GENES4/ANP2003, Sep.

15-19, 2003, Kyoto, JAPAN

Paper 1048

Nuclear and Thermal Hydraulic Design Characteristics of the SMART Core

Chungchan Lee*, Dae-Hyun Hwang, Sung-Quun Zee and Moon Hee Chang
Korea Atomic Energy Research Institute, Daejon, Korea

SMART (System-integrated Modular Advanced ReacTor) is an integral type reactor of 330 MWth. The core
consists of 57 fuel assemblies which are based on the well proven 17x17 array with UO2 fuel rods loaded in the
Korean commercial PWRs. The active height of the core is 2 m and slightly enriched UO2 of 4.95 wt% is used
for a three year operation cycle. SMART is soluble boron free, and the high initial reactivity is mainly
controlled by burnable absorbers. Less burnable absorber is used at the upper part of the core to obtain smooth
axial power distributions.
Control rods have a fine reactivity control capability using linear pulse motors. There are 49 CEDMs, and
they have the capability to shut the reactor down and to maintain it in a subcritical condition at any time during
the core’s life. Core nuclear characteristics are analyzed using the 3 dimensional core depletion and analysis
code system, CASMO-3/MASTER.
Steady-state thermal margin of the SMART core is evaluated by the subchannel analysis code, MATRA
using a local parameter CHF correlation applicable to low mass velocity conditions. According to the results so
far, it is concluded that a 15% operating margin is ensured.

KEYWORDS: SMART, core, UO2, reactivity, soluble boron free, thermal margin, subchannel, SR-1

I. Introduction developed by KAERI. A local parameter CHF correlation is

Korea Atomic Energy Research Institute (KAERI) has developed accounting for a correction factor applicable to
been developing a small-sized integral reactor, SMART1) low mass velocity conditions. The steady state overpower
(System-integrated Modular Advanced ReacTor) and its margin is evaluated for the SMART core employing the
application system since 1997. SMART produces 330 MW digital core protection and monitoring system.
of thermal energy for supplying the energy for sea water In section II, the nuclear and thermal hydraulic design
desalination as well as for electricity generation. bases of the soluble boron-free SMART core are presented.
For system simplification and enhanced safety, an In section III, core nuclear design and thermal margin
integrated system design and soluble boron free operation assessments are described. The conclusion is presented in
concept are introduced. Major components including 4 main section IV.
coolant pumps, 12 steam generator cassettes, a self-
pressurizer as well as active core are placed in a single II. Design Bases
reactor vessel. This integral design eliminates large break Some important design bases of the SMART core design
LOCA. By adopting the soluble boron free operation are summarized. The core consists of 57 fuel assemblies of
concept, there is no need to control the boron concentration the well proven 17x17 array with UO2 fuel rods. Slightly
of the primary coolant. Hence, the system can be further enriched UO2 of 4.95 wt.% is used for the cycle length of 3
simplified. However, an emergency boron storage tank is years. SMART is designed to operate without soluble boron
provided as a diverse shutdown means in case of scram from hot full power to cold zero power including the
failure. The cycle length of the SMART is chosen to be 3 refueling condition. Although an emergency boron injection
years long to increase the plant availability and economy. system is provided for diverse shutdown means, SMART can
Nuclear design has been performed using the CASMO- be safely shut down from any condition and can be
3/MASTER system. CASMO-32) was developed by Studsvik maintained at a subcritical condition without soluble boron.
and is used for group cross section generation. MASTER3,4) The control rod system can provide enough shutdown
(Multi-purpose Analyzer for Static and Transient Effects of reactivity with the most reactive control rod stuck at the fully
Reactors) is a nuclear design and analysis code developed by withdrawn position. The control rod system can also
KAERI based on a two-group diffusion theory in a three- maintain the reactor at a subcritical condition in a cold
dimensional Cartesian or hexagonal geometry. The (20°C) condition without soluble boron. This requirement
uncertainty of the code system is evaluated and approved by can be fulfilled with burnable absorbers and 49 CEAs
the Ministry of Science and Technology of Korea.5) (Control Element Assemblies). The core power distribution
The analysis of thermal hydraulic characteristics has been should be controlled so that the specified acceptable fuel
conducted using a subchannel analysis code MATRA6) design limits are not exceeded.
During normal operation and anticipated operational
* Chungchan Lee, Tel. +82-42-868-8758, Fax. +82-42-868- occurrences, the probability that DNB could not take place is
8990, E-mail: [email protected] at least 95% with 95% confidence. The maximum
temperature of the fuel should be lower than the melting type B. Also, type C fuel has 8 BP rods cutback to
temperature. compensate for the control rod insertion and to reduce local
power peaking.
III. Core Design
1. Nuclear Core Design Description 1.20

SMART core is composed of 57 fuel assemblies, which is 1.15 4 BP Cutback

based on the 17x17 KOFA7) (Korean Standard Fuel
Assembly) that was designed by KAERI/Siemens-KWU and 1.10
8 BP Cutback

K-Infinity
used in the 900 MWe Westinghouse type Korean PWR’s. 1.05

The active fuel height of the SMART is 200 cm. Using a 1.00
4.95 wt.% enrichment of U-235, the SMART core can be A
B
operated for 3 years without refueling.
0.95
C

In Fig. 1, SMART core loading pattern is presented, 0.90

which consists of 3 fuel types. Since the neutron fluxes are 0.85

higher in the core central region than in the peripheral region, 0 10 20 30 40 50 60

Burnup (MW D/Kg)
more burnable absorber rods with a higher concentration are
used in fuel type C than in fuel type B.
Fig. 2 Infinite multiplication factors (k∞) of fuel assemblies

The control rods are grouped in 4 regulating banks and 2

J H G F E D C B A
startup banks. Startup banks are at the fully withdrawn
position during power operation. The regulating banks are
A B A 1 used to control core reactivity during normal operation. Ag-
B B B B B 2 In-Cd is used for the control rods. Figure 3 shows the
control rod pattern. 49 control rods are capable of controling
B B C C C B B 3 the core excess reactivity at all operating conditions.
A B C C C C C B A 4
B B C C C C C B B 5
J H G F E D C B A
A B C C C C C B A 6
B B C C C B B 7 1
S2
B B B B B 8
S1 R1 R1 R1 S1 2
A B A 9
S1 S2 R2 R2 R2 S2 S1 3
R1 R2 R3 R4 R3 R2 R1 4
A 4.95w/o U-235 / 28 Al2O3-B4C Shim / 12 Gd2O3-UO2
S2 R1 R2 R4 R4 R4 R2 R1 S2 5
B 4.95w/o U-235 / 20 Al2O3-B4C Shim / 4 Gd2O3-UO2
R1 R2 R3 R4 R3 R2 R1 6
C 4.95w/o U-235 / 24 Al2O3-B4C Shim / 4 Gd2O3-UO2
S1 S2 R2 R2 R2 S2 S1 7
S1 R1 R1 R1 S1 8
Fig. 1 SMART core loading pattern
S2 9
The number and concentration of the burnable absorber
rods in each fuel type are selected so that reactivity of each
R Regulating Bank
assembly can be as flat as possible. Therefore the power
distribution and critical control rod position do not change S Startup Bank
much during the cycle. Fuel type B and C have less burnable
absorber rods at the top of the core to compensate for control Fig. 3 Control rod groups
rod insertion. Infinite multiplication factors (k∞) of 3 fuel
types are shown in Fig. 2. The critical control rod positions during hot full power
As can be seen in Fig. 2, fuel type A is designed that k∞ is conditions are shown in Fig. 4. The overlap distance
low initially and it rises quickly as burnup increases. This between the groups is determined so that both the power
ensures a sufficient shutdown margin especially at the peaking and ejected rod worth are minimized. The critical
beginning of the cycle because the control rods are not rod position does not change much during the cycle because
installed in the fuel region A. Fuel type B and type C show the core excess reactivity does not change much using the
two reactivity curves, respectively. 4 burnable poison rods fuels shown in Fig.2. A relatively constant control rod
are filled with a dummy pellet of Al2O3 at the top region of position is important for ensuring shutdown margin and
power distribution control.
0.10
2.2

0.05
200 2.0
R2
180 0.00

Power Peaking (Fq)

1.8
160

Axial Offset
-0.05
140 1.6
R3
120
Height (cm)

-0.10
1.4
100

80 -0.15 AO
1.2
R4 Fq
60
-0.20 1.0
40

20 0 200 400 600 800 1000 1200

0 Burnup (EFPD)
0 200 400 600 800 1000 1200
Burnup (EFPD)
Fig. 5 Axial offset and power peaking factor

Fig. 4 Critical control rod position at HFP, Eq. Xe

reload cycle, the average core burnup of the second cycle is
about 31 MWD/KgU. Table 1 shows that the excess
In Fig. 5, the axial offset (AO) and 3-dimensional power reactivity and reactivity defects are fully controlled by the
peaking factor is shown. Axial offset is kept between -20% control rods. It is shown that the minimum shutdown margin
and 0% during the cycle. The behavior of AO is similar with of 1%∆ρ is secured in the cold (20°C) condition.
that of the critical control rod position. The maximum of the Fuel temperature coefficient (FTC) of SMART is about
power peaking factor (Fq) is 2.05. Since the average linear the same as that of a commercial PWR. However, the
power density of SMART is 119.9 W/cm, the sufficient moderator temperature coefficient (MTC) is more negative
margin is secured against the fuel centerline melting limit of at the beginning of the cycle (BOC) than at the end of the
591 W/cm. cycle (EOC). This is due to the soluble boron free operation.
In Table 1, some important nuclear parameters are This negative MTC enhances reactor safety.
summarized. Cycle average burnup of 26.16 MWD/KgU is
based on a single batch operation. For the optional 1.5 batch

Table 1 SMART core nuclear design parameters

Descriptions Values
Burnup and Cycle Length
Cycle Length, Effective Full Power Days 990
Core Average Burnup at EOC (Cycle 1), MWD/KgU 26.2
Reactivity, %∆ρ
Minimum Excess Reactivity at HFPa, Eq. Xe and Sm > 1.0
Excess Reactivity at Cold (20°C) Condition, BOC 14.8
Control Rod Worth, %∆ρ
HFP Scram Rod Worth (from critical) 30
CZPb Total Bank Worth 25
CZP Bank Worth at Stuck Rod Condition 20
Reactivity Defects, %∆ρ
Xenon Worth 1.9
Power Defect (HFP to HZPc) 1.4
Temperature Defect (HZP to CZP) 8.1
keff with all Rods Inserted at Cold (20°C ) Condition < 0.95
Shutdown Margin (%∆ρ) > 1.0
Reactivity Coefficients at HFP
Moderator Temperature Coefficients, pcm/°C -63.
Fuel Temperature Coefficients, pcm/°C -3.0
a
HFP: Hot Full Power
b
CZP: Cold Zero Power
c
HZP: Hot Zero Power
2. Load-follow Operation flow rates including the SMART core operating conditions.
The daily load-follow operation capability of the The CHF data base for rod bundles has been collected from
SMART core is investigated for various scenarios. The open literature. All of the 2518 data points are employed in
simulation is performed by MASTER assuming that the the correlation development. The pressures of the data base
coolant temperature rise is proportional to the core power range from 69 to 171 bars, mass fluxes from 470 to 4973
level. In this study, SMART core parameters are examined kg/m2/s, and the heated lengths from 1.37 to 4.27 m.
during a typical 14-2-6-2 (100%-50%-100% power) daily
load-follow operation. BOC and EOC data are presented in
Figs. 6 and 7, respectively. 2000

The control rods are required to operate only three times P=140 bar, Tin=200 C
a day until the outlet temperature reaches the target if the 1500

CHF (kW/sq-m)
core outlet coolant temperature reaches the programmed set
point (target temperature ±3°C). This is due to the slightly
1000
decreasing Tavg control and strongly negative MTC of
SMART. As the power increases, fuel temperature rises and
core reactivity decreases due to the negative fuel 500

temperature feedback. On the other hand, a decreased Data(Winfrith) AECL-86 KRB-1

average coolant temperature adds positive reactivity to the 0

0 500 1000 1500 2000 2500
core and compensates for the reactivity defect caused by the Mass flux (kg/sq-m/s)
fuel temperature rise.
The figures also show the behavior of the peak local
Fig. 8 Parametric trend of CHF with respect
power and axially integrated peak power as well as the axial
to the mass flux
offset during the daily load-follow operation. Both
normalized peak powers do not increase much from the
The local thermal-hydraulic conditions inside the test
steady state and the axial offset is within the wide operation
bundles are calculated by MATRA code. MATRA is a
region. Hence the load-follow capability of the SMART
thermal-hydraulic analysis code based on the subchannel
core is demonstrated.
approach for calculating the enthalpy and flow distributions
in fuel assemblies for both steady-state and transient
3. Steady State Thermal Margin Evaluation
conditions6). In the subchannel approach, the conservation
SMART is designed to be operated at a low core inlet
equations have been derived assuming an axially dominant
temperature and low core flow rate in comparison with the
one-dimensional flow. Therefore, the accuracy of a
existing PWRs. It is known that CHF reveals a linear
subchannel analysis code is fairly dependent on the
dependency on the mass flux under fixed inlet conditions,
modeling of the interchannel exchanges between the
and its derivative to the mass flux at a low flow condition is
adjacent subchannels such as diversion cross flow and
quite different from that at high flow conditions8). At low
turbulent mixing. The modeling of lateral exchanges of
flow conditions, the CHF mechanism is mainly subject to
mass, momentum, and energy across the imaginary
the liquid film dryout phenomenon under annular flow
interfaces dividing the subchannels is a unique feature of
regimes. In this case the first derivative of CHF with respect
the subchannel analysis. When the subchannel analysis code
to the mass flux, ∂q"C/∂G, is close to the value calculated by
is applied for evaluating the CHFR margin in the SMART
the energy balance relationship under a complete
core, the lateral exchange model affects not only the
evaporation condition. On the other hand, CHF occurs
accuracy of the local parameter CHF correlation but also
mainly due to the departure from nucleate boiling at high
the hot channel conditions used for the evaluation of the
flow conditions and the magnitude of ∂q"C/∂G decreases
thermal margin. Information on the spacer grid, such as the
compared with that at low flow conditions. These
pressure loss coefficient and thermal diffusion coefficient
characteristics have been observed in many experimental
(TDC), is essential for the subchannel analysis of the test
works, and some of the round tube correlations adopted
bundles. Especially, TDC is a parameter of great importance
different equations according to the mass flux conditions8,9).
determining the accuracy of the subchannel analysis. For
It is also observed in rod bundle experiments. Figure 8
the CHF data base employed in this study, the TDC values
shows a comparison of CHF data obtained from a 5-by-5
have been evaluated from the thermal mixing data13) except
test bundle10) with the predictions made by the KRB-1 rod
for the six test bundles (TS 512 ～ TS 517) from the EPRI
bundle CHF correlation11) and the CHF lookup table for the
report14). In this study, TDC values for the EPRI test
round tubes12). As shown in the figure, the KRB-1
bundles are approximated from the grid loss coefficient data.
correlation that has been developed on the bases of CHF
That is, a TDC value of 0.005, which is a typical value for
data at high flow conditions, tends to over predict the CHF
no mixing-vaned grids, is assigned to the test bundles with
values at low mass flux (approximately less than 1000
small grid loss factors, while 0.05 is applied to the test
kg/m2/s) conditions. Thus, a local parameter CHF
bundles with large grid loss factors.
correlation is developed that is applicable to the low mass
 314 200
100 313
180 R2
312
90
311 160
80 310
140
309

cm W ithdrawn
70

Temperature ( C)
308 120 R3
Power (%)

o
60 307
306 100
50
305 80
40 304
60
R4
30 303
302
20 40
301
10 300 20
299
0 0
0 5 10 15 20 25 0 5 10 15 20 25 0 5 10 15 20 25

Time (hours) Time (hours) Time (hours)

1.6
2.0 upper limit = 0.3
0.2

1.4
1.8 0.0

1.2
Fq*Power

Axial Offset
Fr*Power

1.6 -0.2

1.0
1.4 -0.4

0.8
-0.6
1.2
lower limit = 0.84P - 1.0 for p > 0.25
0.6
-0.8
1.0
0 5 10 15 20 25 0 5 10 15 20 25 0 5 10 15 20 25
Time (hours) Time (hours) Time (hours)

Fig. 6 Daily load-follow operation (BOC, 100-50-100% Power)

314 200
100 313
312 180
90 R2
311
160
80 310
309 140
70
cm W ithdrawn
Temperature ( C)

308
Power (%)

120
o

60 307 R3
50 306 100
305
40 80
304
30 303 60
302
R4
20 40
301
10 300 20
299
0 0
0 5 10 15 20 25 0 5 10 15 20 25 0 5 10 15 20 25
Time (hours) Time (hours) Time (hours)

2.2 1.6
0.2
upper limit = 0.3

2.0
1.4
0.0
1.8
Axial Offset
Fq*Power

Fr*Power

1.2
-0.2
1.6

1.0
1.4 -0.4

0.8
1.2 -0.6
lower limit = 0.84P - 1.0 for p > 0.25
1.0 0.6
-0.8

0 5 10 15 20 25 0 5 10 15 20 25 0 5 10 15 20 25
Time (hours) Time (hours) Time (hours)

Fig. 7 . Daily load-follow operation (EOC, 100-50-100% Power)

 The functional form of the CHF correlation, named SR-1, Table 2 Major thermal-hydraulic parameters of SMART
yields PWR
Parameters SMART
F ⋅F (Kori-3)
q "CHF = q "CHF , BASE ⋅ G SG (1) Core thermal power, MW 330 2775
FNU
System pressure, bar 150 155
The functional form of the base equation, q”CHF,BASE, Effective core height, m 2.0 3.66
implies two distinct trends of the CHF with respect to the Core average heat flux, kW/m2 402 599
local quality and the local mass flux. Firstly, CHF decreases Core average LHGR, kW/m 12.0 17.9
as the local quality increases. Secondly, CHF increases with Core inlet temperature, oC 270 291
the local mass flux in low quality conditions while the CHF Core temperature rise, oC 40 37
decreases as the mass flux increases in high quality Core average mass flux, kg/m2/s 1010 3300
conditions. The following equation comprises these
characteristics.
overpower margin estimated from DNB related transients,
q "CHF , BASE = A 1 − A 2 ⋅ G ⋅ χ + A 3 ⋅ G (2)
and UNC is the uncertainty of the core monitoring system.
The coefficients A1, A2, and A3 are dependent on five The SR-1 CHF correlation with the MATRA code is applied
geometry parameters (heated length, hydraulic diameter, to the DNBR calculations. The design limit DNBR is
heated equivalent diameter, grid spacing, and distance from calculated as 1.41 according to the statistical DNB design
CHF to the last grid) and the pressure. The correlation procedure14) by employing the uncertainties of the
coefficients are optimized to fit the experimental data using engineering enthalpy rise factor, engineering heat flux
a nonlinear least square fitting method. FG, FSG, and FNU factor, and TH code uncertainty. The ROPM and UNC are
mean the correction factors for low mass velocity, spacer approximated as 1.18 and 1.09 from the design data of a
grid, and nonuniform axial power shapes, respectively. FSG typical 1000 MWe PWR. The AOPM is calculated from the
is a constant, accounting for the effects of the mixing vanes subchannel analysis of a 1/2 fuel assembly with various
on the CHF. The Tong’s F-factor model is used for the radial peaking factors. According to the results so far
calculation of FNU. The correction factor for low mass achieved, it is concluded that a 15% operating margin is
velocity conditions, FG, is devised to improve the prediction ensured if the radial peaking factor does not exceed 1.72
accuracy in low mass velocity conditions at which the high that is well above of the normalized radial peak power
velocity CHF correlations tend to over-predict the CHF during load-follow operations.
values considerably. The SR-1 correlation is validated by
statistical tests and by the analysis of parametric trends. The IV. Conclusion
correlation limit DNBR is evaluated as 1.331 for the low For a soluble boron free operation, the core design
velocity data base(i.e., the mass velocity is less than 1500 description and nuclear characteristics are presented. The
kg/m2/s) including the SMART core conditions. SMART core consists of 57 fuel assemblies that are based
The thermal hydraulic conditions of the SMART core are on a 17x17 KOFA. The active height of the SMART core is
listed in Table 2 in comparison with those of a typical PWR 2m. Using 4.95wt.% UO2 fuel, a 3-year cycle length is
core. Because of the low LHGR of the SMART core, it is accomplished. Core excess reactivity required for a 3-year
expected that the thermal hydraulic design parameter that cycle length is controlled by 49 CEDMs and burnable
puts a limitation on the thermal power is the minimum absorber rods.
departure from the nucleate boiling ratio (DNBR). This is Load-follow capability of the SMART core is also shown
confirmed from the results of the thermal margin for 14-2-6-2 daily load-follow operations. Core reactivity is
estimations associated with various core protection and controlled by MTC and control rods. Power peaking factors
monitoring systems13). Due to the low head of the canned do not increase much during load-follow operations.
motor reactor coolant pumps, the mass velocity of the The steady state thermal margin of the SMART core is
SMART core is approximately one-third of that of a typical evaluated by employing the MATRA/SR-1 CHF correlation
1000 MWe PWR. The increase of the thermal margin due to system. According to the results so far achieved, it is
the low power density, short heated length of the fuel rod, concluded that a 15% operating margin is ensured.
and low core inlet temperature counterbalances the margin
decrease attributed to the low mass velocity. Acknowledgement
The steady state thermal margin of SMART employing a This study has been carried out under the Nuclear R&D
digital core protection and monitoring system is defined as Program sponsored by the Ministry of Science and
AOPM Technology of Korea.
Thermal Margin = , (3)
ROPM × UNC References
where AOPM is the available overpower margin from a 1) Moon H. Chang, et al., “Advanced Design Features
nominal condition to the fuel design limit from the aspect of Adopted in SMART,” Proc. of International Seminar of
DNB (i.e., design limit DNBR), ROPM is the required Status and Prospects for Small and Medium Sized
 Reactors, May 27-31, Cairo, Egypt, (2001). convection boiling, Advances in Chemical Engineering
2) M. Edenius, B. Forssen, CASMO-3 User’s Manual, 7, 207-293 (1968).
STUDSVIK/NFA-89/3, (1989). 9) Y. Katto, “A generalized correlation of critical heat
3) B. O. Cho, et al., MASTER 2.1 User’s Manual, flux for the forced convection boiling in vertical
KAERI/UM-06/2000, Korea Atomic Energy Research uniformly heated round tubes,” Int. J. Heat Mass
Institute (KAERI), (2000). Transfer 21, 1527-1542 (1978).
4) J. Y. Cho, et al., “Hexagonal CMFD formulation 10) F. H. Bowditch, D. J. Mogford, An experimental and
employing triangle-based polynomial expansion nodal analytical study of fluid flow and critical heat flux in
kernel,” Proc. 2001 M&C Meeting, Salt Lake City, PWR fuel elements, AEEW-R2050 (1987).
Utah, USA, (2001). 11) D. H. Hwang, et. al., “Evaluation of thermal margin in
5) J. S. Song, et al., Verification and Uncertainty a KOFA-loaded core by a multichannel analysis
Evaluation of CASMO-3/MASTER Nuclear Analysis methodology,” J. KNS 27, 518-531 (1995).
System, KAERI/TR-1579/2000/A, Korea Atomic 12) D. C. Groeneveld, et. al., “1986 AECL-UO critical
Energy Research Institute (KAERI), (2000). heat flux lookup table,” Heat Transfer Engineering 7,
6) Y. J. Yoo, et al., “Development of a subchannel 46-62 (1986).
analysis code MATRA applicable to PWRs and 13) Y. K. In, et al., “Assessment of core protection and
ALWRs,” J. KNS 31, 314-327 (1999). monitoring systems for an advanced reactor SMART,”
7) Lothar Heins, et al., Fuel Design Report for 17x17 Annal of Nuclear Energy 29, 609-621 (2002).
Assembly, KRAFTWERK UNION AG, (1987). 14) H. Chelemer, et al., Improved thermal design
8) R. V. Macbeth, The burnout phenomenon in forced- procedure, WCAP-8568 (1975).