show that the temper-
Accepted 27 November 2021
ature distribution obtained from CFD fit well with experiment measurements with a maximum deviation
Available online 23 December 2021
of 10%, which indicates that the CFD program is applicable. Then, the three-dimensional features of the
FAHX in the tube-side and shell-side are analyzed. Results show that peripheral tube temperatures of the
Keywords:
tube bundles are slightly lower than the center tubes, and all tubes temperature are obviously above the
Molten salt reactor
Fluoride salt-to-air heat exchanger
melting temperature of the fluoride-salt
454 °C, which implies that there will be less risk of fluoride-salt
Heat exchanger experiments freezing in the FAHX. Finally, the heat transfer process related flow distributions are discussed, helpful
Computational fluid dynamics guidance are proposed for the optimal design of the molten salt-to-air Heat exchanger.
Thermal performance Ó 2021 Elsevier Ltd. All rights reserved.

1. Introduction (LSTL) (Graydon et al., 2015; Richard, 2014) at ORNL, and the major
purpose of these FAHXs is served as the final heat radiator. Spe-
The Fluoride salt-to-Air Heat exchanger (FAHX) as an essential cially, heat transfer of kinds of gas cooled molten salt exchangers
heat transfer equipment plays an important role in molten salt have been extensively investigated. Experimentally, many studies
related high temperature reactors (Forsberg, 2005). In some exper- investigated air cooled molten salt heat exchanger, to obtain heat
imental reactors, such as Molten Salt Reactor Experiment (MSRE) transfer coefficients of molten salt in both the tube and shell sides,
(Robertson, 1965) at Oak Range National Laboratory (ORNL) and verify the self-made program for the design of molten salt
Thorium Molten Salt Reactor nuclear energy system (TMSR) exchangers (Chen and Wang, 2019), and testify the validity of
(Jiang et al., 2017; Cai et al., 2016) at Shanghai Institute of Applied the classical heat transfer correlation for molten salt (Chen et al.,
Physics, Chinese Academy of Sciences (SINAP, CAS), the FAHX is 2018). Numerically, the thermal–hydraulic characteristics of FAHX
used as final heat removal device. In some commercial reactors, have been also investigated using a system-level analysis program,
such as Fluoride-Salt-Cooled High-Temperature Reactor (in FHRs) such as RELAP5/SCDAPSIM/MOD 4.0 (Lin et al., 2018), and Compu-
(Carpenter et al., 2013; Nakata et al., 2013), the FAHX is employed tational Fluid Dynamics (CFD) program (Cunningham et al., 2014).
to removal a decay heat in a Direct Reactor Auxiliary Cooling Sys- However, there is no analysis about the overall thermal character-
tem (DRACS). Further, aiming at the power cycle for electricity gen- istics considering on effects of both-side fluid coupled heat transfer
eration, Fluoride salt to Gas heat exchangers are potentially and high temperature radiant on the heat transfer process with flu-
compatible with several highly efficient power conversion systems oride salt in FAHX, especially for three-dimensional full-scale flow
such air Brayton cycle and so on (MIT, Uw Madison, UC Berkeley heat transfer simulation which can evaluate the local heat transfer
Facilitators, , 2013). temperature and velocity. In addition, the design and optimization
Up to now, many FAHXs have been used in molten salt experi- of heat transfer equipment will inevitably face a conflict between
mental loops, such as MSRE, Tohoku-NIFS Thermo-fluid loop (TNT) the safety goal of massive heat rejection capability to avoid over-
(Tomoaki et al., 2007; Toda et al., 2002) and Liquid Salt Test Loop heating and overcooling due to the higher melting point for fluo-
ride salt. Therefore, it is important to investigate the
⇑ Corresponding author. performance of the FAHX equipment considering above features.
E-mail addresses: [email protected] (Y.-S. Chen), [email protected] Particularly, the verification of the analysis method of
(J. Tian), [email protected] (H.-H. Zhu), [email protected] (Y. Fu), wangnaxiu@
sinap.ac.cn (N.-X. Wang).

https://doi.org/10.1016/j.anucene.2021.108876
0306-4549/Ó 2021 Elsevier Ltd. All rights reserved.
Yu-Shuang Chen, J. Tian, Hai-Hua Zhu et al. Annals of Nuclear Energy 168 (2022) 108876

thermo-hydraulics is necessary to effectively predict the potential heat exchanger (200 kW capacity), and pipelines. The gas system
to over-cool and prevent the frozen accident. is mainly used to keep the experiment under safe and stable con-
In the present paper, the thermal performance testing of FAHX ditions. All piping and components are wrapped with trace heater
is conducted under forced air draft conditions with steady-state. to heat the loop above FLiNaK melting temperature (
454 °C) all
Then, the thermal and hydraulic performance of the FAHX is sim- the time. Pipelines are sloped to ensure 100% filling and draining
ulated using CFD program and the key parameters calculation val- of salt in the loop. Loop instrumentations include flow-meters,
ues are compared with experiment results in order to apply the level detection devices, pump discharge pressure, and temperature
computation method to the consecutive heat exchanger design. measurements of loop pipes. The photos of Installation status of
Finally, the velocity distribution and the temperature distribution FHTL are shown in Fig. 2.
in the FAHX is computed precisely using a CFD program, and the During the experiment, with the pump, which can deliver a
thermal performances of the heat exchanger are numerically stud- molten salt flow of 20 m3 h1, the molten salt is pumped from
ied and results are discussed. the storage tank through electric-heater with 200 kW capacity
and into the pump sump tank, then introduced into the tube side
2. Experimental studies of FAHX. Meanwhile, the cold air flows through the shell-side of
FAHX, in which the room temperature air is injected into the shell,
2.1. The experimental system the 200 kW of heat loaded by the heater is reject to keep the loop
thermal balance. The temperature of FLiNaK from 550 °C to 600 °C
The Fluoride salt High Temperature Test Loop (FHTL) used to is obtained by controlling the electric heater, and the flow rate
investigate the thermal performance for Fluoride salt-to-Air Heat from 15 m3 h1 to 20 m3 h1 are obtained by controlling the mol-
Exchanger is built at Shanghai Institute of Applied Physics, Chinese ten salt pump frequency. The air flow rate is supplied using a vari-
Academy of Sciences (SINAP, CAS). The schematic diagram of FHTL able speed blower. When the system is highly stable and reaches a
loop is shown in Fig. 1. The facility design is sufficiently versatile to steady-state condition which the fluctuation of the temperature is
allow a variety of testing of components such as air-cooled heat below ± 1 °C for 10 minutes, the flow rate and temperatures for
exchanger, high-temperature instrumentation, improved pump both sides will be collected. The inlet and outlet temperatures of
design, as well as thermal and material testing. the molten salt are measured using PT100 thermocouples with
The molten salt LiF-NaF-KF (46.5–11.5–42 mol %, FLiNaK) is an accuracy of 0.5 K and uncertainty of 1.0%. The inlet and outlet
used in this experiment to eliminate any safety issues. The FHTL temperature of air are measured by K type thermocouples with
consists of main test loop, gas system and related auxiliary sys- accuracy of 1.2 K. Molten salt flow through the heat exchanger is
tems. The main test loop mainly contains molten salt storage tank, measured by ultrasonic wave flow meter with a measuring range
electric heater, molten-salt pump, fluoride-salt valves, air-cooled of 10–20 m3 h1 and an estimated calibration accuracy of ± 1.7%.

Fig. 1. Schematic diagram of Fluoride salt High Temperature Test Loop (FHTL).

2
Yu-Shuang Chen, J. Tian, Hai-Hua Zhu et al. Annals of Nuclear Energy 168 (2022) 108876

Fig. 2. Installation status of Fluoride salt High Temperature Test Loop (FHTL).

Air flow through the heat exchanger is measured by a Coriolis flow efficiently reduce the flow-induced vibration (FIV), two pore plate
meter with an uncertainty of 0.15%. are constructed to support the bundles, as shown in Fig. 3.
Table 1 describes key design parameters of the FAHX. The nom-
inal volume flow rate of FLiNaK is
20 m3 h1 and the nominal
2.2. Fluoride salt-to-Air heat exchanger inlet and outlet temperature of FLiNaK are 570 °C and 550 °C
respectively. The nominal volume flow rate of air across the tubes
In this study, the Fluoride salt-to-Air Heat Exchanger is is 5000 m3 h1 with inlet environment temperature at atmospheric
designed as an eight-pass serpentine shape (S-shape) inclined cir- pressure. Details of the design parameters are shown in Table 1.
cular tube banks in a staggered arrangement with a nominal heat
transfer capacity of 200 kW. Fig. 3 shows the schematic drawing
and fabricated structure of the FAHX, the tube-and-shell type 2.3. Experimental measurements
FAHX is vertically oriented and consists of 2 rows total 11 tubes
(the upper row consists 5 tubes and the lower row consists 6 The data are controlled using an auto-control system and
tubes). The length, outer diameter and thickness of tube are 8 m, acquired every 1 sec in the steady state condition when the exper-
25 mm and 2.5 mm respectively and the material is Hastelloy iment criteria is satisfied. The temperature and volume flow rate
C276. The serpentine shape tubes are attached to tube sheets at under steady-state are obtained for 10 minutes as shown in
either end header tanks. The overall size of the model FAHX Fig. 4. It can be seen the flow rate of salt in tube-side is
(width  length  height) is 0.35 m  1.1 m  1.5 m. In order to 19.8 m3 h1 with average inlet temperature of 542.5 °C, and the

Fig. 3. Schematic of designed serpentine coil fluoride salt-to-air heat exchanger and its fabricated structure photo.

3
Yu-Shuang Chen, J. Tian, Hai-Hua Zhu et al. Annals of Nuclear Energy 168 (2022) 108876

Table 1
Design parameters of fluoride salt-to-air heat exchanger.

Design parameters Values Unit

Power rating 200 kW
Tube OD/ID 25/2.5 mm
Tube material Hastelloy C276 –
Number of tubes 11 –
Tube inclined angle 3 degree
Length of tube 8 m
Molten salt inlet/outlet temperature 570/550 °C
Fluoride salt flow rate 20 m3 h1
Air inlet temperature environment temperature °C
Air flow rate 5000 m3 h1

Fig. 5. The numbering instruction of tube and measuring point location of FAHX.

Table 3
Temperature distribution of tube no.3.

Location (distance from fluoride salt inlet) Temperature (°C)

1 (
0.5 m) 516.5
2 (
1.0 m) 527.3
3 (
1.5 m) 518.2
4 (
2.5 m) 512.7
5 (
3.0 m) 522.0
6 (
3.5 m) 513.0
7 (
4.5 m) 512.5
8 (
5.0 m) 513.1
9 (
5.5 m) 500.1
10 (
6.5 m) 496.4
11 (
7.0 m) 503.3
12 (
7.5 m) 480.8

Fig. 4. Temperature and flow rate results for 10 minutes.

located near the inlet of cold air. For this, 6 thermocouples are
3 1 installed on the outer wall of the tube bundle near the air entrance
volume flow rate of shell-side air across the tubes is 574 m h
area, as shown in the Fig. 6, and the measured temperatures of the
with average inlet temperature of 44.8 °C. In this period all mea-
tubes close to wall (No. 1 and No.3) are obviously lower than other
sured values maintained without any continuous increase or
positions, as listed in Table 4.
decrease as time passes. The experimental conditions for the FAHX
are listed in the Table 2, including the inlet and outlet temperature,
volume flow rate and heat transfer rate of two fluids (FLiNaK and
air) for the shell and tube sides. 3. Numerical calculation and validation
The measured temperature of tube located in the central upper
row (tube No.3) from upper chamber to lower chamber is shown in 3.1. Computational fluid Dynamics modeling
the Fig. 5. 12 thermocouples are installed in the middle of straight
section and bending section respectively, as shown in the Fig. 5 and According to the FAHX in the FHTL system, a Computational
the measure results of the temperature are listed in Table 3. It is Fluid Dynamics (CFD) model is established to simulate to the heat
clearly that the measured temperatures are gradually decreasing
from the inlet to the outlet of tube side, while the bending section
points (point 2, 5, 8 and 11) are slightly higher than the nearby
straight section points.
Due to the relatively high melting point of the fluoride salts
(
454 °C), while fluoride-salt heat transfer tubes in the FAHX
expose to ambient air with temperature obviously below the melt-
ing point of molten salt, the heat transfer pipe of FAHX will
undergo great potential of overcooling, and even damage the heat
transfer equipment, especially for tubes of bottom row which

Table 2
Overall experimental results.

Parameters Tube-side (FLiNaK) Shell-side (Air)

Inlet Temperature (°C) 542.5 44.8
Outlet Temperature (°C) 540.4 412.0
Volume flow rate (m3h1) 19.8 574
Fig. 6. The numbering instruction of measuring point location at bottom tube
Heat Transfer rate (kW) 45.5 43.4
bundle.

4
Yu-Shuang Chen, J. Tian, Hai-Hua Zhu et al. Annals of Nuclear Energy 168 (2022) 108876

Table 4 Table 5
Temperature distribution of the bottom tube Properties used in the CFD modeling calculation.
bundle.
parameters FLiNaK molten salt (Chinese Academy of HN-C276
Location number Temperature (°C) Sciences) alloy
1 470.0 Density [g q ¼ 2:6133  6:431  104 T ðK Þ 8.68
2 480.6 cm3] (500 °C–
3 464.4 550 °C)
4 481.7 Specific heat 1880 408
5 480.8 capacity [J (375 °C–
6 478.1 kg1 k1] 575 °C)
Thermal k ¼ 0:44532 þ 0:0014TðKÞ 17.1–18.57
conductivity (400–
W m1 k1 500 °C)
transfer process from fluoride salt to the atmosphere. To this end,  
Dynamic g ¼ 8:06598 exp  5058:43 þ 3:92342  10
6 0.48
T T2
the CFD modeling includes four regions: fluoride-salt fluid, tube- viscosity [cP]
side solid, air fluid and shell-side solid. The FAHX heat exchanger
is meshed using Workbench’s meshing. The detailed CFD meshing
are shown in Fig. 7 for the fluoride side and the air side. Specifi- surface is well insulated, the heat transfer coefficient from outer
cally, 5 boundary layer nodes are added to the mesh near the tube shell to the environment is assumed to be zero.
wall surface to ensure the first boundary layer nodes are located in To ensure the simulation quality, a detailed grid sensitivity
the viscous sublayer, thus the near-wall flow characteristics can be study is carried out for FAHX. The heat transfer rate as a function
captured. The total meshes are generated mainly using tetrahedral of mesh levels is listed in Table 6. As the quantity of mesh increases
lattices. from 25.10 million to 30.52 million, the heat transfer rate of FAHX
The physical properties of fluoride-salt and HN-C276 alloy used rises from about 42.8 kW to 44.5 kW, while the quantity of mesh
in the current analysis are listed in Table 5, in which the physical continues to increase to 60.01 million, the heat transfer rate is
properties of molten salt were actually measured by the SINAP, almost invariable. Taking the simulation time into consideration,
CAS. In the present study, the whole heat transfer analysis of FAHX the modelling with 30.52 million mesh is adopted in current study.
is simulated by FLUENT solver (ANSYS Inc). A standard k-e turbu-
lence model and DO radiation model are selected for the CFD cal-
culation. The second-order upwind scheme is used to treat the 3.2. Calculation and validation
convection terms, while for viscous terms, second order central dif-
ference is assigned. An inlet volume flow rates of 19.8 m3 h1 at In order to validate the calculation results of CFD simulation,
542.5 °C in the fluoride-side and 574 m3 h1 at 44.8 °C in the the outlet temperature and heat transfer rate of shell-sides and
air-side, and pressure outlet are given to the program as the tube-sides are calculated and compared with measurement results
boundary conditions. In the present analysis, since the air outer from experiment. According to the calculation, the outlet temper-
ature is 442.0 °C in the shell-side and 540.5 °C in the tube-side

Fig. 7. Mesh generation for the CFD analysis.

5
Yu-Shuang Chen, J. Tian, Hai-Hua Zhu et al. Annals of Nuclear Energy 168 (2022) 108876

Table 6
Results of the grid sensitivity study.

Quantity of mesh (million) Heat transfer rate (kW)

25.10 42.8
28.60 43.5
30.52 44.5
60.01 45.1

Table 7
Comparison of experiment and simulation results.

Parameters CFD EXP Deviation (%)

Air Outlet Temperature (°C) 442.0 412.0 7.03
FLiNaK Outlet Temperature (°C) 540.5 540.4 0.02
Shell-side heat Transfer rate (kW) 44.5 43.4 2.50
Tube-side heat Transfer rate (kW) 44.7 45.5 1.77

respectively, and the overall heat transfer rate is 44.5 kW in the

shell-side and 44.7 kW in tube-side respectively. The comparison
of experiment and simulation results is presented in Table 7. It
can be seen that, although the air outlet temperature from CFD is
slightly higher than the experimental value due to mainly the
measurement uncertainty and heat loss, the results from
3-dimensional CFD calculation shows agreement with the experi-
Fig. 9. Temperature contour maps of vertical sections in FAHX.
mental measurements with an acceptable deviation range of less
than 10%.
To further validate the calculation results from CFD simulation,
in Fig. 6) in the FAHX from CFD calculation and experimental mea-
the temperature distributions at different locations in FAHX from
surements. It can be seen that although the calculations are gener-
experiment measurement and CFD simulation is shown in Fig. 8.
ally higher than the measurements, the tendency of temperature at
The wall temperature distribution outside the specific tube (posi-
different location is totally the same with a maximum deviation of
tions shown in Fig. 5) is shown in Fig. 8(a). It is clear that the ten-
6.4%. Among the tube bundle of the same axis level, temperatures
dency of temperature distribution along the tube length from
of peripheral tubes (tube No.1 and No.3) are lower than other
simulation is generally in line with the results from the experiment
tubes. This is because when the air flows through the gap between
with a maximum deviation of 4.43% and an average deviation of
the straight section of tubes and shell wall, it undergoes less air
2.03%. Specifically, the temperature are gradually decreasing from
flow resistance and brings a larger flow velocity, made these tubes
the inlet to the outlet of tube side, while the bending section points
suffering more cooling.
(such as point 2, 5, 8 and 11) is slightly higher than nearby straight
Based on the above studies, although there are some measure-
section points. The main reason is that these bending sections are
ment uncertainties in the experiments, the key parameters and
very close to air-side wall, and less air flow through this area, as a
temperature distribution of FAHX from CFD simulation basically
result, more thermal radiation and less heat transfer will be hap-
fit with the experiments, and thus the numerical calculation is
pened. Fig. 8(b) shows the wall temperature distribution of the
reliable.
bottom tube bundle close to the air entrance area (positions shown

Fig. 8. Temperature distribution of tube No.3 (a) and bottom tube bundle (b) from experiment measurement and CFD simulation.

6
Yu-Shuang Chen, J. Tian, Hai-Hua Zhu et al. Annals of Nuclear Energy 168 (2022) 108876

to salt inlet plane, the temperature of air increases accordingly

from 90.13 °C to 441.27 °C. Besides, the temperature increase near
the air entrance area is more obvious, while for the area closer to
the exit of air, the temperature rise rate is decreasing gradually.
This indicates that the heat transfer rate at the air entrance is bet-
ter than that at the air exit area.
To further investigate the temperature distribution of shell side,
the temperature contour maps of different horizontal planes from
salt inlet to salt outlet (axis height of 0 m, 0.14 m, 0.30 m, 0.46 m,
0.62 m, 0.78 m, 0.95 m, 1.11 m, 1.27 m from salt outlet plane) in
FAHX are shown in Fig. 11. It is clear that, the overall temperature
distribution of air in the shell-side shows a big difference from
inlet horizontal plane to outlet horizontal plane of salt. For inlet
horizontal plane, the hottest air temperature prevails in the region
near the shell-side walls and the maximum value is up to
304.24 °C, while in the central region, cold air about 45.0 °C is dom-
inated. For higher horizontal planes, the situation is opposite, and
the air temperature of center region is comparatively higher than
those region near shell-side walls at the same plane. The main rea-
son for the discrepancy is that the center region near the air-inlet
Fig. 10. The average temperature distribution of air in shell-side at different has more sufficient cold air flow than other higher planes, as the
horizontal planes. height increases, the heat transfer becomes prevailing in the center
area and the air flowing through sidewall region increases due to
the effect of flow resistance.
4. Results and discussions In addition, the solid temperature distribution of FAHX such as
pipe, support plate and air-shell are also shown in Fig. 12 It can be
4.1. Thermal performance of FAHX seen that the temperature distribution of the structure shows an
overall temperature increase trend from the air inlet to the air out-
The overall temperature contour maps of two orthogonal verti- let, which are mainly affected by the air-side temperature
cal section for the FAHX from CFD calculation is shown in Fig. 9. It distribution.
can be seen that the temperature distribution in the tube-side is
almost uniform with slight change from the inlet to the outlet,
while for the shell-side, it shows an inhomogeneous distribution 4.1.2. Temperature distribution of tube side
for both horizontal and vertical sections under the present air Fig. 13 shows the axial temperature distributions along tube
forced convection scenario. length for tubes from inlet chamber to outlet chamber. Since the
symmetry of the structure, half of tubes (No.1, No.2, No.3, No.6,
No.7, and No.8) are presented. It is clear that the overall tempera-
4.1.1. Temperature distribution of shell side ture of salt decreases along tube length from inlet to outlet, how-
The average temperature distribution of air in shell side of the ever, the decreasing rate is different for tubes from center to
FAHX at different horizontal planes from salt outlet to salt inlet periphery. For center fluoride salt tubes (No.1, No.2, No.3, No.7,
(axis height of 0 m, 0.14 m, 0.30 m, 0.46 m, 0.62 m, 0.78 m, No.8), there is no obvious change in the temperature of the location
0.95 m, 1.11 m, 1.27 m from salt outlet plane) is shown in from 0.5 m to 3.0 m, while the temperature have significant reduc-
Fig. 10. It can be seen clearly that the air temperature increases tion as the location changes from 3.0 m to 7.5 m. For the most
as the axial height increases, specifically, from salt outlet plane peripheral tube No. 6, the temperature change tendency is totally

Fig. 11. Contour maps of temperature of cross sections in FAHX.

7
Yu-Shuang Chen, J. Tian, Hai-Hua Zhu et al. Annals of Nuclear Energy 168 (2022) 108876

Fig. 12. Contour maps of the solid temperature distribution of FAHX.

different from the center tubes, specifically, the temperature as well as the larger airflow around, the temperature of tubes
decrease rate drops linearly in the whole flow path, and thereby located in peripheral region have larger decreasing rate along the
the tube temperature is significantly lower than other tubes at flow path of salt as shown in Fig. 11. Therefore, in order to reduce
the same axial height. bypass airflow near two side walls, enhance the effective heat
It can be also seen that the bottom tube bundle (distance from transfer rate over the tube bundle and lower the risk of overcooling
inlet
7.5 m) near the air entrance reaches the lowest values of the most peripheral tubes, additional structures are expected to
about
535 °C, which is still higher than the melt point of FLiNaK. improve the flow distribution of fluid and thermal performance of
Therefore, although the FAHX is working in ambient air which is the FAHX. Fig. 15 proposes one of optimized structures by adding
obviously below the melting point of molten salt, the temperature multiple circumferential horizontal baffles attached to the side
of fluoride-salt at any individual tubes is within the permissible walls, and further optimized analysis will be carried out and ther-
range at current operating conditions and there will be less risk mal performance of FAHX using the baffle structure have also been
of fluoride-salt freezing in the FAHX tubes. underway.

4.2. Velocity distribution and structure optimization of FAHX 5. Conclusion

In order to investigate the impact of structure on the thermal In this paper, a steady-state thermal performance test of fluo-
performance, and further provide useful guidance for the optimiza- ride salt to air heat exchanger (FAHX) with multiple-tube banks
tion of structure, velocity distribution as well as related heat trans-
fer of shell side and tube side is deeply investigated. The partial
enlarged details for contour maps of velocity in vertical section
perpendicular to the tube bending plane are shown in Fig. 14(a).
For the tubes side, central tubes are closer to the inlet nozzle,
therefore, more fluoride salt will flow pass the central tubes com-
pared to the peripheral tubes. For the shell side, as air runs across
the tube bundles, airflow in the passages between two adjacent
tubes is strongly accelerated while behind each tube it becomes
slow, and then the region away from tubes show the velocity max-
imum, as a result, there will be significant heat transfer rate over
these tubes, as expected. In addition, because there is no baffles
in the regions near two side walls, the velocity of air nearby is lar-
ger than those around other area due to the smaller flow
resistance.
The temperature shows a corresponding distribution according
to the velocity distribution, in which, the temperature of air is high
in the wake region and is low in the accelerating region, and large
temperature gradient is formed around the tube bundle as shown
in Fig. 14(b). Meanwhile, due to the smaller flow rate of molten salt Fig. 13. Axial temperature distribution of different tubes in the FAHX.

8
Yu-Shuang Chen, J. Tian, Hai-Hua Zhu et al. Annals of Nuclear Energy 168 (2022) 108876

Fig. 14. Contour maps of velocity and temperature in FAHX.

staggered arrangement was carried out and a three-dimensional tube-side were obtained for the FAHX. Results show that the cur-
calculation using the CFD modeling was performed for heat trans- rent designed FAHX working in ambient air conditions has less risk
fer performance demonstration. Results show that the temperature of fluoride-salt freezing in the heat transfer tubes. In addition,
distribution and heat transfer rate obtained from experimental according to the thermal performance of the FAHX, the heat trans-
data are in good agreement with the CFD calculation within a fer process related flow distribution for both shell-side and tube-
maximum deviation of 10%, which means the current CFD model side are discussed. Base on the studies, optimization of structures
is reliable for the thermal analysis of FAHX. After that, the by adding multiple baffles was proposed to improve the thermal
three-dimensional temperature distributions of the shell-side and performance of the FAHX. Results from current study will provide

Fig. 15. Some optimal designs by adding a baffle in the air bypass section.

9
Yu-Shuang Chen, J. Tian, Hai-Hua Zhu et al. Annals of Nuclear Energy 168 (2022) 108876

important references for future design and construction of molten Test Report of Thermo-physical Properties of Molten Salt in 2MWt Liquid Fuel
Thorium Based Molten Salt Experimental Reactor, internal report, Shanghai
salt-to-air heat exchanger for the TMSR.
Institute of Applied Physics, Chinese Academy of Sciences.
Cunningham, R.B., Ruggles, A.E., Yoder, G.L., 2014. Multidimensional CFD Modeling
Declaration of Competing Interest of a Liquid Salt Pebble Bed Heat Transfer Loop. University of Tennessee, pp. 1–
66. Master’s Thesis.
Forsberg, C.W., 2005. The advanced high-temperature reactor: high-temperature
The authors declare that they have no known competing finan- fuel, liquid salt coolant, and liquid-metal-reactor plant. Prog. Nucl. Energy. 47
cial interests or personal relationships that could have appeared (1–4), 32–43.
to influence the work reported in this paper. Graydon, L. Y. J., Adam, A., Burns, C., David, F., David, H., Roger, K., Fred, P., Kevin, R.,
Dane, W., 2015. High-temperature Fluoride Salt Test Loop. ORNL/TM-2012/430.
Jiang, M.H., Xu, H.J., Dai, Z.M., 2017. Advanced fission energy program-TMSR nuclear
Acknowledgements energy system. Bull. Chin. Acad. Sci. 27, 366–374.
Lin, Hsun-Chia, Lv, Qiuping, Shi, Shanbin, Sun, Xiaodong, Christensen, Richard,
Yoder, Graydon, 2018. Code validation of a scaled-down DRACS model in
This work is performed under the auspices of the ‘‘Strategic Pri- RELAP5/SCDAPSIM/MOD 4.0. Ann. Nucl. Energy 121, 452–460.
ority Research Program” of the Chinese Academy of sciences (Grant MIT, UW Madison, UC Berkeley Facilitators, 2013. Fluoride-salt-cooled, High
No. XDA0201002) and the ‘‘National Natural Science Foundation of temperature Reactor (FHR) Development Roadmap and Test Reactor
Performance Requirements White Paper. Massachusetts Institute of
China” (Grant No. 91326201), and ‘‘Youth Innovation Promotion Technology. UCBTH-12-004.
Association, Chinese Academy of Science” (Grant No. 2020263). Nakata, J., Wakui, M., Mori, M., Sakashita, H., Forsberg, C., 2013. Performance
evaluation of DRACS system for FHTR and time assessment of operation
procedure. 2013 21st International Conference on Nuclear Engineering.
References
Richard, B.C., 2014. Multidimensional CFD modeling of a liquid salt pebble bed heat
transfer loop. A Thesis Presented for the Master of Science Degree The.
ANSYSÒ Academic Research Fluent, Release 14.5, Help System, Fluent User’s Guide, University of Tennessee, Knoxville.
ANSYS, Inc. Robertson, R.C., 1965. MSRE Design and Operations Report Part I Description of
Cai, X.Z., Dai, Z.M., Xu, H.J., 2016. Thorium molten salt reactor nuclear energy Reactor Design, ORNL-TM-728. Oak Ridge National Laboratory, Oak Ridge,
system. Physics 45 (9), 578–590. Tennessee.
Carpenter, D., Cao, G.P., Andreades, C., 2013. ‘‘Fluoride-salt-cooled, High- Toda, Saburo, Chiba, Shinya, Yuki, Kazuhisa, Omae, Masahiro, Sagara, Akio, 2002.
temperature Reactor (FHR) Development Roadmap and Test Reactor Experimental research on molten salt thermofluid technology using a high-
Performance Requirements White Paper,” UCBTH-12-004, Massachusetts temperature molten salt loop applied for a fusion reactor Flibe blanket. Fusion.
Institute of Technology. Eng. Des. 63-64, 405–409.
Chen, Y.S., Wang, N.X., et al., 2019. Thermal sizing design experimental evaluation Tomoaki, S., Kazuhisa, Y., Shin-ya, Chiba, et al., 2007. Heat transfer performance for
of molten salt-to-air heat exchanger. Ann. Nucl. Energy. 132, 504–511. high prandtl and high temperature molten salt flow in sphere packed pipes.
Chen, Y.S., Tian, J., Fu, Y., Tang, Z.F., Zhu, H.H., Wang, N.X., 2018. Experimental study Fusion. Sci. Technol. 52, 618–624.
of heat transfer Enhancement for molten salt with transversely grooved tube
heat exchanger in laminar-transition-turbulent regimes. Appl. Therm. Eng. 132,
95–101.

10