Professional Documents
Culture Documents
= dE E) (
=
dE E E N
i i
) ( ) (
(t+1) N
i
Ni,(t), Geometry,
Fig. 1. Workow of a depletion calculation with the codes of MCNP-ORIGEN-MONTEBURNS.
Table 1
Design parameters of HTR-10.
Parameter Value
Reactor thermal power (MW) 10
Primary helium pressure (MPa) 3.0
Reactor core diameter (cm) 180
Core height (cm) 197
No. of control rod channel in a side reector 10
No. of absorber ball channel in a side reector 7
Nuclear fuel type UO
2
Heavy metal loading per fuel sphere (g) 5
No. of TRISO in a pebble 8335
Enrichment of fresh fuel element (wt%) 17
No. of fuel elements in an equilibrium core 27,000
Table 2
Classication of the double heterogeneity for core modeling.
Parameter Value
Heterogeneity for
pebble
Fuel kernel radius (cm) 0.025
Carbon buffer thickness (cm) 0.009
Inner pyro-carbon thickness (cm) 0.004
Silicon carbide thickness (cm) 0.0035
Outer pyro-carbon thickness (cm) 0.004
TRISO lattice width (cm) 0.0976
No. of TRISO in pebble (#) 8335
Fuel zone (cm) 2.5
Pebble graphite shell thickness (cm) 0.5
Heterogeneity for core Filling fraction [unit-less] (reference) 0.64
(0.61)
BCC lattice width (cm) 6.724
No. of fuel:No. of moderator 57:43
Hexagonal prism height (cm) 19.5959
Hexagonal prism lattice facet to facet
(cm)
16.3923
2162 H. Jeong, S.H. Chang / Annals of Nuclear Energy 35 (2008) 21612171
theory to calculate the ux as a function of energy and position in
the cell. It also has the capability to generate burnup-dependent
microscopic cross-sections for full core depletion (Askew et al.,
1966). CITATION is designed to solve problems by using nite-dif-
ference representation of the neutron diffusion theory; it can treat
up to three space dimensions with arbitrary group-to-group scat-
tering (Fowler et al., 1971). Processing of the WIMSD output sup-
plies the input parameters required for full core modeling with
CITATION; those parameters include a few group constants for
each zone of the fuel element, the energy released per ssion pro-
cess, and the ssion spectrum. WIMSD uses the material composi-
tion in the burnup calculations; in contrast, CITATION uses the
group constants to solve the diffusion equation for the integrated
group uxes, the excess reactivity, and the power prole of the
whole core. However, because the code set of WIMSD and CITA-
TION was designed in the past, the execution condition may not
be compatible with cutting-edge technology. That code, which
uses a reduced model of a ssion product chain, explicitly treats
the buildup of major ssion products with the
135
Xe and
149
Sm
chains in the full core depletion (Mo et al., 1994), and this treat-
ment could diminish the accuracy of the ssion product inventory.
Moreover, a previous study has shown that a Monte Carlo calcula-
tion with the Monte Carlo N-Particle (MCNP) gives more accurate
results than the diffusion calculation of WIMSD-CITATION (Dalle
et al., 2003).
Third, the most promising way of the evaluation method is the
Monte Carlo calculation. The code combination of MCNP-ORIGEN-
MONTEBURNS, which is based on the Monte Carlo method, is suit-
able for a depletion calculation. This code set was chosen in this
study because it is more reliable and applicable than the previous
codes for HTGRs (Chen et al., 2004), though this code set requires
considerable time for execution. The MCNP transport code is a
three-dimensional Monte Carlo transport code that can be used
for calculations involving the transport of neutrons; photons; elec-
trons; or coupled neutrons, photons and electrons. It also has the
capability of calculating eigenvalues for critical systems (Briesmei-
ster et al., 2000). The Oak Ridge isotope generation and depletion
(ORIGEN) code is a zero-dimension isotope generation and deple-
tion code that calculates the buildup, decay, and processing of
radioactive materials (Croff, 1980). ORIGEN solves the Bateman
equation for the concentration, N
i
, of isotopes i, in a material sub-
ject to neutron irradiation. The general differential equation for
depletion can be written as
dN
i
dt
X
M
j1
l
ij
k
j
N
j
/
X
M
k1
f
ik
N
k
r
k
k
i
/ r
i
r
i
N
i
F
i
; i 1; . . . ; M;
1
where
N
i
= atom density of nuclide i;
M = number of nuclides;
l
ij
= fraction of radioactive disintegration by nuclide j which
leads to formation of nuclide i;
k
j
= decay constant of nuclide i;
/ = position and energy aver-
aged neutron ux;
f
ik
= fraction of neutron absorption by nuclide k which leads to
formation of nuclide i;
r
k
= spectrum averaged neutron absorption cross-section of
nuclide k;
r
i
= continuous removal rate of nuclide i from the system;
F
i
= continuous feed rate of nuclide i.
Fig. 2. Double-heterogeneous core modeling by MCNP. The rst heterogeneity is a pebble in gure (c), and the second one is the core in gure (b).
H. Jeong, S.H. Chang / Annals of Nuclear Energy 35 (2008) 21612171 2163
MONTEBURNS, which is a utility code that connects MCNP to
ORIGEN and calculates the depletion with the aid of a sophisticated
MCNP model, was designed to link MCNP with ORIGEN to perform
a Monte Carlo burnup calculation. The depletion calculation starts
from the reactor core modeling by MCNP. MCNP calculates a single
group of microscopic cross-sections and uxes that are used by
ORIGEN in the depletion calculations. The input of MONTEBURNS
consists of burnup condition and material composition with time.
After performing a depletion calculation, MONTEBURNS extracts
the inventory of ssion materials from the ORIGEN output le
and automatically updates the material compositions in the MCNP
input le for the next burnup step (Trellue et al., 2000). In short,
initial MCNP input generates ux and cross-section, and these out-
puts are automatically inserted into ORIGEN by MONTEBURNS.
Then the change of materials change happens with time, and this
information is updated and reected in the new MCNP input by
MONTEBURNS. The input of MONTEBURNS determines how to
connect the MCNP with ORIGEN with time. The overall workow
by this code set is explained in Fig. 1.
The code of MCNP-CINDER-MONTEBBURNS is an alternative
code because MONTEBURNS 2.0 supports CINDER, which is a
four-group, one-point depletion program for ssion products. ORI-
GEN is popularly used worldwide rather than CINDER; thus, the
combined MCNP-ORIGEN-MONTEBURNS code is used to calculate
the inventory of ssion products in this study.
2.1. Core modeling
There are many HTGRs but the core of HTR-10 in China, which is
a kind of a pebble bed reactor, was selected as a reference bench-
mark model. The IAEA previously published the benchmark prob-
lem sets (IAEA, 2003), and this open information offers sufcient
Fig. 3. Five cross-sectional views of a hexagonal core lattice. Gray only is for moderator and another is for fuel.
2164 H. Jeong, S.H. Chang / Annals of Nuclear Energy 35 (2008) 21612171
guidelines for modeling the HTGR core. This study models the core
through a more accurate approach, and deals with the core physics
benchmark problems proposed for the initial core of HTR-10.
The general design information of HTR-10 was referred to in an
IAEA report (2003) and design parameters are summarized in Table
1. A double-heterogeneous MCNP spherical model was constructed
Fig. 5. Discrete fuel management of the MONTEBURNS simulation as indicated by the dotted line in Fig. 4. The black circle is for fuel and the white circle is for the moderator.
The portion of the fuel gradually increases from (a) to (e).
0 200 400 600 800 1000
0
20
40
60
80
100
120
140
1 Stage 2 Stage 3 Stage 4 Stage T
o
t
a
l
N
o
.
o
f
f
u
e
l
p
e
b
b
l
e
p
a
s
s
e
d
t
h
e
c
o
r
e
Operation with full power day
5 Stage
70 (7)
110 (14)
118
125 (25)
No. of fresh fuel=( )
Fig. 4. HTR-10 fuel management with time (Yang et al., 2002). The y-axis means the number of fuel element loaded per EFPD during the stage. The dotted lines separate a fuel
supply interval.
H. Jeong, S.H. Chang / Annals of Nuclear Energy 35 (2008) 21612171 2165
to simulate the core. The rst heterogeneity was in a tri-isotropic
(TRISO) coated fuel particle of a fuel pebble, and the second heter-
ogeneity was implemented in the lattice of the reactor core. The
detailed information about the double heterogeneities is listed in
Table 2. A simple cubic (SC) lattice was used to distribute the indi-
vidual TRISO fuel particles in the fueled region of the pebbles, and
the basic unit of the core lattice was constructed in a hexagonal
(HEX) prism. In previous studies, the core zone was approximated
by using a body-centered cubic (BCC) lattice with moderator peb-
bles of reduced diameter; this process reproduces the specied
fuel-to-moderator pebble ratio of 57:43. The BCC and HEX struc-
tures are both implemented in this study by means of the MCNP4C
and MCNP5 code. The double-heterogeneous geometric congura-
tions of the HTR-10 model are shown in Fig. 2, and the typical
structure of lattice cells with fuel and moderator pebble is shown
in Fig. 3.
2.2. Fuel management with an increase of burnup
Although shorter burnup intervals can increase accuracy, this
benet requires additional execution time. It is common to execute
MONTEBURNS with around 5 GWD/MTU burnup steps (Chen et al.,
2004) to balance the tradeoff between accuracy and execution
time. This study used a burnup interval of around 3.7 GWD/MTU
and 23 burnup steps.
As shown in Fig. 4, Yang et al. (2002) used fuel management
for an operating rate of up to 80 GWD/MTU. For the MONTE-
BURNS simulation, the continuous fuel management of Fig. 4
is divided into the ve discrete periods shown in Fig. 5. The ini-
tial 0.43 ratio of the moderator gradually decreases in the core
and eventually, as shown in Fig. 5e, only the fuel pebbles re-
main. It seems reasonable to assume that every fuel pebble
should be burnt up at around 80 GWD/MTU at the end of the
cycle. The pebble bed reactor is designed to have a continuous
reloading scheme which enables unloaded fuel spheres that
have not reached the target burnup to be returned to the top
of the core.
2.3. Fission products of interest
The selection of an index of ssion products (FPs) when there is
an enormous number of FPs requires criteria for setting the FP of
interest. The criteria are based on a previous collection of FPs de-
ned by the government regulatory organization for an LWR. In
Korea, the collection is based on Article 86 of the Enforcement Reg-
ulations of the Atomic Energy Act. The collection includes the noble
gases of iodine and cesium isotopes as well as Br, Rb, Sr, Y, Ru, Te,
Ba, and Ce (Lee et al., 1999), and the number of isotope nuclides is
40. These FPs produce many different types of gaseous forms:
namely Xe, Kr, Ba, BaO, Sr, SrO, Ce, CeO, CeO
2
, Te, Te
2
, TeO
2
, I, I
2
,
Cs, Cs
2
, CsI, Cs
2
I
2
, Rb, and Rb
2
(Minato et al., 1994). Note that the
combined FPs are perceived as detached individual elements (for
example CsI is perceived as Cs and I).
3. Results
The data must be examined before the results can be analyzed.
The usual MONTEBURNS output fails to provide sufcient infor-
mation for the FP of interest. If the Perl script le of MONTEBURNS
is modied, most items in the FP inventory can be gathered from
the ORIGEN output. MONTEBURNS Output Extractor (MOTEX),
which is developed by the aid of Spreadsheets (Excel) and the
commercial program Visual Basic, facilitate the extraction and
processing of the outputs from MONTEBURNS and ORIGEN. MO-
TEX also helps to make out the input deck for the MONTEBURNS
simulation.
With MOTEX the output results of MCNP, ORIGEN and MONTE-
BURNS can be analyzed. The code work is veried in the following
section.
80 100 120 140 160 180 200
0.85
0.90
0.95
1.00
1.05
1.10
1.15
k
e
f
f
Height [cm]
at 300K
IAEA,2003(China) : VSOP, He
IAEA,2003(Turkey): VSOP, He
Jing,2002 : VSOP, Air
Seker,2003 : MCNP4B, He
KAIST : MCNP5, He
Fig. 6. Variation of k
eff
with fuel loading (IAEA, 2003; Seker and Colakm, 2003).
Table 3
Critical height prediction based on the two types of core lattice, HEX and BCC.
Method Experiment No. of source histories per cycle/skipped cycles/total cycles = 5000/10/60
MCNP5 MCNP5 MCNP4C MCNP4C
CPU type Intel Intel Intel AMD
OS Windows XP Redhat Linux Windows XP Windows XP
Core lattice Random HEX BCC HEX BCC HEX BCC HEX BCC
Critical height (cm) 123.06 124.44 121.73 124.92 121.96 122.65 120.15 122.14 119.95
2166 H. Jeong, S.H. Chang / Annals of Nuclear Energy 35 (2008) 21612171
3.1. Validation of MCNP modeling
The validity of MCNP modeling work can be determined by
examining the multiplication factor, k
eff
, with the core height.
How the modeling works is explained in an IAEA report of HTR-
10 benchmark problems (IAEA, 2003), and the question of valida-
tion is explained in terms of the estimation of the critical height.
A previous analysis of the HTR-10 core used a BCC lattice to model
the core. The radius of the moderator sphere must be smaller than
the original to match the actual fuel-to-moderator ratio. Reducing
a radius was simple to use but an accurate and sophisticated meth-
od was developed and used in this study. The basic unit of the
developed method for the core lattice was constructed in a hexag-
onal prism. Table 3, which compares each simple BCC and accurate
HEX method, conrms that the HEX lattice provides more accurate
results.
Fig. 6 shows a comparison of the results of the MCNP simulation
with previous studies for the same conditions. The MCNP calcula-
tions yield a slightly higher k
eff
, though the curve trend is identical.
Table 4, which lists the results of several works that focus on this
problem, shows that the current work can estimate the critical
height of the core. The HEX scheme used here has the potential
to be used for new core modeling in the future. The results in
Fig. 6 and Table 4 show that the MCNP modeling of the HTR-10
core in the current work seems reasonable.
The remarkable point about the condition of executing MCNP is
gured out. As shown in Table 5, the outputs generated from the
identical MCNP input vary in relation to the conditions of imple-
mentation; the variation is due to the calculation conditions listed
in each column of Table 5. Each version of MCNP may use a differ-
ent algorithm for generating a random number and the executing
conditions may also affect the output. It would be interesting to
see whether the following factors are the main causes of the differ-
ence in the multiplication factor: the version of MCNP, the central
processing unit (CPU), the operating system (OS) type in the com-
puter system, the number of histories per cycle and the total cycles
in the input deck of MCNP. When the number of histories was
10,000 for case 11, the k
eff
parameter was assumed to be wholly
converged since the standard deviation of the k
eff
variation was
0.00057. If the k in the case 11 is dened as the standard value,
the difference compared with the standard is indicated in the last
column in Table 5. Given that the conditions for the converged va-
lue should be the same as those in case 11, a considerable amount
of time is needed for the MCNP execution. That is the reason the
number of histories and total cycles are reduced. Thus, the pre-
ferred execution circumstances include an Intel CPU, a Windows
XP OS for accessibility, and fewer histories per cycle, a less active
cycle for minimizing execution time. According to Table 5, the cal-
culation conditions cannot be ignored as well. But, in spite of a few
errors, the results of case 5 seem plausible and, accordingly, the
conditions of case 5 can be used for all the MCNP input decks of
this work. Consequently, for any MCNP modeling in the future,
careful consideration should be given to the conditions of
implementation.
3.2. Fission product inventory with burnup
Fig. 7 shows that the reactor core maintains super-criticality
and is operated with full power. Over time, with the management
of the fuel feed and discharge (which is described in Fig. 4), k
eff
gradually declines. The occasional swing during the simulated
operation may be due to the failure of approaching the nal con-
vergence value or the lack of balance in the ratio of the fuel feed
and discharge. It would be interesting to see if the initial peak
when the burnup starts has the highest k
eff
. In terms of neutronics,
the highest peak should exist at the beginning as well, because a
real reactor is expected to operate smoothly, without a hitch.
Gathering the published inventory data for general HTGRs was a
difcult task. Over 10,000 documents were searched with key-
words for the source term and ssion products. However, only
one paper by Liu and Cao (2002) contained relevant data for
HTR-10, and one paper is insufcient for open information. The
inventories of each isotope in this work and the work by Liu are
compared in Fig. 8. The inventories were compared in relation to
the end of the cycle in the equilibrium core of HTR-10. The refer-
ence inventory by Liu and Cao was calculated solely with ORIGEN.
However, the use of ORIGEN alone can cause erroneous results due
to wrong neutron cross-sections and ux information. In fact, Liu
and Cao made a mistake to take the wrong constant neutron ux.
The ux has to vary when the composition of material changes
with time. Although the trend of the two curves in Fig. 8 seems
Table 4
Comparison on the calculation of the critical height (IAEA, 2003).
Institution Method/code Critical height (cm)
INET Experimental 123.06
INET Diffusion&Transport/VSOP94 125.80
HU Diffusion&Transport/VSOP94 119.27
NRG Diffusion&Transport/PANTHER 122.1
FZJ Monte Carlo/TRIPOLI4 117.37
MIT Monte Carlo/MCNP4B 127.5
KAIST (this work) Monte Carlo/MCNP (HEX) 123.54
Monte Carlo/MCNP (BCC) 120.95
Table 5
Differences in the value of k
eff
with various executing conditions. Initial 10 cycles are skipped.
At the calculated full loading height, 180.114 cm
Case MCNP ver. CPU/OS No. histories Total cycles k
eff
Dk
eff
1 MCNP5 Intel/WinXP 2000 40 1.13198 0.00153
2 MCNP5 Intel/Linux 2000 40 1.13048 0.00003
3 MCNP4C Intel/WinXP 2000 40 1.13242 0.00197
4 MCNP4C AMD/WinXP 2000 40 1.13546 0.00501
5 MCNP5 Intel/WinXP 5000 60 1.13140 0.00095
6 MCNP5 Intel/Linux 5000 60 1.13089 0.00044
7 MCNP4C Intel/WinXP 5000 60 1.13443 0.00398
8 MCNP4C AMD/WinXP 5000 60 1.13799 0.00754
9 MCNP5 Intel/WinXP 10,000 60 1.13004 0.00041
10 MCNP4 Intel/WinXP 10,000 60 1.13596 0.00551
11 MCNP5 Intel/WinXP 10,000 210 1.13045 Reference
12 MCNP5 Intel/Linux 5000 210 1.13084 0.00039
13 MCNP4C Intel/WinXP 5000 210 1.13547 0.00502
14 MCNP4C AMD/WinXP 5000 210 1.13794 0.00749
H. Jeong, S.H. Chang / Annals of Nuclear Energy 35 (2008) 21612171 2167
similar, the difference in the absolute magnitude of the two curves
comes from the ignorance of Liu with regard to the unique charac-
teristics of the core itself.
The activity [Bq] of all the 41 isotope nuclides over time is pre-
sented in Table 6. As mentioned above, each burnup interval was
originally around 3.7 GWD/MTU but was reclassied as 10 GWD/
MTU. It is natural for an isotope with a short half-life to produce
higher activity like that of
134
I. There is an overall increase in FPs,
according to fuel depletion and the growing fuel fraction. As bur-
nup increases, the amount of FPs becomes stable. Only a few FPs
rise steadily due to their long half-life. However, these results high-
light the need for further investigation because of the paucity of
data available for comparison. A considerable amount of research
must be conducted in the future.
It is also possible to conrm in Fig. 9 that the inventories of
actinides change as the burnup increases. There is also a rise in
the number of actinides and FPs as the core becomes lled with
fuel and the moderator is diminished. Furthermore, the current
work lays the foundation for future spent fuel analysis at the nal
point of the horizontal axis.
4. Discussion
The generation of FPs depends on the burnup, and the burnup
dependency of the neutron reaction cross-sections is explained in
Fig. 10. This MCNP simulation shows that the thermal neutron ux
per source particle at the peak decreases as the burnup increases.
Each line, square, circle and triangle means the specic burnup,
the beginning of 0 GWD/MTU, the middle of 40 GWD/MTU, and
the end of 80 GWD/MTU, respectively. The gray dotted line is for
the ssion cross-section. The peak of the neutron ux in the ther-
mal energy region is almost identical to the peak of the ssion
cross-section. As the burnup increases, the probability of ssion
in
235
U diminishes due to the lower generation of neutron ux;
thus, a higher fuel fraction is required to maintain criticality.
The neutron streaming effect in the coolant channels should
also be considered in three-dimensional diffusion calculations with
the introduction of a direction-dependent diffusion coefcient. The
inuence of neutron streaming on core reactivity needs to be quan-
tied. Such quantication could provide a useful guideline for han-
dling the streaming effect. However, the neutron streaming effect
H
3
K
r
8
3
m
K
r
8
5
K
r
8
5
m
K
r
8
7
K
r
8
8
X
e
1
3
1
m
X
e
1
3
3
X
e
1
3
3
m
X
e
1
3
5
X
e
1
3
5
m
I
1
3
1
I
1
3
2
I
1
3
3
I
1
3
4
I
1
3
5
R
b
8
8
C
s
1
3
4
C
s
1
3
7
S
r
8
9
S
r
9
0
A
g
1
1
0
m
1E12
1E13
1E14
1E15
1E16
1E17
1E18
I
n
v
e
n
t
o
r
y
[
B
q
]
Liu, 2002
KAIST
* The equilibrium core of HTR10 at the end of cycle
Fig. 8. Comparison of the inventory of 22 ssion products.
0 20 40 60 80
1.02
1.04
1.06
1.08
1.10
1.12
1.14
| | Deviation
k
e
f
f
Burn-up [GWd/MTU]
1 Stage 2 Stage 3 Stage 4 Stage 5 Stage
Fig. 7. Variation of k
eff
with increase of burnup when reactivity is not controlled.
2168 H. Jeong, S.H. Chang / Annals of Nuclear Energy 35 (2008) 21612171
was estimated to be less than 67 pcm in a previous study (Kim
et al., 2007); note that the reactivity variation was from 305 pcm
to 5823 pcm as a result of the conversion of the k
eff
values in Table
5 for changes in the execution conditions, such as the MCNP
Table 6
The inventory [Bq] of ssion products in the equivalent core of HTR-10.
Burnup [GWd/MTU] Inventory [Bq] 0 9.85 20.96 32.07 39.48 50.59 61.70 69.85 80.00
H3 0 3.31E+12 5.74E+12 6.33E+12 6.55E+12 7.03E+12 1.07E+13 1.44E+13 1.87E+13
Kr83m 0 1.56E+17 1.53E+17 1.59E+17 1.58E+17 1.57E+17 1.88E+17 1.89E+17 1.88E+17
Kr85 0 1.44E+15 2.48E+15 2.72E+15 2.81E+15 3E+15 4.53E+15 6.1E+15 7.93E+15
Kr85m 0 3.68E+17 3.61E+17 3.74E+17 3.72E+17 3.69E+17 4.41E+17 4.42E+17 4.42E+17
Kr87 0 7.17E+17 7.02E+17 7.26E+17 7.24E+17 7.16E+17 8.56E+17 8.58E+17 8.56E+17
Kr88 0 1.06E+18 1.04E+18 1.07E+18 1.07E+18 1.06E+18 1.26E+18 1.27E+18 1.26E+18
Xe131m 0 1.48E+16 1.48E+16 1.47E+16 1.48E+16 1.49E+16 1.84E+16 1.88E+16 1.87E+16
Xe133 0 2.82E+18 2.8E+18 2.87E+18 2.88E+18 2.87E+18 3.44E+18 3.48E+18 3.48E+18
Xe133m 0 8.58E+16 8.53E+16 8.82E+16 8.82E+16 8.81E+16 1.06E+17 1.07E+17 1.07E+17
Xe135 0 9.62E+17 1E+18 1.04E+18 1.07E+18 1.12E+18 1.45E+18 1.47E+18 1.48E+18
Xe135m 0 5.23E+17 5.21E+17 5.41E+17 5.41E+17 5.41E+17 6.49E+17 6.55E+17 6.56E+17
Xe138 0 2.23E+18 2.2E+18 2.28E+18 2.28E+18 2.26E+18 2.71E+18 2.73E+18 2.73E+18
Br84 0 1.54E+18 1.51E+18 1.56E+18 1.56E+18 1.54E+18 1.84E+18 1.85E+18 1.85E+18
I131 0 1.39E+18 1.39E+18 1.42E+18 1.42E+18 1.42E+18 1.71E+18 1.73E+18 1.73E+18
I132 0 2.13E+18 2.12E+18 2.19E+18 2.19E+18 2.18E+18 2.62E+18 2.64E+18 2.65E+18
I133 0 2.94E+18 2.92E+18 3.02E+18 3.02E+18 3.01E+18 3.61E+18 3.64E+18 3.64E+18
I134 0 3.44E+18 3.4E+18 3.53E+18 3.53E+18 3.52E+18 4.21E+18 4.25E+18 4.25E+18
I135 0 2.66E+18 2.64E+18 2.74E+18 2.74E+18 2.74E+18 3.28E+18 3.31E+18 3.31E+18
Rb88 0 1.05E+18 1.03E+18 1.07E+18 1.06E+18 1.05E+18 1.26E+18 1.26E+18 1.26E+18
Cs134 0 1.74E+15 5.9E+15 8.48E+15 9.4E+15 1.08E+16 1.79E+16 2.72E+16 4.19E+16
Cs136 0 1.48E+16 1.75E+16 1.85E+16 1.88E+16 1.95E+16 2.55E+16 2.92E+16 3.3E+16
Cs137 0 1.87E+16 3.26E+16 3.62E+16 3.76E+16 4.05E+16 6.16E+16 8.33E+16 1.09E+17
Sr89 0 9.25E+17 1.02E+18 9.7E+17 9.75E+17 9.78E+17 1.29E+18 1.39E+18 1.39E+18
Sr90 0 1.19E+16 2.05E+16 2.25E+16 2.32E+16 2.48E+16 3.76E+16 5.08E+16 6.64E+16
Sr91 0 1.65E+18 1.62E+18 1.68E+18 1.67E+18 1.65E+18 1.98E+18 1.98E+18 1.98E+18
Y91 0 1.15E+18 1.31E+18 1.25E+18 1.25E+18 1.26E+18 1.68E+18 1.84E+18 1.85E+18
Y91m 0 1.62E+16 1.59E+16 1.65E+16 1.64E+16 1.63E+16 1.94E+16 1.95E+16 1.95E+16
Y93 0 1.9E+18 1.87E+18 1.93E+18 1.93E+18 1.91E+18 2.29E+18 2.3E+18 2.29E+18
Ru103 0 9.21E+17 1.01E+18 9.85E+17 9.97E+17 1.01E+18 1.32E+18 1.41E+18 1.41E+18
Ru106 0 4.55E+16 7.93E+16 8.85E+16 9.32E+16 1.03E+17 1.54E+17 2E+17 2.43E+17
Ag110m 0 8.03E+11 2.65E+12 3.85E+12 4.36E+12 5.24E+12 8.67E+12 1.3E+13 1.94E+13
Te129 0 4.63E+17 4.64E+17 4.8E+17 4.8E+17 4.8E+17 5.8E+17 5.88E+17 5.88E+17
Te129m 0 5.64E+16 5.95E+16 5.8E+16 5.85E+16 5.9E+16 7.59E+16 8E+16 7.92E+16
Te131 0 1.32E+18 1.31E+18 1.36E+18 1.36E+18 1.36E+18 1.63E+18 1.64E+18 1.64E+18
Te131m 0 1.8E+17 1.81E+17 1.88E+17 1.88E+17 1.88E+17 2.26E+17 2.29E+17 2.29E+17
Te132 0 2.03E+18 2.02E+18 2.08E+18 2.08E+18 2.08E+18 2.5E+18 2.52E+18 2.52E+18
Ba140 0 2.63E+18 2.61E+18 2.64E+18 2.65E+18 2.64E+18 3.21E+18 3.26E+18 3.24E+18
La140 0 2.62E+18 2.6E+18 2.62E+18 2.63E+18 2.62E+18 3.19E+18 3.24E+18 3.23E+18
Ce141 0 2.34E+18 2.44E+18 2.37E+18 2.39E+18 2.4E+18 3.08E+18 3.23E+18 3.19E+18
Ce143 0 2.67E+18 2.64E+18 2.73E+18 2.73E+18 2.71E+18 3.25E+18 3.27E+18 3.27E+18
Ce144 0 5.65E+17 8.64E+17 8.84E+17 8.94E+17 9.28E+17 1.34E+18 1.69E+18 1.99E+18
0 20 40 60 80
10
100
1000
10000
100000
I
n
v
e
n
t
o
r
y
[
g
]
Burn-up [GWd/MTU]
ACT+FP
U235
U236
U238
Np237
Pu239
Pu240
Pu241
Fig. 9. Variation of inventory of actinides and ssion products with increase of burnup.
H. Jeong, S.H. Chang / Annals of Nuclear Energy 35 (2008) 21612171 2169
version and the OS. The neutron streaming effect on reactivity
seems less signicant than other environmental factors; therefore,
this effect can be ignored when analyzing the physics of the HTGR
core.
For the present work, although considerable effort was made to
obtain accurate and objective data, the task of predicting an exact
inventory is fraught with difculties. First, the reactivity could not
be kept constant and no consideration was given to reactivity con-
trol or neutron absorber balls. No reactivity control may have a cer-
tain effect on the generation of FPs over time. Second, cross-
sections for adjusting to the condition of real reactors need to be
provided, and they can be generated by using the NJOY code for
different temperatures with an existing library, such as the evalu-
ated nuclear data le (ENDF). However, generating cross-section is
not handled here, and the cross-sections used in this study had a
default temperature of 300 K. The temperature dependence of k
eff
at the full loading is listed in Table 7 along with the compensation.
Third, there is uncertainty over the feed-discharge rate; the contin-
ual process of fuel feeding and discharging is impossible to de-
scribe with exactitude. Fourth, as indicated in Table 5, the
conditions of execution produce another uncertainty in the results.
Consequently, any future MCNP modeling should pay attention to
the environmental conditions of execution because, as demon-
strated above, the condition effects cannot be ignored. These ve
problems need to be explored in the future but are beyond the
scope of the present paper. Further studies are needed on different
large-scale assessments.
5. Conclusion
The design of the fourth generation HTGR has not yet been def-
initely determined. This article has attempted to sketch out the
main method and provide fundamental information with regard
to the safety assessment of any new HTGR design. It should be
stressed that the evaluation of the FP inventory is the rst step
on the path to nuclear safety.
The main purpose of this article has been to explore the
inventory of FPs in a HTGR, especially a pebble bed reactor.
The study has attempted to establish a method of evaluation
by using the combined code of MCNP, ORIGEN, MONTEBURNS,
and MOTEX. As for the core modeling used by MCNP, a previous
analysis of the HTR-10 core used a BCC lattice. An accurate,
sophisticated method was developed by using a hexagonal prism
as the basic unit for the core lattice. A comparison of the results
for each simple BCC and accurate HEX method conrms that the
HEX lattice provides more accurate results. Based on the model-
ing, the generation of FPs with an increase of burnup has been
simulated. Although there are few reference studies available,
the results of the current study seem highly probable. Besides
focusing on the FP inventory, this paper also investigated the
actinides in a fuel pebble at the end of the cycle. This investiga-
tion would be informative to take into account of approaching
the spent fuel. The inventory simulation could be applied to
spent fuel management in the future, particularly as a means
of indicating the initial amount of actinides and FPs.
This article lays the foundation for future work on the safety
assessment of HTGRs, though an exact FP inventory could not be
obtained due to the uncertainty of reactor operation and fuel man-
agement, such as the feeding and discharging of fuel. If the prob-
lems raised by this approach to the FP inventory can be solved in
the future, this work will hopefully serve as a platform for the anal-
ysis of the source term of HTGRs in the future.
Acknowledgment
This paper was carried out under National Research Laboratory
(NRL) for Development of Safety Analysis Tools for Hydrogen Pro-
duction High Temperature Gas-cooled Reactors in Korea.
References
Askew, J.R., Fayers, F.J., Kemshell, P.B., 1966. A general description of the lattice code
WIMS. J. Brit. Nucl. Energy Soc. 5 (4), 564.
Briesmeister, J.F., 2000. MCNP A General Monte Carlo N-Particle Transport Code.
Version 4C, LA-13709-M.
Chen, J., 2004. On-line interrogation of pebble bed reactor fuel using passive
gamma-ray spectrometry. Doctoral dissertation, University of Cincinnati.
Croff, A.G., 1980. A Users Manual for the ORIGEN2 Computer Code, ORNL/TM-7175.
1E-9 1E-8 1E-7 1E-6 1E-5 1E-4 1E-3 0.01 0.1 1 10
0.0
5.0x10
-6
1.0x10
-5
1.5x10
-5
2.0x10
-5
2.5x10
-5
10
0
10
1
10
2
10
3
10
4
1.60e-5
1.96e-5
N
e
u
t
r
o
n
f
l
u
x
[
#
/
c
m
2
/
s
o
u
r
c
e
p
a
r
t
i
c
l
e
]
Energy [MeV]
0 GWD/t
40 GWD/t
80 GWD/t
2.27e-5
U
2
3
5
f
i
s
s
i
o
n
C
r
o
s
s
-
s
e
c
t
i
o
n
[
b
a
r
n
s
]
Fig. 10. The decrease of ssion cross-section with increase of burnup. It means the burnup dependency of FP inventory.
Table 7
The k
eff
depends on temperature at helium atmosphere (IAEA, 2003). But, there are
differences between the results by three institutions.
T (C) k
eff
INET FZJ NRG
20 1.119747 1.12665 1.11759
120 1.110435 1.11331 1.10846
250 1.095961 1.09588 1.09629
2170 H. Jeong, S.H. Chang / Annals of Nuclear Energy 35 (2008) 21612171
Dalle, H.M. et al., 2003. Benchmark Measurements and Calculations of U3Si2-Al
MTR Fuel Plates with Burned Fuel. IAEA Posters of an International Conference
(Research Reactor Utilization, Safety, Decommissioning, Fuel and Waste
Management), pp. 254269.
Fowler, T.B., Vondy, D.R, Cunningham, G.W., 1971. Nuclear Reactor Core Analysis
Code: CITATION, ORNL-TM-2496, ORNL and Union Carbide Corporation.
IAEA, 2003. Evaluation of high temperature gas cooled reactor performance:
benchmark analysis related to initial testing of the HTTR and HTR-10, TECDOC-
1382, IAEA.
Kim, K.S. et al., 2007. Development of a physics analysis procedure for the prismatic
very high temperature gas-cooled reactors. Ann. Nucl. Energy 34, 849860.
Lee, K.J. et al., 1999. Development of Source Term Evaluation Method for Korean
Next Generation Reactor IV, KINS/HR-241, KINS.
Liu, Y., Cao, J., 2002. Fission product release and its environment impact for normal
reactor operations and for relevant accidents. Nucl. Eng. Des. 218, 8190.
Minato, K. et al., 1994. Fission product behavior in Triso-coated UO
2
fuel particles. J.
Nucl. Mater. 208, 266281.
Mo, S.C., 1994. Methodology and application of the WIMS-D4M ssion product data.
International Meeting on Reduced Enrichment for Research and Test Reactors.
Morris, R.N., Petti, D.A., Powers, D.A., Boyack, B.E., 2004. TRISO-Coated Particle Fuel
Phenomenon Identication and Ranking Tables (PIRTs) for Fission Product
Transport Due to Manufacturing, Operations, and Accidents, NUREG/CR-6844,
vol. 1, US-NRC.
ORNL, 2004. How to create ORIGEN-ARP libraries, SCALE workshop, ORNL.
Seker, V., Colakm, U., 2003. HTR-10 full core rst criticality analysis with MCNP.
Nucl. Eng. Des. 222, 263270.
Soffer, L., Burson, S.B., Ferrell, C.M., Lee, R.Y., Ridgely, J.N., 1995. Accident Source
Terms for Light Water Nuclear Power Plants, NUREG-1465, US-NRC.
Stoker, C.C., Reitsma, F., Karriem, Z., 2002. Creation of the equilibrium core PBMR
ORIGEN-S cross section library. In: Proceedings of the HTR-2002 Conference on
High Temperature Reactors, Petten, NL.
Trellue, H.R., Poston, D.I., 2000. Users Manual Version 2.0 for Monteburns, Version
5B, LANL preprint LA-UR-99-4999.
World Nuclear Association, 2008. Nuclear Power Reactors. Available from: http://
www.world-nuclear.org/info/inf32.html>.
Yang, Y., Luo, Z., Jing, X., Wu, Z., 2002. Fuel management of the HTR-10 including the
equilibrium state and the running-in phase. Nucl. Eng. Des. 218, 3341.
H. Jeong, S.H. Chang / Annals of Nuclear Energy 35 (2008) 21612171 2171