Nuclear Engineering and Design 248 (2012) 216–225

Contents lists available at SciVerse ScienceDirect

Nuclear Engineering and Design

Pump-induced pulsating pressure distributions in a system-integrated modular

a
Yonsei University, 50 Yonsei-ro, Seodaemun-gu, Seoul 120-749, Republic of Korea
b
Korea Atomic Energy Research Institute, Daeduk-daero 989-111, Deockjin-dong, Yuseong-gu, Daejoen 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: This study investigated pump pulsating pressure distributions in a scaled-down SMART research reactor
Received 4 July 2011 developed by KAERI. The local pressures were measured in the research reactor to calculate the pump-
Received in revised form 15 March 2012 induced loads. The calculated pressures agreed with the measured pressures at all 13 points in the reactor
Accepted 20 March 2012
model within an error of 3%. The pressure variation was controlled by changing the number of pulsating
pumps. When the number of pulsating pumps was one, two, or three, the relative pressure variation
was 25%, 50%, and 75% of what was measured with four pulsating pumps. The pressure variation due
to out-of-phase pulses could be deduced by superposition of the imposed pulses. The results indicated
that the numerical method used in this study is suitable for calculating pump-induced loads in a nuclear
reactor design.
© 2012 Elsevier B.V. All rights reserved.

1. Introduction inlet (SGI) region, eight stream generators (SGs), four reactor
coolant pumps (RCPs), core, flow mixing header assembly (FMHA),
Nuclear power will be valuable in the future due to the limited and flow skirt, are located in the reactor vessel, as shown in
quantity of fossil fuels such as oil, gas, and coal (Heising-Goodman, Fig. 1.
1981; Hammond, 1996). Scientists have been experimenting with SMART has no large piping among its main components.
nuclear fusion energy, but it is not yet useful for power generation. Therefore, pressure distributions and pressure drops induced by
Presently, nuclear power provides approximately 6% and 13% of pumping flow as well as pump-induced pulsating pressure varia-
the world’s energy and electricity, respectively. Because nuclear tions are very important. These pressures are related to forces on
power use will increase in the future, it is vital to minimize its the reactor internals; to ensure the safety of the SMART system,
dangerousness. pump pressure loads must be predicted in an actual hydrodynamic
Small nuclear reactors are powerful enough to generate electric- design.
ity and desalinate seawater; therefore, they have been developed An emergency in nuclear power plants may occur following an
in many countries. In 1996, the Korea Atomic Energy Research earthquake, tsunami, or failure of cooling systems. Many studies
Institute (KAERI) designed a new advanced integral reactor named have focused on incorporating ways to predict accident risks dur-
SMART, which stands for System-integrated Modular Advanced ing the development of a nuclear power plant (Strupczewski, 2003).
ReacTor. This consists of a small energy-conversion system and In the SMART system, thermo-hydrodynamics also plays a key role
an advanced integral pressurized-water reactor (PWR) (Chang and in safety. Many researchers (Yang et al., 2005; Chung et al., 2008;
Kim, 1997). SMART differs from a commercial loop pressurized- Lee et al., 2009) have evaluated the SMART design through safety
water reactor in that it is compact and has no large pipe systems analyses. They have investigated its safety and its performance
penetrating the reactor vessel. Its reactor coolant and major under conditions of a control-rod withdrawal event, pressure
primary circuit components are contained in the reactor pres- changes, and hot shutdown. Chang et al. (2000) studied how a
sure vessel. All main components, including the steam generator preliminary SMART system performed during an over-pressurizing
accident, and found that SMART has inherent over-pressure protec-
tion capability. Chung et al. (2006) studied the thermal-hydraulic
characteristics of the natural circulation loop in the SMART heat
∗ Corresponding author. Tel.: +82 2 2123 2828; fax: +82 2 312 2159.
removal system and found that most of the heat was removed at
E-mail addresses: [email protected] (K.M. Kim), [email protected]
(B.I. Lee), [email protected] (D. Lee), [email protected] (H.H. Cho), the steam generators by heat transfer. Recently, many researchers
[email protected] (J.S. Park), [email protected] (K.H. Jeong). have investigated pressure distributions in entire nuclear reactor

0029-5493/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.nucengdes.2012.03.048
 K.M. Kim et al. / Nuclear Engineering and Design 248 (2012) 216–225 217

Fig. 1. Layout of the SMART reactor vessel (Lee et al., 2009) and the scaled-down model.

systems (Durmayaz and Yavuz, 2001; Becker and Laurien, 2003) 2. Research methods
using commercial codes. Lee et al. (2010) conducted experiments
to measure time-averaged pressure distributions using a 1/10- 2.1. Scaled-down SMART model
scale SMART model. Kim et al. (2011) compared the experimental
pressure data with calculated steady-state pressure data using a Flow-induced loads must be estimated in nuclear reactor inter-
numerical code and the same model. Kim et al. (2010, under review) nals to obtain a standard design approval. Because it is very difficult
predicted local pressure distributions as well as temperature dis- to measure flow-induced loads in actual reactor internals, they are
tributions, and evaluated flow-induced loads in a full-sized SMART often predicted using a commercial computational fluid dynamics
system. (CFD) codes. However, these results may be affected by turbulence
In general, Weaver et al. (2000) reviewed flow-induced forces models, wall functions, and other numerical schemes, so the calcu-
generated in a nuclear reactor, including the steam generator tube lations must be verified experimentally.
bundles, fuel rod bundle, shell cavity, internal structure, control This study developed a scaled-down and simplified model to
rod guide tube, and leakage flow path. Au-Yang et al. (1995) con- predict flow-induced loads in SMART. This scaled-down design
ducted experiments to develop an empirical correlation to estimate simplified some components of the reactor due to its very complex
turbulence-induced forces and to check the effects of viscosity on actual geometry. First, it ignored the leakage flow rate between
turbulence using a 1/9-scale PWR model. Smith and Derksen (1998) components. Second, it simplified the reactor core, steam gener-
developed a test rig to measure the hydrodynamic loads at the ators, and control rod guides using hydraulic diameters or orifice
rod bundle of a pressurized heavy-water nuclear reactor. They pro- plates, because the main focus was on quantifying pressure drops.
posed that the bundle excitation force induced by the flow could Finally, it ignored the pressurizer region because there was no fluid
be reduced by flow straighteners. In nuclear power plants, failures flow.
of the reactor internals are generally the result of vibration caused Fig. 1 shows a conceptual sketch of SMART and the modified
by pump-induced acoustic pressure. Thus, it is necessary to predict SMART test model. Fig. 1a presents a layout of the SMART reac-
the pump-induced pressure to avoid problems related to vibration tor vessel and coolant flow direction by pump (Lee et al., 2009).
in the structure of the reactor internals. It is not easy to measure Fig. 1b depicts the flow directions in the scaled-down model. The
pressure loads due to the complex geometry of nuclear reactor two geometries differ in terms of pump regions: the pumps in the
internals, so many researchers (Cheong, 2000; Araseki et al., 2000; test model are placed outside and connected to the four inlets by
Spence and Amaral-Teixeira, 2008) have investigated pressure pul- hoses.
sations using numerical methods supported by experimental tests. The test model was made from acrylic. The height of the vessel
However, these researchers examined only pulsations caused by was 1.5 m, and the diameter of the vessel was 0.6 m. It included a
pumps. core support barrel, a flow mixing header assembly, a UGS support
The pump-induced pressure affects directly the internals barrel, a flow skirt, eight steam generators, and a reactor core inside
because the SMART reactor is that the pumps are located inside the the vessel. Tap water was used as the working fluid, and entered the
system, though the pump-induced pressure drops in the pipe in model through four annulus-shaped inlet ducts, flowed through the
case of a general reactor. We must predict how the pump-induced inner components, and finally left the vessel through the outlets.
pressure works. Therefore, the primary objective of the present All components were sealed using O-ring and flange rubber and
study was to develop a numerical method that can be applied to acrylic bonds to prevent leakage flow between components.
the design of the actual SMART reactor, specifically with regard As marked by the arrows in Fig. 1, fluid was supplied to the test
to pump pulsation. We compared calculated pressure results with reactor from four inlets located in the side of the upper shell and
measured pressure results, and investigated the hydrodynamic connected to the four pumps with hoses. In this experimental rig,
characteristics in a scaled-down research reactor with pump pul- an additional pressure pulse generator was located between one
sation. inlet and a hose, because the invert pumps did not generate large
218 K.M. Kim et al. / Nuclear Engineering and Design 248 (2012) 216–225

pulses. After the flow passed the pulse generator, it exhibited low adjusted from both the butterfly type damper size and the rotor
frequency and high amplitude pulses. The additional pressure pulse speed.
generator consists of a circular passage, a butterfly type damper, The flow was mixed in the inner volume of the upper shell
and stepping motor. The butterfly type damper and the stepping and moved downward through eight steam generators that were
motor control the frequency and amplitude. The butterfly type located at the circumference of the middle part of the shell and
damper has two wings. The pressure amplitude and frequency are divided into four pairs. After that, it moved into the flow-mixing

Fig. 2. Configuration of the test loop.

Fig. 3. Main internals in the scaled-down SMART model.

Fig. 4. Mesh shapes and measurement points.

header assembly, which consisted of a fourth level connected to for the height difference between the measurement point and the
each pair of steam generators and dispersed the fluid to maintain a connection point of the pressure transducer. When the pressure
uniform temperature in the inlet of the reactor core. Next, the fluid transducer was connected to the measuring point in such a manner,
moved to the flow skirt and the reactor core, located in the inner trapped air was removed from the connecting line.
shell side of the flow-mixing header assembly. Finally, it passed Signal data from pressure transducers were digitized simultane-
the upper guide structure assembly and passed through the out- ously using an NI USB-6259, which is a data acquisition system that
let of the test model, which was connected to the water bath with can simultaneously convert 1.25 million samples per second from
hoses. 16-bit analog inputs. Data were recorded on a PC using a data acqui-
sition program. During the experiments, signals from the pressure
transducers were acquired at a sample rate of 2000 samples per
2.2. Experimental methods second because the frequency range of interest was below 1000 Hz.
The water temperature was measured using a handheld thermome-
Figs. 2 and 3 present schematic diagrams of the test equipment ter (OMEGA, HH85) connected to a J thermocouple; the uncertainty
configuration. The open-loop test model consisted of a water bath, was 0.1 ◦ C.
four invert pumps, four flow meters, and the test reactor. These During the experiments, the temperature of the test room was
were connected consecutively with several hoses. The water bath maintained at room temperature and the fluid temperature was
was connected to the pumps that supplied the fluid passing through
the test reactor. The pumps maintained the water temperature
at 27 ◦ C (300 K) during the experiments, because pump operation
increased the water temperature. The pumps were placed on racks
to fill the test reactor with water. The upper side was open to the
atmosphere, so the static pressure at the top of the water bath was
atmospheric. The pumps were controlled to maintain a specified
flow rate by the flow controllers, and were placed outside the test
reactor because there was no space inside the vessel. Three turbine
flowmeters and one electromagnetic flowmeter were installed to
measure the flow rate of the working fluid: the flowmeter uncer-
tainty rate was about 0.5% of the full measurement range.
Piezo-resistive pressure transducers (Sensys, PSHK0001BCPG)
were used to measure the mean pressure and the unsteady fluctu-
ating pressure. A total of 26 measurement points were machined
on the outer surface of the vessel, and three measurement points
were machined on the inner structures. The three measurement
points located on the inner structure were connected to the out-
side of the vessel by extension tubes because a pressure transducer
could disturb the flow. The value of the measured pressure will not
differ if there is no gas in the tube and the system compensates Fig. 5. Experimental and numerical averaged static pressure and pressure drop.
220 K.M. Kim et al. / Nuclear Engineering and Design 248 (2012) 216–225

Fig. 6. Experimental fast Fourier transform (FFT) results for points 1, 4, and 9. (a) Pump pulsation peaks; (b) pressure reduction at each location.

Fig. 7. Calculation of pressure distributions by one pulsating and three constant pressure pumps. (a) Resultant pressure variations; (b) decrement of pump pulsation pressure.
 K.M. Kim et al. / Nuclear Engineering and Design 248 (2012) 216–225 221

Fig. 8. Pressure ratios of the maximum variation to the inlet variation at each location.

Fig. 9. Pressure changes by a pulse with ±10% variation and the same phase at each point.

maintained at 300 K. The system had a reactor flow of 3.3 kg/s; two-equation turbulence model that uses two transport equations
it discharged to the steam generators, entered the flow-mixing to represent the flow turbulence properties. Two-equation mod-
header assembly and flow skirt, moved upward through the core, els account for the convection and diffusion of turbulent energy.
and entered the pumps. The first variable is the turbulent kinetic energy, k and the second
variable is the turbulent dissipation, ε. The turbulent dissipa-
tion variable determines the scale of the turbulence, whereas the
2.3. Numerical methods kinetic energy variable determines the energy in the turbulence.
The k–ε model is known to be suitable for free-shear layer flows
The flow domain was the same as that of the experimen- with relatively small pressure gradients. This model also gives
tal setup. Fig. 4 shows the numerical grids used for the test reasonable results for internal and wall-bounded flows in cases
reactor model; these were generated using a GAMBIT grid genera- where the mean pressure gradients are small. We selected this
tor. The grid consisted of approximately 31 million unstructured model because the pressure gradients in this reactor are relatively
cells. The grid number was determined from the largest mesh small.
size that could be used without any variation in results; 31 mil- As shown in Fig. 4, numerical data were obtained at 13 points
lion cells were determined to be adequate to obtain reasonable located before and after the main components. The locations of
data. the points were the same as those provided in the experimental
A commercial CFD code, FLUENT 12.1.2, was used to calculate apparatus.
the pressure distributions in the nuclear reactor. This code solves
the three-dimensional Reynolds-averaged Navier–Stokes (RANS)
equations using a second-order upwind scheme and the SIM- 3. Results and discussion
PLE pressure-correction algorithm. The fluid was considered to be
incompressible water with constant physical properties, Newto- The pressure at the measuring points changed instantly and
nian, non-reactive, and unsteady. The RANS turbulence modeling rapidly due to the pump operation and turbulence perturbation.
approach involved solving the continuity, momentum conserva- Thus, the pressure is presented as averaged values over a 60-
tion, and energy equations. s period. Fig. 5 shows the averaged static pressure difference
The RNG k–ε model with a standard wall function was selected between a location (point X) and the outlet (point 12) for each
for the turbulence model. The k–ε model is a commonly used component of the reactor internals, and the partial pressure drop
222 K.M. Kim et al. / Nuclear Engineering and Design 248 (2012) 216–225

Fig. 10. Pressure distributions with all four pumps operating with the same pulse phase. (a) Pressure distribution of pulsating, nominal, maximum, and minimum inputs.
(b) Ratio of the pulsating pressure amplitude to the averaged pressure at each location.

divided by the entire pressure drop for the experimental setup. the flow passed through the reactor internal. The peak amplitude
The figure also includes calculated pressure data for comparison was 112 Pa in the inlet of the reactor, and changed to 38 Pa due
with experimental results. Among the main reactor internals, the to the mixing of fluid from four pumps in the chamber. After the
greatest pressure drops occurred in the steam generators and the steam generators and the flow-mixing header assembly, the peak
core, with drops of 35.1% and 36.7% of total pressure, respec- values were reduced to 29 Pa and 22 Pa, respectively. Finally, the
tively. The smallest pressure drop appeared in the flow skirt. peak amplitude at the outlet (point 12) was 4 Pa. These results indi-
The calculated average pressures matched the measurement pres- cate that the peak amplitude was reduced after passing through the
sures very well at all 13 points in the reactor, with an error reactor internals.
within 3%. Fig. 7 shows the calculated pressure distributions induced by
Fig. 6 presents the experimental Fast Fourier Transform (FFT) one pulsating and three constant pressure pumps. The simulation
results for five locations (points 0, 1, 4, 6, and 12). As shown in is calculated under pressure inlet condition, which is the pres-
Fig. 6(a), peaks were generated by vortex flow, and by the pumps sure pulse of 12 Hz and the amplitude of 120 Pa. Fig. 7(a) displays
and pulse generator. Among the peaks, two generated by the pulse pressure variations over time at all locations: increased pressure
generator had high amplitudes of more than 70 Pa. The peak fre- variations occurred at four locations (points 0, 1, 4, and 6). The
quencies were 11.7 Hz and 23.4 Hz. Fig. 6(b) presents the changes time duration of high and low values was the same at all the loca-
in amplitude of a pressure pulse at 11.7 Hz; the peak decreased as tions; only the peak amplitudes changed. The reason for this is that
 K.M. Kim et al. / Nuclear Engineering and Design 248 (2012) 216–225 223

Fig. 11. Ratios of the pressure variation for different number of pulsating pumps.

Fig. 12. Ratios of the pressure variation for different phases of pulses.

pressure is transferred as acoustic waves in water. Fig. 7(b) presents pressure amplitude near the flow-mixing header assembly was the
the FFT results from the calculated data; the results are almost result of the non-uniform pressure distributions due to the outlet
identical to the experimental values presented in Fig. 6(b). In the structure of the header assembly, as shown in Fig. 3(c). The pres-
numerical results, the amplitude of the peak at 12 Hz was reduced sure distributions are explained in detail in a previous study (Kim
by the reactor internals. et al., 2011).
Fig. 8 presents ratios of the pulsating pressure amplitude to the Figs. 9–12 present the results for pulses at 20 Hz and ±10% pres-
average pressure difference normalized to the value at the inlet, sure variation of the average pressure difference in the inlet. The
and compares the experimental and numerical results. The experi- figures clearly reveal the pressure changes. Fig. 9 shows the pres-
mental and numerical results matched well. The ratio at point 2 was sure changes over time at each location when four pumps had ±10%
reduced to approximately 25% of that in the inlet (point 0) by the pressure variation of the same phase. As mentioned in the dis-
mixing of the four flow inlets. After passing the steam generators, cussion of Fig. 7(a), the time duration of the highest and lowest
the ratios from point 4 to point 8 were about 40% of the ratio in the pressures at each point was identical because pressure is trans-
inlet. The ratios after the core decreased to less than 30%. The high ferred as an acoustic wave in water, which travels very quickly.
224 K.M. Kim et al. / Nuclear Engineering and Design 248 (2012) 216–225

Fig. 10 presents the pressure characteristics when four pumps approximately 10%. This result indicates that pressure variation
are operated at identical pulse phases. The arrows in Fig. 10(a) could be predicted without any analysis of unsteady pressures.
indicate the ranges in pressure variations at a location, and rectan- 3. The pressure variation changed with the number of pulsating
gular and triangular symbols indicate the results when four pumps pumps. When the number of pulsating pumps was one, two, or
are operated at constant pressure with values 10% higher and 10% three, the relative pressure variation was 25%, 50%, and 75% of
lower than the nominal inlet pressure, respectively. The pressure what was measured with four pulsating pumps. The pressure
variations were within 10% of the high and low constant-pressure variation due to the out-of-phase pulses could be deduced using
distributions. This result indicates that the pressure variation can by summing the pulses.
be predicted without any analysis of unsteady pressure when all 4. Because the peaks of the pulses at all locations were observed
pumps have the same amplitudes and phases. Fig. 10(b) presents at the same time and the pressure rise and fall due to the
the magnitude of the pressure variation and the ratio of the magni- pump-induced pulsation was linear, we could predict the pump-
tude to the average pressure difference at each location for identical induced loads by superposition of the pressures without an
amplitudes and phases at all pumps. The magnitude of the pressure unsteady analysis of the pump pulsation.
variation decreased as the flow passed through the reactor inter-
nals. This trend was similar to that observed in the average pressure
difference under nominal operation. The ratios of the magnitude to Acknowledgments
the average difference at each location were approximately 10%.
This result also indicates that the pressure variation can be pre- This work was supported by the Korea Atomic Energy Research
dicted without any analysis of unsteady pressure when all pumps Institute and the National Research Foundation of Korea (NRF) grant
have the same amplitudes and phases. funded by the Korea government (MEST) (No. 2011-0017673).
Figs. 11 and 12 present the relative pressure variation ratios
for an operating case when the four pumps have the same ampli- References
tudes and phases. As shown in Fig. 11, pressure variation changed
with the number of pulsating pumps: when the number of pulsat- Araseki, H., Kirillov, I.R., Preslitsky, G.V., Ogorodnikov, A.P., 2000. Double-supply-
ing pumps was one, two, or three, the relative pressure variation frequency pressure pulsation in annular linear induction pump. Part I.
Measurement and numerical analysis. Nucl. Eng. Des. 195, 85–100.
was 25%, 50%, and 75% of what was measured with four pulsat- Au-Yang, M.K., Brenneman, B., Raj, D., 1995. Flow-induced vibration test of an
ing pumps. Fig. 12 presents the pressure variation for out-of-phase advanced water reactor model. Part 1. Turbulence-induced forcing function.
pulses. In the case of four different phases, one positive cosine wave, Nucl. Eng. Des. 157 (1–2), 93–109.
Becker, S., Laurien, E., 2003. Three-dimensional numerical simulation of flow and
one negative cosine wave, one positive sine wave, and one negative
heat transport in high-temperature nuclear reactors. Nucl. Eng. Des. 222,
sine wave were imposed at each pump. In the case of three dif- 189–201.
ferent phases, two positive cosine waves, one positive sine wave, Chang, M.H., Kim, S.H., 1997. The Republic of Korea’s R&D Program and activities
and one negative sine wave were imposed. In the case of two dif- for nuclear desalination. In: Proceedings of a Symposium on Desalination of
Seawater (IAEA), Taejon, Korea.
ferent phases, two positive cosine waves and two positive sine Chang, M.H., Sim, S.K., Lee, D.J., 2000. SMART behavior under over-pressurizing
waves were imposed. The magnitude of the pressure variation was accident conditions. Nucl. Eng. Des. 199, 187–196.
deduced from the sum of the pulses. Thus, the pressure variation Cheong, J.S., 2000. An analytical prediction on the pump-induced pressure pulsation
in a pressurized water reactor. Ann. Nucl. Energy 27, 1373–1383.
disappeared under operations with different and opposite phases. Chung, Y.-J., Kim, S.H., Chung, B.-D., Chung, M.-K., Zee, S.-Q., 2006. Two phase natural
In summary, the pulses peaked at all locations at the same time, circulation and the heat transfer in the passive residual heat removal system of
and the pressure rise and fall due to the pump-induced pulsation an integral type reactor. Ann. Nucl. Energy 33, 262–270.
Chung, Y.-J., Lee, S.W., Kim, S.H., Kim, K.K., 2008. Passive cooldown performance of
was linear. Therefore, we can predict pump-induced loads from a 65 MW integral reactor. Nucl. Eng. Des. 238, 1681–1689.
the superposition of pressures without analyzing unsteady pump Durmayaz, A., Yavuz, H., 2001. Exergy analysis of a pressurized-water reactor
pulsation. nuclear-power plant. Appl. Energy 69, 39–57.
Hammond, G.P., 1996. Nuclear energy into the twenty-first century. Appl. Energy 54
The results indicate that we can predict the reduction in pres-
(4), 327–344.
sure variation caused by changing the number of pulsating pumps Heising-Goodman, C.D., 1981. Supply of appropriate nuclear technology for the
and introducing out-of-phase pulses. Our calculated data for local developing world: Small power reactors for electricity generation. Appl. Energy
8 (1), 19–49.
pressure distributions matched with the measured data very well.
Kim, K.M., Lee, B.I., Cho, H.H., Kim, K.R., Park, J.S., 2010. Thermo-hydraulic analysis
Therefore, this method and these data are suitable for calculating for evaluating flow induced loads in the SMART reactor. In: Proceedings of KSME
flow-induced loads and pump pressure loads during the design of Annual Spring Meeting 2010, Seoul, Korea.
a nuclear reactor, and can provide the exact pressure force acting Kim, K.M., Lee, B.I., Cho, H.H., Park, J.S., Chung, Y.-J., 2011. Numerical study on
thermo-hydrodynamics in the reactor internals of SMART. Nucl. Eng. Des. 241
in nuclear reactors during the initial design phase. (7), 2536–2543.
Kim, K.M., Lee, B.I., Cho, H.H., Park, J.S., Jeong, K.H. Experimental and numerical study
on local pressure distributions in a system-integrated modular reactor. Ann.
Nucl. Energy, under review.
4. Conclusions Lee, B.I., Kim, K.M., Cho, H.H., Jeong, K.H., Park, J.S., 2010. Measurement of pres-
sure distributions on the shell surface of the downscaled SMART reactor. In:
Proceedings of KSME Annual Fall Meeting 2010, Seoul, Korea.
The objective of this study was to develop a numerical method Lee, S.W., Kim, S.H., Chung, Y.J., 2009. Development and steady state level experi-
for application to the actual SMART reactor by investigating the mental validation of TASS/SMR core heat transfer model for the integral reactor
pressure characteristics induced by pump pulsation in a scaled- SMART. Ann. Nucl. Energy 36, 1039–1048.
Smith, B.A.W., Derksen, D.D., 1998. Measurement of steady and unsteady hydrody-
down SMART research reactor. We obtained the following results:
namic loads on a nuclear fuel bundle. J. Fluids Struct. 12 (4), 475–489.
Spence, R., Amaral-Teixeira, J., 2008. Investigation into pressure pulsations in a cen-
trifugal pump using numerical methods supported by industrial tests. Comput.
1. The calculated pressure matched the measured pressure very Fluids 37, 690–704.
Strupczewski, A., 2003. Accident risks in nuclear-power plants. Appl. Energy 75,
well at all 13 points in the reactor, within an error range of 3%.
79–86.
This result indicated that the numerical method could be applied Weaver, D.S., Ziada, S., Au-Yang, M.K., Chen, S.S., Paidoussis, M.P., Pettigrew,
to an actual nuclear reactor. M.J., 2000. Flow-induced vibrations in power and process plant components:
2. In test cases with 10% higher and lower pressure amplitudes progress and prospects. J. Pressure Vessel Technol. 122 (3), 339–348.
Yang, S.H., Chung, Y.-J., Kim, H.C., 2005. Assessment of the MDNBR enhancement
and the same phases at all pumps, the ratio of the magni- methodologies for the SMART control rods banks withdrawal event. Ann. Nucl.
tude to the average pressure difference at each location was Energy 32, 1567–1583.
 K.M. Kim et al. / Nuclear Engineering and Design 248 (2012) 216–225 225

Kyung Min Kim received his PhD in Mechanical Engi- Hyung Hee Cho received his PhD in Mechanical Engi-
neering from Yonsei University, Seoul, Korea in 2008. neering from University of Minnesota, Minneapolis, MN
Since March in 2010, he is the research professor in in 1992. In 1995, he joined the Department of Mechan-
school of mechanical engineering, Yonsei University. His ical Engineering, Yonsei University, Seoul, Korea, where
research activities and interests are in the area of ther- he is currently a full professor in the School of Mechan-
mal design, management, and optimization in energy and ical Engineering. His research interests include thermal
power applications. management and cooling of high temperature/heat flux
devices. Dr. Cho is a Fellow of ASME, a Scientific Council
Member of ICHMT (International Centre for Heat and Mass
Transfer) and an Associate Member of the Korea Academy
of Science and Technology.

Byoung In Lee received his MS in Mechanical Engineer- Jin Seok Park received his MS in Mechanical Engineering
ing from Pusan National University, Busan, Korea in 2001. from KAIST, Daejeon, Korea in 1986. In 1986, he joined
From 2001, he has been working for the Samsung Elec- Korea Atomic Energy Research Institute (KAERI), Daejeon,
tronics Co., Ltd. From 2009, he is doing a PhD course in Korea, where he is currently a Principal Researcher.
Department of Mechanical Engineering, Yonsei Univer-
sity, Seoul, Korea.

Donghwi Lee received his BS in Mechanical Engineering Kyeong Hoon Jeong received his PhD in Mechanical Engi-
from Yonsei University, Seoul, Korea in 2010. From 2010, neering from Chonbuk National University, Jeonju, Korea
he is doing a PhD course in the Department of Mechanical in 1995. In 1985, he joined Korea Atomic Energy Research
Engineering, Yonsei University, Seoul, Korea. Institute (KAERI), Daejeon, Korea, where he is currently a
Principal Researcher. He participated in reactor internals
design and research of fluid–structure interaction.