Progress in Nuclear Energy 71 (2014) 39e51

Contents lists available at ScienceDirect

Progress in Nuclear Energy

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

Development of thermal hydraulic and margin analysis code for

steady state forced and natural convective cooling of plate type fuel
research reactors
Daeseong Jo*, Jonghark Park, Heetaek Chae
Korea Atomic Energy Research Institute, 1045 Daeduk-Daero, Dukjin-Dong, 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: Thermal hydraulic and Margin Analysis code for Plate type fuel assembly, TMAP-1.0, was developed to
Received 11 March 2013 analyze the thermal hydraulic characteristics and thermal margins for steady state forced convective and
Received in revised form natural circulation cooling in research reactors facilitated with plate type fuel assemblies. Forced
10 October 2013
convective cooling during normal operation at nominal power, and natural convective cooling during a
Accepted 11 November 2013
training operation at low power, were considered for the model. Since the fuel cooling channels in plate
type fuel assemblies are independent such that there is no cross flow from one channel to another, a one-
Keywords:
dimensional analysis is suitable to predict the thermal hydraulic characteristics through the fuel cooling
Research reactor
Plate type fuel
channels. The flow directions considered in the present model were downward during forced convective
Forced convection cooling and upward during natural convective cooling. The most unfavorable condition in the core was
Natural convection simulated by introducing engineering hot channel factors as well as radial and axial power peaking
Thermal margin factors. Thermal margins such as Onset of Nucleate Boiling (ONB) temperature margin and Departure
from Nucleate Boiling Ratio (DNBR) were estimated for reactor safety and design purposes. Furthermore,
coolant, cladding outer surface, and fuel centerline temperatures were estimated along the axial direc-
tion with and without the engineering hot channel factors applied. For forced convective cooling, the
prediction resulting from the present analyses were compared with experimental data taken by Sudo
et al. (1984) and predictions by a subchannel analysis code, MATRA. For natural convective cooling,
coolant velocities resulting from the balance of buoyancy and friction forces predicted by the present
analyses were compared with predictions by RELAP5/MOD3. The comparisons showed good agreements
for both forced and natural convective cooling.
Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction 2002). The operating mode and design of research reactors are
very different from one reactor to another, and are dependent on
Interest in safety issues of research reactors is constantly the application. Also, different fuel types such as bare rod, rod with
increasing owing to their scientific and industrial applications. The fins, plate, or curved plate are used in research reactors. To achieve
primary purpose of research reactors is not power generation, but a high neutron flux, an effective fuel geometry is desired to remove
various activities such as research, training, education, material heat deposited into the fuel and cladding. For example, HANARO
tests, radioisotope production, neutron transportation mutation (High-flux Advanced Neutron Application ReactOr) uses rod type
doping, neutron beam applications, and others. The IAEA research fuel with 7 fins located circumferentially to increase the heat
reactor database (RRDB) in 2011 states that nearly 700 research transfer areas. Most of the research reactors use straight plate type
reactors have been built: 232 operational, 288 decommissioned, fuel. Some reactors use curved plate type fuel if the channel flow
148 shutdown, and 12 temporarily shutdown. The RRDB can be velocity is close to the critical velocity. From a thermal hydraulic
accessed via the IAEA’s Nucleus portal at “nucleus.iaea.org/RRDB/”. design point of view, an upward flow direction is preferred, but a
Table 1 lists general information of research reactors including downward flow provides advantages such as accessibility to the
reactor configurations and operational conditions (Park et al., pool top owing to the low pool top radiation, a simple locking
system of the fuel assemblies and irradiation blocks, easy handling
* Corresponding author. of irradiation targets and fuel assemblies, and less flow resistance
E-mail address: [email protected] (D. Jo). during shutdown rod insertion.

0149-1970/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.pnucene.2013.11.006
40 D. Jo et al. / Progress in Nuclear Energy 71 (2014) 39e51

Table 1
General information of research reactors.

Parameter HANARO OPAL TRR-II PARR-1 JRR-3 JRR-4 RSG-GAS MNR JHR

Power [MW] 30 20 20 10 20 3.5 30 2 100

Site Korea Australia Taiwan Pakistan Japan Japan Indonesia Canada France
Reactor type Open pool Open pool Open pool Open pool Open pool Open pool Open pool Open pool Open pool
Fuel type Rod with fins Plate Plate Plate Plate Plate Plate Curved Plate Curved plate
Fuel material U3Si U3Si2 U3Si2 U3Si2 U3Si2 U3Si2 U3O8 U3Si2 U3Si2
Fuel enrichment (% U-235) 19.75 19.77 19.75 19.99 20 19.75 19.75 19.75 19.75
Fuel density [g/cm3] 3.15 4.8 4.8 3.28 4.8 3.8 2.96 0.74 4.8e6.0
No. of standard fuel assemblies (SF) 20/12 16 25 29 26 20 40 28 37
No. of plates per SF 36/18 21 21 23 21 15 21 18 6/17
No. of control fuel assemblies (CF) e e 6 5 6 e 8 6 e
No. of plates per CF e e 21 13 16 e 15 10 e
Fuel meat thickness [mm] 6.35/5.49 (dia.) 0.61 0.51 0.51 0.51 0.50 0.54 0.51 0.51
Active length [mm] 700 615 600 600 750 600 600 600 600
Coolant channel thickness [mm] 2.5 2.45 2.58 2.1 2.35 4.1 2.4 3.0 1.84
Averaged heat flux [kW/m2] 121 kW/m 691.3 409 181 360 150 415 45 1900
Coolant H2O H2O H2O H2O H2O H2O H2O H2O H2O
Flow direction Upward Upward Downward Downward Downward Downward Downward Downward Upward
Total flow rate [kg/s] 700 525 600 260 663 135 900 75 1200
Coolant velocity in core [m/s] 7.3 8.2 6.24 2.35 5.83 1.45 5.56 0.69 15
Inlet coolant temperature [
C] 35 38 40 38 35 35 40 30 25
Averaged core DT 10.0 9.1 7.8 9.1 8.7 e 8.0 6.4 16.0
Pool depth [m] 13.4 14.1 11.5 10 8.5 9.4 13 9.35 e
Reflector D2 O D2O D2O Gr Be/D2O Gr Be Gr/Be Be/D2O

Because of various research reactor designs regarding opera- approach to calculate the enthalpy and flow distribution in the fuel
tional mode, fuel type, and applications, every research reactor has assemblies and reactor core for both steady state and transient
unique thermal hydraulic characteristics. To analyze the thermal conditions (Hwang et al., 2008; Yoo et al., 1999; Kim et al., 2002).
hydraulic characteristics of research reactors as well as support the MATRA uses nearly the same governing equations as those of
design and safety analyses, computer codes have played an COBRA-IV-I, which were derived for various channel configurations
important role in nuclear technology. A lot of effort has been put or rod arrays, but there are some implementations i.e., the non-
into development and commercialization of some computer codes square-lattice pressure drop model, wide-range applicable void
for the utilization of nuclear power reactors such as RELAP, PARET, fraction model, improved lateral transport model, and others.
and TRAC. Unfortunately these codes were originally designed for Although there are many computer codes, they are not directly
power reactors in which a fuel rod type and open channel lattice applied to a newly designed research reactor. Since every research
were used, and its application to research reactors is not straight-
forward or even proper sometimes (Bousbia-Salah and
Hamidouche, 2005). Even though some of these programs have Ambient
included some packages in their latest versions to be applicable for Pool Flap
the case of research reactors, the use of these codes is still so- PCS
valve
phisticated and requires a lot of time and energy for input prepa- pipe
ration and output processing. For previous reasons, many research PCS
Screen
institutes have tried to develop simpler thermal hydraulic codes to
design, license, and evaluate the performance of research reactors
under various conditions. AECL (Atomic Energy of Canada Limited) Flap
Upper guide valve
developed CATHENA, which is a two-fluid one-dimensional code
structure
used for dynamic simulations of reactor transients involving ther-
mal hydraulics and kinetics. It was originally designed for the
CANDU reactor and was subsequently modified for use with
MAPLE-X10 (Lee and Wilkin, 1996). JAERI (Japan Atomic Energy Top of fuel
PCS
Research Institute) developed COOLOD-N2, which was applied to
pipe
evaluate steady state thermal hydraulic analyses for JRR-3
Fuel plate
(Kaminaga, 1997). This code is a revised version of the COOLOD-N (unheated + active)
code, and is applicable not only for research reactors in which
plate-type fuel is adopted, but also for research reactors in which
rod-type fuel is adopted. The COOLOD-N2 code can calculate Bottom of fuel
coolant and fuel temperatures under forced convection cooling
Grid plate
mode and natural convection cooling modes. The COOLOD-N2 code
is available for Nuclear energy Agency (NEA) data bank member Lower
countries. KAERI (Korea Atomic Energy Research Institute) devel- plenum
oped MATRA-h code to analyze the thermal hydraulic characteris-
Lower PCS
tics of HANARO (Park et al., 2008; Lim et al., 1999). MATRA-h is an plenum pipe
interim version of the subchannel analysis code MATRA, which was
developed for thermal hydraulic analyses based in the subchannel Fig. 1. Schematic drawing of open pool type research reactor.
 D. Jo et al. / Progress in Nuclear Energy 71 (2014) 39e51 41

Outer coolant channel Inner coolant channel channels (inner and outer channels) as shown in Fig. 2. The inner
coolant channels, which have fuel plates on both sides, can be
modeled as a coolant channel heated from both sides. The outer
coolant channels, which have a fuel plate on one side and an
adjacent assembly side plate (or a wall of reflector block or a wall of
guide tube of a control absorber assembly) on the other side, can be
modeled as a coolant channel heated from one side.
The average and hot channels are modeled by applying a radial
power peaking factor. The hottest channel in the core, where the
lowest thermal margins are, is simulated by additionally intro-
ducing engineering hot channel factors.

2.2. Governing equations

Side plate Fuel plate

2.2.1. Forced convective cooling
Because the coolant channels of a plate type fuel assembly are
Fig. 2. Schematic drawing of fuel assemblies.
independent (no cross flow from one channel to another), steady
reactor has a unique design or features, the present study is focused state thermal hydraulics through the coolant channels can be
on developing a thermal hydraulic and thermal margin analysis analyzed by a steady state one-dimensional computational calcu-
code of an open pool type research reactor with plate type fuel lation. The coolant temperature distribution along the axial direc-
assemblies, TMAP-1.0. The computer code was able to analyze both tion from the inlet is determined by solving the simplified one-
forced and natural convective cooling. The total power peaking dimensional energy equation as shown in Eq. (1) (see Appendix A).
factors and engineering hot channel factors were implemented to
dTb
simulate the average and hot channel. Upper guide structure, _ p
mC ¼ Ph q00 (1)
plenum, and flap valves were modeled to generate a flow path dz
during natural convective cooling. For forced convective cooling, Since the coolant is incompressible, the coolant velocity change
the prediction resulting from the present analyses was compared through the core owing to the coolant temperature change (or the
with experimental data taken by Sudo et al. (1984) and predictions coolant density change) is determined by the conservation of mass
from the subchannel analysis code MATRA. For natural convective given as
cooling, coolant velocities resulting from the balance of buoyancy _
m
and friction forces predicted by the present analyses were v ¼ (2)
rAf
compared with predictions by RELAP5/MOD3.
The pressure drop is estimated as
2. Model and methodology

v2
dp ¼ rg  FDIR  dz þ f dz (3)
2.1. Model description 2gDh
The first term considers the hydrostatic pressure change due to
A typical open pool type research reactor, illustrated in Fig. 1, was
the elevation change, and the second term considers the friction
modeled to analyze the steady state thermal hydraulic characteris-
loss through the distance dz. FDIR is 1 for downward flows, and
tics and thermal margins for reactor safety and design purposes. The
FDIR is 1 for upward flows.
reactor is submerged under water. The reactor may be operated at
The friction factor f in Eq. (3) is estimated as
forced and natural convection cooling modes. The forced convection
cooling by the Primary Cooling System (PCS) is a normal operation f ¼ 0:316Re0:25 for Re  2500 (4)
mode when the reactor is operating at nominal power. The natural
convective cooling by natural circulation is a training mode when
the reactor is operating at low power. Also, the cooling capacity by f ¼ Cf Re1 for Re < 2500 (5)
natural convective cooling is one of the important design consid- Eq. (4) is the Blasius friction factor for Re larger than 2500. The
erations during normal shutdown to effectively remove decay heat friction factor constant Cf in Eq. (5) developed by Eckert and Irvine
generated from the core. The coolant flowing downward through the (1957) for incompressible fluids flowing through rectangular cross
core during forced convection cooling, and the coolant flowing up- sectional ducts is listed in Table 2. For flows through pipes, the
ward through the core during natural convection cooling are friction factor constant Cf becomes 64.
considered in this model development. The pressure drop through a sudden contraction and expansion
The flow path for the forced convective cooling is generated as: is determined as
PCS / pool / screen / upper guide structure / top of
1 2
fuel / fuel plate / bottom of fuel / grid plate / lower dp ¼ K rv (6)
2
plenum / PCS pipe / PCS. The flow path for the natural circu-
lation cooling is generated through the flap valve: pool / flap K factor for a sudden expansion or contraction is determined as
valve / PCS pipe / lower plenum / grid plate / bottom of
fuel / fuel plate / top of fuel / upper guide  2
K ¼ 1  AAu for a sudden expansion
structure / screen / pool. The flap valve is closed by the pressure d

 2 (7)
difference between the pool pressure and inside the outlet pipe
K ¼ 1 1 for a sudden contraction
pressure at the same elevation in the pool. If the pressure difference Cc
is less than the designed value, the flap valve is open and a flow
path is generated via the flap valve. The core consists of a number of where Au and Ad are the upstream and downstream cross sectional
fuel assemblies, and each fuel assembly has several coolant areas, and Cc is the contraction coefficient listed in Table 3.
42 D. Jo et al. / Progress in Nuclear Energy 71 (2014) 39e51

Table 2 where kc is the thermal conductivity of the cladding and thc is the
Friction factor constant. thickness of the cladding.
Channel ratio (width/thickness) Cf The fuel centerline temperature is estimated by solving one-
1 58
dimensional heat conduction equation with heat generated by
2 63 the fuel meat as
3 69 !
4 72.5 thf
00
5 77 Tf ¼ Tw;i þ q (12)
6.3 80 4kf
8 83
11 85 where kf is the thermal conductivity of the fuel and thf is the
15 88
thickness of the fuel.
18 89
100 96 If there is an oxide layer, then the cladding inner surface and the
fuel centerline temperatures are increased by DToxide. DToxide can be
estimated as


thoxide
Table 3 DToxide ¼ q00 (13)
koxide
Contraction coefficient.

Area ratio (Ad/Au) Cc where koxide and thoxide are the thermal conductivity and the
0.1 0.624 thickness of the oxide layer, respectively. The temperature profile in
0.2 0.632 a fuel plate is shown in Fig. 3. The flow conditions on the left hand
0.3 0.643 side and the right hand side of the fuel plate are equal, so that the
0.4 0.659 maximum fuel temperature appears in the middle of the fuel meat.
0.5 0.681
0.6 0.712
0.7 0.755
0.8 0.813 2.2.2. Natural convective cooling
0.9 0.892 Unlike the forced convection, the coolant velocity through the
1.0 1.000 core is not an input parameter. The coolant velocity is determined
when the friction force and buoyancy force are balanced (Jo et al.,
2012). The buoyancy force generated due to the density differ-
The friction factor through a circular metal wire screen can be ence within the core is given as
estimated as  
FB;core ¼ rf  rpool gAf Lf (14)


2
A A1
Kscreen ¼ 1:3 1  0 þ 1 (8)
A1 A0 The friction force generated within the core is given as


1
where A0 and A1 are the open area in the screen and upstream flow FF;core ¼ rf v2f Kenter þ f Lf=D þ Kexit Af (15)
2 h
area, respectively.
The cladding outer surface temperature Tw,o is determined as The first and third terms in Eq. (15) account for the entrance and
exit losses, while the second term is the pressure drop due to the
q00 friction through the coolant channel. A conservative value for the
Tw;o ¼ Tb þ (9)
h entrance loss coefficient Kenter is 0.5, and the exit loss coefficient
The convective heat transfer coefficient h in Eq. (9) can be Kexit is 1.0. The resultant buoyancy force and friction force used to
determined as determine the coolant velocity under a natural circulation also
include forces generated through structures such as the upper
kth guide structure, plenum, PCS pipe, screen, and flap valve. The force
h ¼ Nu (10) balance can then be expressed as
Dh

FB;core þ FB;ugs ¼ FF;core þ FF;ugs þ FF;plenum þ FF;PCS pipe þ FF;screen þ FF;expansion þ FF;contraction (16)
|fflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflffl} |fflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl}
¼ FB ¼ FF

where kth is the thermal conductivity of the fluid, Dh is the hy- Since the buoyancy force is proportional to the power and the
draulic diameter, and Nu is the Nusselt number. friction force is proportional to the coolant velocity, the balance
The cladding inner surface temperature Tw,i is estimated by between the buoyancy and friction force at a given power is
solving one-dimensional heat conduction equation with no heat determined by changing the coolant velocity through the core.
generation as Until the buoyancy force and friction force are balanced, the coolant
velocity is iteratively estimated with a guessing value. If FB > FF,


thc then the new guessing value is larger than the old one. If FB < FF,
Tw;i ¼ Tw;o þ q00 (11) then the new guessing value is smaller than the old one. The
kc
 D. Jo et al. / Progress in Nuclear Energy 71 (2014) 39e51 43

Nu ¼ 0:915Gz0:4 for Gz  40
(18)
Nu ¼ 4:0 for 16 < Gz < 40
Coolant

Coolant
For upward flows, Sudo et al. (1985b) proposed a convective heat
transfer correlation given as
Fuel
Cladding

Cladding
Nu ¼ 2:0Gz0:3 for Gz  40
(19)
Nu ¼ 6:0 for 16 < Gz < 40
Temperature
where Gz ¼ Re Pr Dh =z and z is the distance in the axial direction.
The thermal margins such as a minimum ONB temperature
Tf +ΔToxide margin and DNBR need to be evaluated for the reactor safety and an
Tf optimum thermal hydraulic core design. The IAEA guidebook on
research reactor conversions (IAEA, 1980) recommends Bergles-
Tw,i+ΔToxide Rohsenow correlation (1964) for the determination of an ONB
Tw,i heat flux in plate type research reactor coolant channels as follows,

Tw,o q00ONB ¼ 1082P 1:156 ½1:8ðTONB  Tsat ÞP0:0234

2:16
(20)

Tb where P is in bar, q00 is in W/m2, and T is in
C.

Eq. (20) is rearranged for the wall temperature at which the ONB
occurs, which is given as

 P0:0234
5 q00ONB 2:16
TONB ¼ Tsat þ (21)
Fig. 3. Temperature profile in a fuel plate. 9 1082P 1:156
To determine the ONB temperature from Eq. (21), the ONB heat
pressure drops through the average channel and hot channel flux is needed. Unfortunately, both the ONB temperature and heat
should be the same. Therefore, the coolant velocity at the inlet of flux are unknown. Therefore, a single-phase temperature change
the hot channel can be iteratively evaluated by matching the against heat flux, given as Eq. (22), is used as an auxiliary equation
pressure drops. to determine the ONB temperature.
Once the coolant velocity through the core is determined, the
q00ONB
coolant temperature and pressure are also estimated using Eqs. (1) TONB ¼ Tb þ (22)
and (3). Other parameters such as the pressure drop as well as wall h
and fuel centerline temperatures are estimated in the same manner As shown in Fig. 4, the single phase heat transfer curve does not
described in Section 2.2.1. exist where the surface temperature exceeds the saturation tem-
perature Tsat. The single phase heat transfer curve can be extended
from a to a0 in order to obtain a graphical solution of Eq. (21) (Collier
2.3. Heat transfer package
and Thome, 1996). The surface temperature at the intersection of
the single phase heat transfer curve and fully developed boiling
Since a single-phase heat transfer is strongly dependent on the
curve is the ONB temperature. The ONB temperature margin is the
flow conditions and geometry, a different single phase heat transfer
temperature difference between the ONB temperature and wall
correlation may be applied to each flow region. Nusselt number for
temperature.
rectangular channels have been empirically developed by Colburn
(1933), Dittus and Boelter (1930), Sieder and Tate (1936), and
Sudo et al. (1985b). In general, DittuseBoelter heat transfer corre-
Surface heat flux

Fully developed
lation (1930) is widely and extensively used for many applications, Single-phase liquid Partial Subcooled
subcooled
Non-Boiling region boiling region
and was developed for turbulent flows (10,000 < Re < 12,000 and boiling region
0.7 < Pr < 120). In 1985, Sudo et al. developed single phase forced
convection heat transfer correlations in narrow vertical rectangular
channel with a gap of 2.25 mm, width of 50 mm, and length of
750 mm. This development considered both laminar and turbulent
flow regions with upward and downward flow for the plate type
fuel element of the research reactor, JRR-3, to be upgraded from 10
to 20 MW thermal power. Sudo et al. (1985b) compared predictions a’
by DittuseBoelter, SiedereTate, and Colburn correlations with their
Single-phase heat
experimental data, and concluded that DittuseBoelter correlation Fully developed transfer curve
shown in Eq. (17) is still applicable to upward flows for Re larger boiling curve (Eq.22)
than 4000, and applicable to downward flows for Re larger than (Eq.21) a
3000.

Nu ¼ 0:023Re0:8 Pr 0:4 (17)

Tsat TONB Surface temperature
For downward flows, Sudo et al. (1985b) proposed a convective
heat transfer correlation given as Fig. 4. Convective boiling curve.
44 D. Jo et al. / Progress in Nuclear Energy 71 (2014) 39e51

For reactor design and safety analysis purposes, CHF is poten- 2.4. Hot channel factors
tially a limiting thermalehydraulic design constraint. CHF correla-
tions applicable to plate type research reactors with low pressure Since a number of uncertainties arise concerning variations in
and low temperature conditions are very limited. A set of CHF flow distribution, fuel and coolant channel geometry, power and
correlations for narrow vertical rectangular flow channels was flow measurements, and physics power calculation, hot channel
proposed by Sudo et al. (1985a), which was used for the core factors are introduced for conservatism in designing reactors in
thermalehydraulic design and safety analyses of the upgraded JRR- thermal hydraulic analyses. Also, the most unfavorable conditions
3. They found that different CHF correlations should be applied to (hottest spot) in the core occur with hot channel factors. Therefore,
upward and downward flows. Kaminaga et al. (1998) improved the the selection of hot channel factors for thermal hydraulic analyses
early CHF correlations with more experimental data. CHF correla- that may have a significant impact on the thermal margins is
tions developed by Kaminaga et al. (1998) are given as important.
0:611
 In general, hot channel factors are applied as four separate
* 5000 *
q00*
CHF;1 ¼ 0:005 G 1 þ * DTsub;out (23) components corresponding to bulk temperature rise Fb, heat flux
G
Fq, heat transfer to the coolant Fh, and film temperature rise Ff. Fb is
the uncertainty in the bulk coolant temperature rise from the

Af * reactor inlet to the local elevation of concern, Ff is the uncertainty in
q00*
CHF;2 ¼
*
DTsub;in G (24)
Ah the local film temperature rise at the location of concern on the fuel
plate surface, Fq is the uncertainty in heat flux at the local fuel plate
pffiffiffiffiffiffiffiffiffiffi   surface of concern, and Fh is the uncertainty in heat transfer coef-
Af W=l ficient at the location of concern on the fuel plate surface. Fq is a
q00*
CHF;3 ¼ 0:7 1 þ 3 DT *
(25)
Ah   
1=4 2 sub;in
multiplier of the heat flux, and Fh is a divisor of the nominal value of
1 þ rg =rf
h as a smaller value of film coefficient h will result in larger film
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi temperature rise at the fuel plate surface.
where q00* CHF ¼ qCHF =ðhfg lðrf  rg Þrg g Þ, G ¼ G=ð lðrf  rg Þrg g Þ,
00 *
All hot channel factors are 1.0 for a best-estimate analysis and
DTsub
* ¼ ðCp DTsub Þ=hfg , l ¼ ðs=ððrf  rg ÞgÞÞ1=2 , Af is the flow area, should be larger than 1.0 to include uncertainties in the limiting
and Ah is the heated area. analysis. The hot channel factors are generated by a random error
G*1, G*2, and G*3 shown in Fig. 5 are the region boundaries for high, combining method, by a systematic error combining method, or by
medium, and low mass fluxes. A different set of CHF correlations is a combination of the two methods. The systematic error combining
applied to upward and downward flows. For downward flows, the method is a simple multiplication of uncertainties and assumes
criterion to select the appropriate correlation is described in Eq. that all of uncertainties occur at the same time and location. The
(26). For upward flows, the criterion to select the appropriate cor- random error combining method is a statistical combination of the
relation is described in Eq. (27). DNBR is defined as the ratio of the uncertainties and assumes all of uncertainties do not occur at the
DNB to the maximum fuel plate heat flux. same time and location. In reality, some of the uncertainties may
For downward flows happen at the same time and location, but not all of them. There-
  fore, a combination of the two methods is appropriate.
If G*  G*1 ; then q00CHF ¼ min q00CHF;1 ; q00CHF;2 Table 4 lists the uncertainties of the design parameters and the
  (26) selection of hot channel factors (Jo et al., 2011). The uncertainty of
If G* < G*1 ; then q00CHF ¼ max q00CHF;2 ; q00CHF;3
the power distribution by nuclear physics calculation is 10%. The
uncertainties of the fuel homogeneity and loading are 20% and 3%,
For upward flows
respectively. The uncertainty of the channel spacing is 12%. The
  uncertainty of the flow distribution between channels is 10%. The
If G*  G*1 ; then q00CHF ¼ min q00CHF;1 ; q00CHF;2 measurement uncertainties of the power and flow are 5% and 2%,
  (27)
If G* < G*1 ; then q00CHF ¼ max q00CHF;1 ; q00CHF;3 respectively. The uncertainty of the heat transfer correlation is 20%.
Fuel meat thickness, fuel geometry and loading, channel spacing,
and flow distribution are treated as random error combining pa-
where G*1 ¼ ð0:005=ððAf =Ah ÞDTsub;in
* ÞÞ1=0:389 .
rameters. Power and flow measurements, and the heat transfer
For G < G1, DTsub;out in Eq. (23) should be set to zero (Sudo and
* * *
coefficient are as treated as systematic error combining parameters
Kaminaga, 1993).
(Woodruff, 1997). With the considerations of random and

Upward flow Table 4

Downward flow Uncertainties of design parameters and selection of hot channel factors.

Uncertainty Tolerance fraction Fq Fh Fb Ff

Fuel meat thickness 0.08 1.000 e e 1.000

e e
qCHF”

U-235 homogeneity 0.12 1.120 1.120

qCHF,1"
U-235 loading per plate 0.03 1.030 e 1.030 1.030
Channel spacing 0.12 e 1.043 1.214 1.043
qCHF,2" Flow distribution 0.10 1.100 1.079 1.100 1.079

Random error combined 1.159 1.090 1.259 1.183

qCHF,3"
Power measurement 0.05 1.050 e 1.050 1.050
Flow measurement 0.02 e 1.016 1.020 1.016
Low mass flux Medium mass flux High mass flux Heat transfer coefficient 0.20 e 1.200 e 1.200

Systematic error combined 1.050 1.219 1.071 1.280

G2* G3* G1*
Product of random and 1.217 1.329 1.348 1.514
systematic errors
Fig. 5. CHF correlations for upward and downward flows.
 D. Jo et al. / Progress in Nuclear Energy 71 (2014) 39e51 45

systematic errors, the engineering hot channel factors can be The operational conditions include the average heat flux, mass
simply determined as product of both errors as listed in Table 4: flux, mass flow rate, inlet pressure and temperature, flow direction,
Fq ¼ 1.217, Fh ¼ 1.329, Fb ¼ 1.348, Ff ¼ 1.514. total power, etc. Regarding the average heat flux, the actual heat
The bulk temperature rise DTb in the hot channel is calculated in flux along the axial direction is generated as
terms of Fb given as
Q q00i ¼ Pn;i  q00avg
DTb ¼ Fb (28) (36)
_
Cp m
where Q is heat input into the channel, including a nuclear power The actual heat flux generated by Eq. (36) is once more corrected
peaking factor, Cp is heat capacity of the coolant. with the total power. For the hot channel, the radial power peaking
The peak heat flux q00max can be calculated using Fq as given as factor is multiplied with the actual heat flux. Geometry conditions
include the fuel and channel size (height, width, and thickness),
q00max ¼ Fq FR FZ q00avg (29) total number of assemblies and fuel plate per assembly. Regarding
where q00avg is the core average heat flux, FR is radial power peaking the channel height (Hch), the actual location in the axial direction is
factor, and FZ is axial power peaking factor. The nuclear power generated as
peaking factor is the product of FR and FZ.
The maximum film temperature rise can be given using the hot
zi ¼ zn;i  Hch (37)
channel factor Fh and Ff as
Other preliminary calculations used to estimate the axial node
Fh q00max Fh Fq FR FZ q00avg Ff FR FZ q00avg size and number, hydraulic diameter, heated area, fluid flowing
DTw ¼ ¼ ¼ (30)
h h h area, radial peaking factor, and so on are performed for the prep-
arations of the main processor.
where DTw is the maximum film temperature rise and h is the heat
The flow direction FDIR is checked before Cell 2 is performed;
transfer coefficient based upon the nominal channel dimensions
the flow direction is downward for forced cooling, and upward for
and nominal coolant velocity.
natural cooling. If the flow direction is downward (FDIR ¼ 1), the
The local value of the coolant temperature Tb,hc with hot channel
coolant temperature, pressure, and velocity through the core are
factors applied is determined as
calculated by marching from the inlet to the outlet. After that,
pressure drops for other structures such as the upper guide struc-
Tb;hc ¼ ðTb  Tin ÞFb þ Tin (31)
ture, plenum, and PCS pipe are estimated. Last, thermal margins
The local value of the fuel plate surface temperature Tw,hc with such as coolant, wall, and centerline temperatures, ONB tempera-
hot channel factors applied is determined as ture margin, and DNBR with respect to the hot channel factors are
estimated.
Tw;hc ¼ ðTw  Tb ÞFf þ Tb;hc ¼ ðTw  Tb ÞFf þ ðTb  Tin ÞFb þ Tin If the flow direction is upward for natural convection (FDIR ¼ 1),
(32) first, some of the input parameters such as the mass flux, inlet
The local value of the fuel centerline temperature Tf,hc with hot pressure, and total power are reset for average channels. Second, the
channel factors applied is determined as coolant, wall, and centerline temperatures and pressure through
  the core are calculated using a guessing value of the coolant velocity
Tf;hc ¼ Tw;hc þ Tf  Tw (33) at the inlet. Third, pressure drops generated by other structures
are calculated. If the buoyancy force due to the density difference
The local value of DNBR with hot channel factors applied is deter- and the friction force due to the pressure losses are not balanced, a
mined as new guessing value of the coolant velocity is set as follows:
For the first iteration (j ¼ 1),
q00CHF
DNBR ¼ (34)
q00 Fq vjþ1 ¼ 0:95  vj for FF > FB
(38)
vjþ1 ¼ 1:05  vj for FF < FB

3. Code structure For an iteration number larger than 1 (j > 1),

The steady state thermal hydraulic margin analysis code for vj1  vj
plate type fuel assembly TMAP-1.0 is written in MATLAB version vjþ1 ¼ vj  ðF  FF Þj (39)
ðFB  FF Þj1  ðFB  FF Þj B
7.2.0.232. The structure of the code, shown in Fig. 6, consists of
three cells: Cell 1 is the pre-processer to import inputs from a pre- The pressure, temperature, and velocity through the core and
named Excel file (“input.xls”), Cell 2 is the main-processer to structure calculations for natural circulation are iteratively per-
calculate thermal hydraulics through the core and calculate the formed until the buoyancy and friction forces are balanced. If
pressure drop across structures for forced convention (FC) and balanced, then the coolant velocity estimated is used as the inlet
natural convection (NC), and Cell 3 is the post-processer to plot and condition for the hot channel. For the hot channel, the coolant
export the results to an Excel file. velocity, temperature, and pressure along the axial direction are
First, Cell 1 reads the weighted power distribution (Pn,i) and calculated using Eqs. (1)e(3). Last, the thermal margins are calcu-
normalized axial location (zn,i) and the operational, geometry, and lated with and without hot channel factors applied in Cell 2.
material conditions. The summation of the weighted power with In Cell 3, the results estimated from Cell 2 are saved as an Excel
respect to the normalized axial distance should be equal to 1.0; file, and the coolant, wall, centerline, and ONB temperatures with
P and without hot channel factors applied are plotted as a function of
Pn;i Dzn;i
i
the axial location. For a natural convection, the buoyancy and
P ¼ 1:0 (35) friction forces against the iteration number are plotted to show the
Dzn;i
i convergence.
46 D. Jo et al. / Progress in Nuclear Energy 71 (2014) 39e51

Start

Input.xls Read weighted power and

(Sheet2) normalized axial location Reset inputs for natural
circulation NC
Input.xls Read operational conditions
(Sheet1) and geometry information Guess a value of the coolant
velocity
Input.xls Heating and correlation
(Sheet1) option Set inlet conditions and heat
flux for average channel
Unit conversions to SI
Solve for coolant temperature
Preliminary calculations for
heat flux, axial node size,
peaking factor, hydraulic Calculate Nu and f using
Cell1 diameter, etc correlations

Solve for wall and fuel

Yes temperatures, pressure
FDIR =0 through the core
No
No
Cell2
Location = oulet
FC Set inlet conditions
Yes
Solve for coolant temperature Solve for pressure drops
and velocity through the structures

Calculate Nu and f using Calculate friction and

correlations buoyancy forces

Solve for wall and fuel Reset a

temperatures, pressure No
(FF-FB) < DELTA new
through the core guess
Yes
Set the coolant velocity
No
Location = oulet estimated as the inlet
condition
Yes
Solve for pressure drops Solve for coolant temperature
through the structures

Calculate Nu and f using

Calculate thermal margins correlations

Solve for wall and fuel

temperatures, pressure
through the core
Output.xls Print results

Yes No
Location = oulet
Plot results
Cell3

Stop

Fig. 6. Flowchart diagram of TMAP.

4. Code verification The verifications of the functions used in TMAP-1.0 were simply
done to compare the values obtained by the function run and the
TMAP-1.0 uses MATLAB built-in and user-defined functions to exact values (Jo et al., 2011). Also, TMAP-1.0 performs iterations to
read the input, interpolate the data, estimate the water and steam evaluate the ONB temperature, average channel inlet velocity for
properties, and calculate the friction factors and heat transfer natural convection cooling, and hot channel inlet velocity for nat-
related values. TMAP-1.0 uses a steam table that uses formulations ural convection cooling. Tables 5e7 show the convergences of the
approved by IAPWS in 1995 for general and scientific use. The iterations. The stop criterion of 0.002 was applied to all iterations.
formulations are thermodynamically rigorous, accurate, and The iteration to evaluate the ONB temperature stopped at 5 steps by
applicable over a wide range of conditions (IAPWS, 1995). achieving the difference less than the stop criterion. The iterations
 D. Jo et al. / Progress in Nuclear Energy 71 (2014) 39e51 47

Table 7
Table 5
Hot channel inlet velocity iteration for natural convection (at 200 kW).
ONB temperature iteration (at 5 MW).
Iteration Vin [m/s] abs(DPhot  DPaverage)
Iteration Qonb [W/m2] Tonb by Eq. (21) [
C] Tonb by Eq. (22) [
C] Difference
[
C] 1 0.076163 33.16257
2 0.114244 25.54031
0 971428.26 115.7393 47.1338 68.6255
3 0.095204 2.68984
1 1110863.39 125.5867 115.7393 9.8474
4 0.104724 11.59382
2 1119916.19 126.2260 125.5867 0.6393
5 0.099964 4.50450
3 1120482.49 126.2660 126.2260 0.0400
6 0.097584 0.92200
4 1120517.84 126.2685 126.2660 0.0025
7 0.096394 0.88005
5 1120520.04 126.2687 126.2685 0.0002
8 0.096989 0.02192
9 0.096691 0.42883
10 0.096840 0.20340
11 0.096914 0.09072
to calculate average and hot channel inlet velocities for natural
12 0.096951 0.03440
convection stopped at 4 and 15 steps, respectively. For the average 13 0.096970 0.00624
channel, the frictional and buoyant forces were compared until the 14 0.096979 0.00784
difference is less than the stop criterion. For the hot channel, the 15 0.096975 0.00080
pressure drops between the average and hot channels were
compared.
heated from both sides (fuel plates). The amount of heat released
5. Analysis results from each fuel plate, which is deposited into a single coolant
channel, is half of the heat generated by a fuel plate. Changes in
5.1. Downward forced convection coolant temperature, velocity, and local pressure along the axial
direction are estimated by solving full three-dimensional mass,
5.1.1. Sudo et al. (1984) experiment momentum, and energy transport equations. The geometry, flow
In 1984, Sudo et al. performed experimental studies of forced and heating conditions described as a reference in Table 8 are used
convection heat transfer characteristics of a rectangular channel for for this simulation. Fig. 8 shows the comparisons of MATRA and
upward and downward flows. Experiments were carried out in TMAP-1.0 predictions. The coolant temperature and velocity and
which the flow velocity was changed from 0.07 to 7 m/s for a flow local pressure change along the axial direction predicted by TMAP-
channel thickness of 2.25 mm. The flow channel was made of two 1.0 agree well with those by MATRA analyses.
flat heating plates (750 mm long, 1 mm thick, and 40 mm wide) and
the axial heat flux distribution was uniform. Their experimental 5.2. Natural convection
studies reported the coolant and wall temperature, channel thick-
ness, flow velocity, Nu and Re numbers, and so on along the axial 5.2.1. Comparison with RELAP/MOD3
direction. RELAP5/MOD3 model originally developed to support the
In this section, four cases are reviewed and used for a compar- analysis of postulated accidents for reactor safety is used to analyze
ison study against the predictions by TMAP-1.0. In Table 8, the the thermal hydraulic characteristics for natural convection.
conditions of the reference case are typical operational conditions Recently, RELAP5/MOD3 was used to simulate NUR research reactor
of open pool type research reactors. The four cases are chosen ac- behavior under natural convection and was assessed against
cording to the mass flux, average heat flux, inlet temperature, and experimental data at a power of 100 kW (Azzoune et al., 2010). NUR
flow direction similar to the reference. research reactor is an open pool type research reactor utilized with
Fig. 7 shows the coolant and wall temperatures along the axial plate type fuel elements, which is almost identical to the reactor
direction. The temperatures predicted by TMAP-1.0 are indicated as described in Fig. 1. A RELAP5/MOD3 model used to analyze the
solid lines, and the temperatures measured by Sudo et al. (1984) are cooling capacity of the plate type research reactor during natural
indicated as dots. The predictions by TMAP-1.0 are higher than the convective cooling, equivalent to the reactor configuration shown
measured values, except for the measurements near the exit. in Fig. 1, was developed by Park (2011) for safety analyses of a
Because the uncertainty of the heat transfer coefficient is consid-
ered as 20% to determine the engineering hot channel factors listed
in Table 4, the predictions sit within the uncertainty. Table 8
Comparison of conditions.

5.1.2. Comparison with MATRA Reference Case1 Case2 Case3 Case4

The results of the thermal hydraulic analysis by TMAP-1.0 are Power per each 12.83 9.49 18.21 9.79 29.75
also compared with the results obtained by a subchannel analysis plate [kW]
code MATRA. Even though MATRA is a computer code for sub- Mass flux 2326.3 884.0 2405.5 3666.2 5173.3
[kg/(m2 s)]
channel analyses, it simulates a single channel with a hot spot.
Avg. heat flux 164 158.27 303.52 163.17 495.87
An equivalent MATRA model can be modeled as a single channel [kW/m2]
Inlet temperature 38.7 12.8 21.3 20.2 34.3
[
C]
Table 6 Flow direction Downward Downward Downward Downward Downward
Average channel inlet velocity iteration for natural convection (at 200 kW). Heated length 640 750 750 750 750
[mm]
Iteration Ffriction Fbuoyancy abs(Fbuoyancy  Ffriction) Inlet velocity
Heated width 62.1 40 40 40 40
0 12.81873 6.13312 6.68561 0.10073 [mm]
1 11.82327 6.48045 5.34282 0.09570 Channel width 66.6 50 50 50 50
2 8.21261 8.35798 0.14537 0.07566 [mm]
3 8.30101 8.29451 0.00650 0.07619 Channel thickness 2.35 2.25 2.25 2.25 2.25
4 8.29722 8.29721 0.00001 0.07616 [mm]
48 D. Jo et al. / Progress in Nuclear Energy 71 (2014) 39e51

Fig. 7. Coolant and wall temperature distributions: (a) Case 1, (b) Case 2, (c) Case 3, and (d) Case 4.

research reactor utilized with plate type fuel assemblies. The node At 50 kW, the coolant velocity through the core predicted by
diagram of the developed RELAP5/MOD3 model is illustrated in TMAP-1.0 is 0.038 m/s, while it is 0.039 m/s by RELAP. The tem-
Fig. 9. The axial and radial power shapes used in this simulation are perature rise from the inlet to the outlet of the average channel is
the same as the input of TMAP-1.0. The elements 210, 220, and 230 5.42
C by TMAP-1.0 and is 4.66
C by RELAP. With the same given
represent the hot channel, average channel, and bypass channel, power of 50 kW, the higher coolant velocity results in a lower
respectively. The elements 200, 280, and 290 represent the upper coolant temperature rise. The coolant velocities through the core
guide structure, upper plenum, and lower plenum, respectively. At predicted by TMAP-1.0 are slightly lower than those by RELAP. As
t ¼ 0 s, the PCS pumps stop. Then, first the pressure difference the power increases, the difference in the coolant velocity between
between the pool pressure (130) and the PCS pipe pressure (312) at TMAP-1.0 and RELAP increases. This may be caused by the friction
the same elevation decreases. Second the flap valves (313 and 315) factors used in TMAP analysis conservatively set at the inlet and
are open to generate the flow path for the natural circulation. Last, outlet of the fuel assembly. The thermal margins estimated by
the flow direction through the hot and average channel is reversed TMAP-1.0 listed in Table 10 provide a maximum allowable power
from downward to upward. This transition of the flow reversal is cooled by natural circulation of approximately 0.27 MW where the
completed within several tens of seconds. minimum ONB temperature margin is close to 3.0
C.
Fig. 10 shows a comparison of the coolant velocity through the
core as a function of power at 3600 s after the PCS pumps stop. As 6. Discussions
expected, the coolant velocity increases with the power. However,
the coolant velocity through the core generated by natural circu- A computer code, TMAP-1.0, was developed to analyze the
lation depends on the flow path. The upward flow path generated thermal hydraulic characteristics and thermal margins for steady
by the temperature difference shown in Fig. 9 is from the core (210) state forced convective and natural circulation cooling in research
to the pool (140), and the downward flow path is from the pool reactors facilitated with plate type fuel assemblies. Upper guide
(140) to the pools at a lower elevation. Hence, the pool (150) do not structure, plenum, and flap valves were modeled to generate a flow
contribute to the natural circulation flow path generated by the path during natural convective cooling. Since the fuel cooling
temperature difference between the core and pool. To be equivalent channels in a plate type fuel assembly are independent such that
to TMAP-1.0 model, the results shown in Table 9 consider the there is no cross flow between channels, a one-dimensional anal-
natural circulation flow path from the core (210) to the top of the ysis is suitable to predict the thermal hydraulic characteristics
pool (150) since the hotter coolant from the core flows toward the though the fuel cooling channels. Two different types of fuel
pool surface until the temperature difference is negligible. cooling channels, inner and outer coolant channels, were modeled
 D. Jo et al. / Progress in Nuclear Energy 71 (2014) 39e51 49

150

140

200 313
130 312
202 315

P
2 2 2 2 2 2
1 2 3 4 5 6 191 C
0 0 0 0 0 0 S

280

290 304

Fig. 9. RELAP model for natural circulation.

The prediction resulting from the present analyses was

compared with experimental data taken by Sudo et al. (1984) and
predictions by a subchannel analysis code MATRA for forced
convective cooling. The wall temperatures resulted from TMAP-1.0
analyses were slightly higher than those measured by Sudo et al.
(1984). The coolant temperature, local pressure, and coolant ve-
locity along the axial direction predicted by TMAP-1.0 were almost
identical to those resulting from MATRA analyses. For natural
convective cooling, coolant velocities resulting from the balance of
buoyancy and friction forces predicted by the present analyses
were compared with the predictions by RELAP5/MOD3. At 50 kW,
the coolant velocity through the core predicted by TMAP-1.0 was
0.038 m/s, while it was 0.039 m/s by RELAP. As the reactor power
increased, the difference in the velocities increased. However, the
difference was less than 4%. As a result of the comparison studies,
TMAP-1.0 code developed in the present study can be used to
analyze thermal hydraulic characteristics and to evaluate the

Fig. 8. Distributions along the axial location: (a) coolant temperature, (b) local pres-
sure, and (c) coolant velocity.

to represent the core. The inner channel, which has fuel plates on
both sides, is modeled as a coolant channel heated from both sides.
The outer channel, which has a fuel plate on one side and a side
plate the other, is modeled as a coolant channel heated from one
side. The most unfavorable condition (hottest spot) in the core was
simulated by introducing radial and axial power peaking factors as
well as engineering hot channel factors. Thermal margins such as
ONB temperature margin and DNBR were estimated for reactor
safety and design purposes. Furthermore, coolant, cladding outer
surface, and fuel centerline temperatures were estimated along the
axial direction with and without engineering hot channel factors
applied. Fig. 10. Velocities through the core vs. power.
50 D. Jo et al. / Progress in Nuclear Energy 71 (2014) 39e51

Table 9 Greeks
Comparison of the coolant velocity vs. power. m viscosity, kg/m/s
Power [MW] Coolant velocity through the core r density, kg/m3
By TMAP By RELAP
h bubble detachment parameter

0.01 (10 kW) 0.018 m/s 0.018 m/s

0.05 (50 kW) 0.038 m/s 0.039 m/s Subscripts
0.10 (100 kW) 0.051 m/s 0.053 m/s B buoyancy
0.20 (200 kW) 0.070 m/s 0.072 m/s
b bulk
0.30 (300 kW) 0.083 m/s 0.086 m/s
0.40 (400 kW) 0.094 m/s 0.098 m/s
CHF critical heat flux
0.50 (500 kW) 0.104 m/s 0.108 m/s c cladding
d downstream
F friction
f flow or fuel
g gas
Table 10 h heated or hydraulic
Mass flow rate and thermal margins vs. power.
hc hot channel factor
Power [MW] Mass flow Minimum ONB margin [
C] Minimum DNBR [e] i index for axial location
rate [kg/s] in inlet
0.01 (10 kW) 1.096 72.19 106.02 j index for iteration
0.05 (50 kW) 2.305 58.29 33.41 ONB onset of nucleate boiling
0.10 (100 kW) 3.147 44.46 20.20
oxide oxide layer
0.20 (200 kW) 4.276 19.89 12.18
0.25 (250 kW) 4.715 8.35 10.34 out outlet
0.27 (270 kW) 4.876 3.80 9.78 pool pool
0.28 (280 kW) 4.954 1.55 9.52 R radial
sat saturation
sub subcooling
u upstream
thermal margins for research reactors with plate type fuel w,i inner wall
assemblies. w,o outer wall
z axial

Nomenclature Appendix A
A area, m2
Cc contraction coefficient General transport equation derived from continuum mechanics
Cf friction factor constant can be expressed as the following differential form:
Cp heat capacity at constant pressure, kJ/kg/K
v4
D diameter, m þ V$ð4!
v Þ ¼ V$J þ 4_ g (A.1)
vt
F force, N, or hot channel factor
Fb hot channel factor of bulk temperature rise The first term on the left hand side describes time rate of change
Ff hot channel factor of film temperature rise of 4 per unit volume, the second term on the left hand side de-
Fh hot channel factor of heat transfer scribes convection of 4 by material motion per unit volume, the first
Fq hot channel factor of heat flux term on the right hand side describes influx of 4, and the second
FDIR flow direction term on the right hand side describes generation of 4 per unit
f friction factor volume. The quantities of 4, J, and 4_ g are determined for mass,
G mass flux, kg/m2/s . momentum, and energy transport equation as listed in Table A.1.
Gz Graetz number ¼ Re Pr D z
Table A.1
H channel height, m Quantities of mass, momentum, and energy equation.
h heat transfer coefficient, kW/m2/K
4 J 4_ g
hfg heat of vaporization, kJ/kg
K friction factor Mass r 0 0
k thermal conductivity, W/m/K Momentum r!v PI þ s r!g
! Q ! !
Energy rðu þ v2 =2Þ q þ $v r!v $ g þ q_
P perimeter, m
p pressure, bar
Pr Prandtl number For energy equation, Eq. (A.1) becomes as follows:
q00 heat flux, kW/m2 .      
Re Reynolds number ¼ rvD m vr u þ v2 =2
L length, m þ V$ r u þ v2 =2 !
v
vt
_ ! !
m mass flow rate, kg/s . ¼ V$ q  V$ðP v Þ  V$ðs v Þ þ r!
! ! v $ g þ q_ (A.2)
Nu Nusselt number ¼ hD k
T temperature,
C The first and second terms on the left hand side are time rate of
th thickness, m change of total energy and convection of total energy, respectively.
v velocity, m/s The terms on the right hand side are heat conduction, work done by
W channel width, m pressure, work done by viscous force, work done by gravity, and
z axial location, m heat generation, respectively. By eliminating the kinetic energy
 D. Jo et al. / Progress in Nuclear Energy 71 (2014) 39e51 51

from Eq. (A.2), then Eq. (A.2) is rearranged for the internal energy Hwang, D.H., Seo, K.W., Kwon, H., 2008. Validation of a subchannel analysis code
MATRA version 1.0. KAERI/TR-3639/2008.
given as
IAEA TECDOC-233, 1980. Research Reactor Core Conversion from the Use of High
Enriched Uranium to the Use of Low Enriched Uranium Fuels Guidebook.
vru !
þ V$ðru!
v Þ ¼ V$ q  PV$!
v  s : V!
v þ q_ (A.3) IAPWS formulation 1995 for the thermodynamic properties of ordinary water
vt substance for general and scientific use, The International Association for the
Properties of Water and Steam, Frederica, Denmark, September 1996.
Jo, D., Park, J., Chae, H., 2011. SVVR: Thermal Hydraulic and Margin Analysis Code
There is no heat generation in the fluid and the dissipation is
for Plate Type Fuel Assembly (TMAP version 1.0). KAERI technical report JR-
negligibly small compared with heat flux and pressure work. Then SW-CA-009.
Eq. (A.3) is further simplified as Jo, D., Park, S., Park, J., Chae, H., Lee, B., 2012. Cooling capacity of plate type research
reactors during the natural convective cooling mode. Progr. Nucl. Energy 56,
vru ! 37e42.
þ V$ðru!
v Þ ¼ V$ q  PV$!
v (A.4) Kaminaga, M., 1997. Steady-state thermal hydraulic analysis and flow channel
vt blockage accident analysis of JRR-3 silicide core. JAERI-Tech 97-015.
Kaminaga, M., Yamamoto, K., Sudo, Y., 1998. Improvement of critical heat flux
Using i ¼ u þ P=r and di ¼ Cp dT, the simple one-dimensional correlation for research reactors using plate-type fuel. J. Nucl. Sci. Technol. 35,
energy equation for incompressible fluids can be derived as 943e951.
Kim, W.S., Kim, Y.G., Kim, Y.J., 2002. A subchannel analysis code MATRA-LMR for

 wire wrapped LMR subassembly. Ann. Nucl. Energy 29, 303e321.
vT vT Ph q00
rCp þv ¼ (A.5) Lee, A.G., Wilkin. G.B., 1996. Results of a survey on accident and safety analysis
vt vz A codes, benchmarks, verification and validation methods. AECL-11483.
Lim, I., Park, C., Chae, H., Lee, C., 1999. Analyses of HANARO bundle experimental
data using MATRA-h: Revision. KAERI/TR-1383/99.
Finally, the one-dimensional energy balance equation for steady Park, C., Chae, H., Lee, B., Cho, Y., Jeong, H., Seo, C., 2002. Status and design char-
state conditions is expressed as acteristics of research reactor. KAERI/AR-657/2002.
Park, J., Park, D., Kim, H., Chae, H., 2008. Safety analysis process for the HANARO.
vT FNCA 2008 Workshop on Research Reactor Technology Group. October 16e20,
_ p
mC ¼ Ph q00 (A.6) 2008. Vietnam, Dalat.
vz Park, S., 2011. Analysis of Loss of Coolant Accidents. KAERI Technical Report JR-077-
TR-SA-007.
Sieder, E.N., Tate, G.E., 1936. Heat transfer and pressure drop of liquids in tubes.
Indust. Eng. Chem. 28, 1429e1435.
References Sudo, Y., Kaminaga, M., 1993. A new CHF correlation scheme proposed for vertical
rectangular channels heated from both sides in nuclear research reactors.
Azzoune, M., Mammou, L., Boulheouchat, M.H., Zidi, T., Mokeddem, M.Y., Belaid, S., J. Heat Transfer 115, 426e434.
Bousbia Salah, A., Meftah, B., Boumedien, A., 2010. NUR research reactor safety Sudo, Y., Miyata, K., Ikawa, H., Ohgawara, M., Kaminaga, M., 1984. Core heat
analysis study for long time natural convection (NC) operation mode. Nucl. Eng. transfer experiment for JRR-3 to be upgraded at 20MWt: Part1 Differences
Design 240, 823e831. between upflow and downflow for rectangular vertical flow channel. JAERI-
Bergles, A.E., Rohsenow, W.M., 1964. The determination of forced-convection sur- M 84-149.
face-boiling heat transfer. Trans. ASME J. Heat Transfer 86, 365e372. Sudo, Y., Miyata, K., Ikawa, H., Ohgawara, M., Kaminaga, M., 1985a. Experimental
Bousbia-Salah, A., Hamidouche, T., 2005. Analysis of the IAEA research reactor study of differences in DNB heat flux between upflow and downflow in vertical
benchmark problem by the RETRAC-PC code. Nucl. Eng. Design 235, 661e674. rectangular channel. J. Nucl. Sci. Technol. 22, 604e618.
Colburn, A.P., 1933. A method of correlating forced convection heat transfer data Sudo, Y., Miyata, K., Ikawa, H., Ohkawara, M., Kaminaga, M., 1985b. Experimental
and a comparison with fluid friction. Trans. AIChE J. 29, 174e210. study of differences in single-phase forced-convection heat transfer charac-
Collier, J.G., Thome, J.R., 1996. Convective Boiling and Condensation, third ed. Oxford teristics between upflow and downflow for narrow rectangular channel. J. Nucl.
University Press, USA. Technol. 22, 202e212.
Dittus, F.W., Boelter, L.M.K., 1930. Heat Transfer in Automobile Radiator of the Woodruff, W.L., 1997. Evaluation and selection of hot channel (peaking) factors for
Tubular Type, vol. 2. University of California at Berkley Publ. Eng, pp. 443e461. research reactor applications. ANL/RERTR/TM-28.
Eckert, E.R.G., Irvine, Jr., T.F., 1957. Incompressible friction factor, transition and Yoo, Y.J., Hwang, D.H., Sohn, D.S., 1999. Development of a subchannel anal-
hydrodynamic entrance length studies of ducts with triangular and rectangular ysis code MATRA applicable to PWRs and ALWRs. J. Korean Nucl. Soc. 31,
cross sections. WADC tech. Report 58-85. 314e327.