REACTOR PHYSICS CALCULATIONS ON MOX FUEL

IN BOILING WATER REACTORS (BWRs)

Christophe Demazière
Chalmers University of Technology
Department of Reactor Physics
SE-412 96 Gothenburg
Sweden

Abstract

The loading of MOX (Mixed Oxide) fuel in BWRs (Boiling Water Reactors) is considered in this
paper in a “once-through” strategy. The fuel assemblies are of the General Electric 8×8 type, whereas
the reactor is of the General Electric BWR/6 type. Comparisons with traditional UOX (Uranium
Oxide) fuel assemblies revealed that the loading of MOX fuel in BWRs is possible, but this type of
fuel creates new problems that have to be addressed in further detail. The major ones are the SDM
(Shutdown Margin) and the stability of the cores at BOC (beginning of cycle), which were
demonstrated to be significantly lowered. The former requires a new design of the control rods,
whereas a modification of the Pu isotopic vector allows improving the latter. Another issue with the
use of the MOX fuel assemblies in a “once-through” strategy is the increased radiotoxicity of the
discharged fuel assemblies, which is much higher than of the UOX fuel assemblies.

801
Introduction

In the past few years, the possibility of the use of MOX fuel in commercial nuclear power stations
has been demonstrated in some western countries. Whereas MOX fuel has been loaded mostly in
PWRs (pressurised water reactors), such a type of fuel is considered to be used also in BWRs (boiling
water reactors).

This paper investigates the possibility of loading MOX fuel assemblies in BWRs, both in a partial
or full MOX loading (the purpose of this paper is not to compare the possibility of loading MOX fuel
assemblies in BWRs to the loading of such fuel assemblies in PWRs). These fuel assemblies are
assumed to be used in a “once-through” strategy, i.e. the discharged MOX fuel assemblies are
considered as nuclear wastes. The multirecycling of Pu is out of the scope of this paper. Several
aspects are thus studied: the fuel inventory (major/minor actinides, and fission products), the fuel
performance in terms of Pu utilisation, the reactivity coefficients (both at the assembly and the core
level), the point-kinetic parameters and the stability of the core. For that purpose, a generic model of a
General Electric BWR/6 has been chosen. The fuel assemblies are all of the General Electric 8×8 type,
but three different lattices are presented in this paper: a UOX lattice used as a reference lattice, and a
UOX and a MOX lattice used for the mixed UOX/MOX loading (the full MOX loading was carried
out with the same type of MOX fuel assemblies).

In the first section of this paper, the assemblies and the core are described, and so are the
hypotheses used for the assembly/core calculations. In the second section, an analysis of the fuel
assemblies is carried out, and their main properties are summarised. In the third section, the core
characteristics are analysed.

Description of the fuel assemblies and the core for the study

In this section, the fuel assemblies and the reactor, which were chosen for this study, are
presented. The hypotheses used to perform the assembly and core calculations are also emphasised.

Fuel assemblies

In order to study the ability of BWR cores to be loaded with MOX assemblies and to develop a
core equilibrium model, several types of assemblies are necessary. First of all, one needs fresh UOX
fuel, which will constitute the first loading (cycle 1).1 One also needs fresh UOX fuel that will be used
for a full UOX loading and that will be loaded according to the same strategy from cycle to cycle until
the equilibrium core is reached. These assemblies are referenced in the following as the u00 lattice.
Since the main aim of this paper is to compare the MOX cores to the UOX cores, it was decided to use
typical UOX assemblies for these aforementioned ones. They are of the General Electric 8×8 type and
are based on Ref. [1].

From this full UOX equilibrium, a MOX equilibrium must be reached, either in a partial MOX
loading or a full MOX loading. This requires defining both a UOX and a MOX assembly (referenced
in the following as the u01 and m01 lattices respectively), since the properties of one assembly depend
on its environment.2

1. The characteristics of these bundles are not touched upon in the paper.
2. The same MOX bundle was used for both the partial and the full MOX loading.

802
 For the MOX assembly, it is necessary to define the quality of plutonium, which is used in this
study. The plutonium isotopic vector corresponds to the reactor-grade plutonium (64.1% fissile) that
can be recovered from spent UOX assemblies. This quality is defined in Table 1 below.

Table 1. Nominal Plutonium isotopic vector

Pu isotope Mass fraction (%)
238
Pu 2.7
239
Pu 56.0
240
Pu 25.9
241
Pu 8.1
242
Pu 7.3

The lattices for the full UOX loading, for the mixed UOX/MOX and full MOX loadings are
presented in Figure 1. The lattice for the full UOX loading is taken from Ref. [1], and represents a
typical General Electric 8×8 UOX fuel assembly that will be considered as a reference fuel assembly
for the rest of this study. The UOX and MOX lattices used in the mixed UOX/MOX loading (and the
full MOX loading for the MOX fuel assembly) were determined via an iterative optimisation process
(see Ref. [2] for further details) so that the desired cycle length was achieved, the power peaks inside
the fuel assemblies and between the fuel assemblies (in case of a mixed UOX/MOX loading) were
avoided, there was no mismatch between the Gadolinium and the fuel depletion, and finally the
control rod efficiency was good enough.

It has to be pointed out that the fuel density has been reduced from 10,3 g/cm3 for UOX
assemblies to 9,78 g/cm3 for MOX assemblies in order to keep the optimum of the moderator to fuel
ratio.3 This fuel density allows having an under-moderated lattice at EOL (end of life).

Figure 1. Enrichment (%) of the lattices in 235U and/or Pu

(South-East corner of the control blades)

m01 lattice
u00 lattice u01 lattice
(both mixed UOX/MOX and
(full UOX loading) (mixed UOX/MOX loading)
full MOX loadings)

Gd Gd

Gd Gd

W Gd W Gd W

Gd W Gd W W

Gd Gd

Gd Gd

3. This optimum is larger for MOX fuel since the multiplication factor is higher due to 239Pu.

803
Core

The nuclear reactor is supposed to be of the General Electric BWR/6 type. The cycle
characteristics, assuming that they can be applied to each cycle in the equilibrium model, are described
in Table 2 below.

Table 2. Cycle characteristics for a general electric BWR/6

Cycle length 369 days (≈ 12.3 months)
Capacity factor 85%
314 EFPD
Cycle length at full power
(effective full power days)
Power density 50 kW/l
Core rated thermal power 3 380 MWth
Core rated flow 47.25×106 kg/hr
Heated fraction of flow 85%
Number of assemblies 764
Assembly nominal active fuel height 381 cm (cold)
Fraction of discharged assemblies ≈ 1/4
UOX assemblies
(for both the full UOX and the mixed UOX/MOX loading)
Fuel density 10.3 g/cm3
EOC core-averaged burn-up 19 061 GWd/tHM
MOX assemblies
(for both the mixed UOX/MOX and the full MOX loading)
Fuel density 9.78 g/cm3
EOC core-averaged burn-up 20 075 GWd/tHM
Number of MOX assemblies
280
(for mixed UOX/MOX loading)

In order to be able to compare the different cores (core with a full UOX loading, core with a
mixed UOX/MOX loading, core with a full MOX loading), one has to define a common “strategy” for
them, i.e. one has to apply the same rules to reload the cores. Furthermore, only the cores that have
reached equilibrium, i.e. cores that have the same properties from cycle to cycle if one applies the
same reloading pattern, can be used for these comparisons.

The loading strategy was carried out according to the Control Cell Core (CCC).4 Once the loading
pattern has been chosen, it remains to determine how the fuel will be depleted. The depletion was
carried out in accordance with the Haling principle (see Ref. [3]), which is a hypothetical depletion,
but which can be applied to any loading.5

4. In the CCC, the control of the reactor is limited to a fixed group of control rods (rods in the A2 control
cells) and low reactivity fuel assemblies surround the control rods.
5. The Haling principle states “The minimum peaking factor for a given fuel loading arrangement is achieved
by operating the reactor so that the power shape (power distribution) does not change appreciably during
the operating cycle”.

804
 Based on this Haling principle, an equilibrium model was developed for each loading.6 The only
constraint that was used in this model was to reach a target effective multiplication factor of
1.00000 (±100 pcm) at EOC (end of cycle). This was realised by modifying the reload batch size
(approximately one fourth of the core) and by readjusting the loading pattern accordingly. Optimising
the power distribution and readjusting the enrichment of the assemblies were clearly out of the scope
of this study.

Codes

The codes used to perform all the calculations were CASMO-4, TABLES-3, and SIMULATE-3
from Studsvik Scandpower (see Refs. [4-6]). The neutron transport calculations were carried out with
70 groups, and the cross-section library was based on ENDF/B-IV. The CASMO-4 models rely on the
method of the collision probabilities, the method of the response matrix, and the method of the
characteristics, whereas the SIMULATE-3 models are based on the 2-group nodal diffusion equations
(see Ref. [7]).

Fuel assembly analysis: fuel inventory/performance and reactivity coefficients

Although this paper focuses on the use of MOX fuel in BWRs, i.e. the characteristics of cores
loaded with MOX fuel, it is essential to study such fuel at the assembly level via CASMO-4. The main
properties of interest are of course the fuel inventory and the performance of the fuel assemblies. But
the reactivity coefficients can also be studied for each assembly type assuming an infinite loading
pattern constituted by identical fuel assemblies.

Regarding the fuel inventory, Figure 2 gives the mass distribution of the minor and the major
actinides, 135Cs, 137Cs, and 151Sm7 at BOL (beginning of life) and EOL, for a void content of 40%. As
can be seen on this Figure, the main difference between the UOX fuel assemblies (u00 and u01
lattices) and the MOX assembly (lattice m01) is the relatively high content of some of the minor
actinides in the MOX fuel assemblies at EOL, the concentration of which is 5 to 15 times higher than
for the UOX fuel assemblies. Concerning the U-Pu isotopes, one notices that roughly more than half
the amount of 239Pu and of 235U has been burnt in the MOX fuel compared to their initial amount. The
amount of the other Pu isotopes is larger at EOL than at BOL due to the Pu build-up by neutron
capture/radioactive decay. The MOX fuel reduces considerably the amount of 236U and 237Np
compared to the UOX fuel. Nevertheless, one notices also that the 135Cs, and 151Sm fission products
have much higher concentrations at EOL for MOX assemblies than for UOX assemblies, and their

6. An equilibrium cycle consists of establishing a loading pattern using a fresh batch of fuel (fixed number of
assemblies at a fixed enrichment) and then depleting and shuffling using the same fuel design and loading
pattern for multiple cycles. Equilibrium is reached when exposure and power distributions converge to one
solution, which does not change in succeeding cycles.
7. Regarding the radiotoxicity of the wastes, several fission products, which can be grouped into three
categories (see Ref. [8]), have to be studied. The first group is constituted by 99Tc, 129I, 93Zr, 107Pd, and
135
Cs, and represent the long-lived isotopes that can be transmuted in a fast spectrum much faster than their
natural decay (only 135Cs is accessible in CASMO-4 using the ENDF/B-IV library). 90Sr, 137Cs, and 151Sm
represent isotopes that are not worth transmuting since their natural decay is similar or much faster than
their transmutation (only 137Cs and 151Sm are accessible in CASMO-4 using the ENDF/B-IV library).
Finally, the last group (79Se, 126Sn, and 94Nb) is made of isotopes that cannot be transmuted rapidly because
of their relatively small cross-sections (none of these isotopes is available in CASMO-4 using the ENDF/B-
IV library). Fortunately, the radiotoxicity of these isotopes is rather small since their yield is limited.

805
concentration is roughly twice as high as for UOX assemblies (the amount of 137Cs is almost identical
in the UOX and MOX fuel assemblies). Regarding the radiotoxicity of these fuel assemblies, it can be
seen on Figure 3 that the radiotoxicity of the MOX fuel assembly (lattice m01) is roughly one order of
magnitude higher than the radiotoxicity of the UOX fuel assemblies (Figure 3 gives the radiotoxicity
for the u01 lattice, which is identical to the one for the u00 lattice). The main actinides contributing to
the radiotoxicity are of course the Pu isotopes,8 the concentration of which is much higher in the MOX
than in the UOX assemblies. The effect of the Am and Cm isotopes (mainly 241Am and 244Cm) is also
roughly ten times higher in the MOX than in the UOX assemblies. The radiotoxicity of 237Np,
although lower in the MOX assembly than in the UOX assembly just after discharge, becomes higher
a few hundreds of years after discharge due to the α decay of the accumulated 241Am. Taking all these
facts into consideration, the use of MOX fuel in BWRs as a means of burning the fissile Pu is possible
since the amount of 239Pu is reduced by half. Nevertheless, the radiotoxicity of such discharged fuel
assemblies is roughly ten times higher than the one of conventional UOX fuel assemblies.

Figure 2. Fuel inventory for the different lattices (calculations performed at 40% of void)

At BOL At EOL
3.5 1.4
u00 lattice u00 lattice
u01 lattice u01 lattice
m01 lattice m01 lattice
3 1.2

2.5 1
Mass fraction (%)

Mass fraction (%)

2 0.8

1.5 0.6

1 0.4

0.5 0.2

0 0
Am 2m

Am 2m
N 7
Pu 9

Am 241

C 43

N 7
Pu 9

Am 241

C 43

1
Pu 8
Pu 9
Pu 0
Pu 1
Am 42

C 42
C 43
C 44
C 45

C 6
C 5
Sm 37

Pu 8
Pu 9
Pu 0
Pu 1
Am 42

C 2
C 43
C 44
C 45

C 6
C 5
Sm 37
4
5
6

N 9

4
5
6

N 9
3
3

15

3
3

15
23
23
24
24

24

23
23
24
24

24

24

3
23
23
23
23

23
23
23
23
p2
p2

24

p2
p2

24

2
2

2
2
2
2

s1
s1

2
2
2

s1
s1
m
m
m
m
m

m
m
m
m
m
U
U
U
U

U
U
U
U

Regarding the behaviour of the fuel assemblies in case of a mixed UOX/MOX loading, power
discontinuities can occur at the boundaries between UOX and MOX assemblies because of the larger
absorption cross-section of plutonium fuel. It is thus necessary to check the properties of the UOX and
MOX segments (u01 and m01 respectively) in a checkerboard pattern, via the use of CASMO-4 (see
Table 3 and Figure 4).

8. The radiotoxicity of some of the fission products such as 90Sr is known to be actually larger than the one of
the Pu isotopes just after the fuel discharge and to decrease significantly after a few hundreds of years, but
these fission products are not present in the ENDF/B-IV library.

806
 Figure 3. Evolution of the radiotoxicity of the discharged fuel assemblies
u01 lattice m01 lattice
9 9
10 10

Total Total
8 FP 8 FP
10 10
U U
Np Np
7 Pu 7 Pu
10 Am 10 Am
Cm Cm
Radiotoxic inventory (Sv/tHM)

Radiotoxic inventory (Sv/tHM)

6 6
10 10

5 5
10 10

4 4
10 10

3 3
10 10

2 2
10 10

1 1
10 10

0 0
10 10
0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7
10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 1
Time (year) Time (year)

Table 3. Comparison between the UOX and the MOX assemblies in a checkerboard pattern
At 0 GWd/tHM At 40 GWd/tHM
UOX MOX UOX MOX
Burn-up (GWd/tHM) 0.000 0.000 38.302 41.788
νΣ f Σ a (1) 1.13827 1.11642 0.87403 0.93249
2-group k∞ (1) 1.14816 1.10338 0.87952 0.92650
2 2
2-group M (cm ) 86.11 83.77 87.69 87.84
Fission fraction (1) 1.023 0.977 0.941 1.059
Absorption fraction (1) 0.943 1.057 0.959 1.041
Nu*fission fraction (1) 0.952 1.048 0.927 1.074
Power fraction (1) 1.009 0.991 0.938 1.062

One notices that the UOX assembly contributes to most of the energy release at BOL, whereas the
contribution of the MOX assembly becomes more significant during the depletion and even larger than
the UOX one. This is mostly due to the higher macroscopic absorption cross-section in the MOX
segment at BOL compared with the UOX segment. Consequently, the depletion rate is higher in the
MOX segment; this is why the corresponding burn-up is larger at EOL. Since the infinite
multiplication factor for the MOX segment at hot conditions without control rod is larger than the one
for the UOX segment at high burn-up, the relative importance of the MOX assembly becomes larger
during the depletion. As expected also, the flux is harder in the MOX assembly because the thermal
absorption is much larger (the corresponding macroscopic absorption cross-section is larger). Finally,
the power distribution is relatively good during the depletion. The power peak for the MOX assembly
is moving from the periphery at BOL towards the centre of the assembly at EOL, since the depletion
rate is larger along the (wide) water gaps. For the UOX assembly, the power peak located along the
water gaps at BOL totally vanishes during the fuel depletion.

807
 Figure 4. Power distribution in the UOX and MOX segments in a checkerboard pattern
(40% of void)
At BOL At EOL
1.2 1.04
1.2
Gd Gd 1 Gd Gd
1 1.02
Gd Gd
1.1
W 0.8 Gd W 0.8 W Gd W 1
W W Gd W W Gd
1
0.6 Gd 0.6 Gd 0.98
Gd Gd Gd Gd
0.4 0.9
0.4 0.96
MOx UOx MOx UOx

1.2 1.04
1.2
1 Gd Gd Gd Gd
1 1.02
Gd Gd
1.1
Gd W 0.8 W 0.8 Gd W W
1
W Gd W W Gd W
1
Gd 0.6 Gd 0.98
0.6
Gd Gd Gd Gd
0.4 0.9
0.4 0.96

UOx MOx UOx MOx

Concerning the reactivity coefficients of each assembly modelled in an infinite lattice, the
Doppler and the void coefficients of reactivity are presented in Figures 5 and 6, respectively. If one
relies on the classical four-factor formula, these reactivity coefficients can also be analysed in a
qualitative manner. Regarding the fuel temperature coefficient, the most important effect is due to the
decrease of the resonance escape probability in the resonance group when the fuel temperature
increases (increase of the resonance integral). This is a negative reactivity effect. On the other hand,
there is a positive contribution of 239Pu and 241Pu for MOX fuel. This positive effect is triggered by the
larger increase of the fission rate than the capture rate when the fuel temperature rises. 240Pu
contributes also to the Doppler coefficient but in a negative manner (strongly at BOL, slightly at EOL)
to the resonance escape probability because of the increase of the capture rate solely when the fuel
temperature increases. It thus results that the Doppler coefficient is slightly more negative for MOX
assemblies than UOX assemblies. The analysis of the void coefficient is slightly more complicated
since many competing effects are involved when the void content changes. When the void content
increases, the neutron mean free path increases, the fraction of neutrons captured in the resonances
during the slowing-down increases, and so does the neutron age. All these effects have a negative
reactivity contribution. On the other hand, two main effects induce a positive contribution when the
void content rises: the decrease of the fraction of thermal neutrons captured in the moderator, and the
increase of the fast fission factor due to the increase of the flux around the fission energies. In case of
MOX fuel, the harder spectrum (higher flux in the resonance region) gives an increase in the change of
the capture rates, so that the variation of the resonance escape probability becomes more negative. For
the same reason, this spectrum hardening strengthens the changes in the fast fissions, so that the fast
fission factor becomes more positive. Nevertheless, the effect on the resonance escape probability is
larger than the effect on the fast fission factor, and consequently the void coefficient is lowered (more
negative). It thus results that the void coefficient is more negative for MOX assemblies than for UOX
assemblies (except at high void fraction, where the effect on the fast fission factor becomes larger for
MOX fuel than for UOX fuel).

808
Core analysis: reactivity coefficients and stability

In the previous section, the reactivity coefficients have been calculated for each assembly via
CASMO-4. In this Section, the reactivity coefficients are calculated for the whole core via
SIMULATE-3. The following Table 4 summarises the main core characteristics for the full UOX core,
the mixed UOX/MOX core, and the full MOX core.

In summary, one can say that the uniform Doppler coefficient is almost unchanged in case of
MOX assemblies, whereas the void coefficient (most of the MTC – Moderator Temperature
Coefficient) is more negative, as expected. The power coefficient is slightly larger (more negative),
and so is the flow coefficient (more positive). The pressure coefficient is also slightly smaller at BOC
(Beginning Of Cycle), but larger at EOC.

The most significant changes are the variation of the effective fraction of delayed neutrons (due to
239
Pu and 240Pu, 210 pcm and 490 pcm respectively, compared with 235U, 650 pcm), the variation of the
prompt neutron lifetime, and the reduction of the control rod efficiency. This reduction is so strong
that the shutdown margin becomes smaller than 1% at BOC. The high value for the thermal absorption
cross-section of the MOX fuel causes a reduction in the mean free path of thermal neutrons, which
reduces the thermal neutron absorption probability for the control rod.

It was decided, for the UOX/MOX core exclusively, to model a variation of the Pu isotopic vector
in order to study the evolution of the previous parameters. A variation of 50% of each Pu isotope
(relative to their nominal value) was carried out. A variation of the Pu-2xx isotope is referenced as mxx
and pxx, for -50% and +50% of variation respectively in the following Figures. All the absolute results
are given, and also their variation relative to the full UOX core (for the UOX/MOX core and the full
MOX core) or relative to the nominal UOX/MOX core (for a Pu isotopic vector variation). Regarding
the variation of the Pu quality, one notices that there is no significant change with the Pu isotopic
vector (only a significant reduction of 239Pu gives a decrease of the power coefficient) except for the
SDM (Shutdown Margin), as depicted in Figure 7. The SDM can become greater than 1% if one
reduces the 239Pu amount and/or one increases the 240Pu amount. This is the result of the large
absorption cross-section of 239Pu (≈1 035 barn in the thermal region), whereas the 240Pu absorption
cross-section is relatively small (≈197 barn in the thermal region). Consequently, any reduction in the
absorption cross-section tends to increase the control rod efficiency, as explained before. The
influence of the other Pu isotopes is lower, since their nominal proportion is relatively small compared
with 239Pu and 240Pu.

Another parameter that is essential to estimate in case of BWRs is the stability of the reactor. It is
well known that three main types of instability may occur in a BWR (see Refs. [9,10]): the
fundamental mode or in-phase oscillation (in which the fundamental flux is oscillating over the whole
core), the first azimuthal mode or out-of-phase oscillation (in which the first azimuthal flux is
oscillating over the whole core), and the channel instability or density wave oscillation (in which the
flux is oscillating only in a certain channel). In this study, only the first type of instability is
investigated in a qualitative manner via the so-called March-Leuba model (see Ref. [11]). This model
allows estimating the transfer function of the reactor (from reactivity perturbation to normalised
neutron density variation), which can be reduced to a second-order model in the frequency range of
interest for the in-phase type of oscillations. From this transfer function, it is possible to calculate the
Decay Ratio (DR), defined as the ratio between two consecutive maxima of the impulse response of
the normalised neutron density (the DR gives a measurement of the damping of the system). The full
derivation of the DR ratio from the March-Leuba model is given in Ref. [2].

809
 The DR was found to be sensitive to several parameters used in the March-Leuba model. The first
one is related to the fuel density, and it was shown that a lower fuel density renders the core slightly
more unstable. As pointed out previously (see Table 2), the fuel density of MOX fuel assemblies is
lower than the one of UOX fuel assemblies, and thus the reactor is expected to be less stable in case of
a mixed UOX/MOX loading and a full MOX loading. Likewise, it was demonstrated that the DR was
larger (core more unstable) when the effective fraction of delayed neutron was smaller, due to the
proximity of prompt-criticality. MOX cores are thus expected to be less stable since the effective
fraction of delayed neutrons is smaller than full UOX cores (see Table 4). Finally, the DR was proven
to be sensitive to a parameter called in Ref. [2] the amplification factor of the void response, the
parameter C. This parameter is a destabilising factor if it is larger since the void response is a large
delayed negative response that could be out of phase with the initial reactivity perturbation, and thus
could enhance it. MOX cores are thus less stable since the void coefficient is larger (less negative).
Core calculations revealed that C was very sensitive to the Pu quality (mostly 239Pu and 240Pu) at BOC.
As can be seen on Figure 8, the DR decreases significantly at BOC if one reduces the 240Pu content
or/and one increases the 239Pu content. This is due to the fact that in these cases, the void coefficient
and consequently the C parameter are less negative, thus improving the stability. This stability
improvement disappears during the cycle because of the fuel depletion.9

Figure 5. Comparison between the Doppler coefficients of reactivity for the different assemblies
At BOL At EOL
0 0
u00 u0
u01 u0
m01 m0

−0.5 −0.5

−1 −1
Doppler (pcm/K)

Doppler (pcm/K)

−1.5 −1.5

−2 −2

−2.5 −2.5

−3 −3
void=0% void=40% void=80% void=0% void=40% void=80%

9. It was also shown that the DR was dependent on the residence time of steam bubbles in the fuel channels.
This does not depend on the fuel type loaded in the core (UOX or MOX), but more on the thermal-
hydraulic design of the fuel assemblies.

810
 Figure 6. Comparison between the void coefficients of reactivity for the different assemblies
At BOL At EOL
0 0
u00 u0
u01 u0
m01 m0
−20
−20

−40

−40
Void coefficient (pcm/%)

Void coefficient (pcm/%)

−60

−60 −80

−100
−80

−120

−100
−140

−120 −160
void=0% void=40% void=80% void=0% void=40% void=80%

Table 4. Summary of the reactivity coefficients and the associated parameters

for the three types of cores

Mixed UOX/MOX
Full UOX core Full MOX core
core
BOC EOC BOC EOC BOC EOC
Moderator temperature
-55.44 -57.01 -57.69 -60.37 -64.46 -62.84
coefficient (pcm/°F)
Uniform doppler coefficient
-1.12 -1.08 -1.10 -1.15 -1.19 -1.16
(pcm/°F)
Power coefficient
-41.80 -45.81 -42.66 -47.37 -45.52 -46.85
(pcm/%power)
Flow coefficient
31.63 36.57 32.53 37.66 34.72 37.22
(pcm/%flow)
Pressure coefficient
9.32 8.43 7.96 8.82 8.78 9.00
(pcm/psia)
Effective fraction of
604 532 536 491 408 417
delayed neutrons (pcm)
Prompt neutron lifetime
39.6 41.4 34.1 35.9 24.5 26.5
(µs)
Control rod worth (pcm) 35 023 35 965 32 132 33 251 26 854 28 164
Shutdown margin (%) 2.426 3.756 0.737 3.039 0.705 2.343
Average discharge burn-up 29.9/35.6
– 28.3 – – 33.2
(GWd/tHM) UOX/MOX

811
 Figure 7. Analysis of the shutdown margin
SDM at BOC
4
Abs
Diff
2

SDM (%)
0

−2

−4
full UOxUMOxm38 p38 m39 p39 m40 p40 m41 p41 m42 p42 full MOx

SDM at EOC
6
Abs
Diff
4
SDM (%)

−2

−4
full UOxUMOxm38 p38 m39 p39 m40 p40 m41 p41 m42 p42 full MOx

Figure 8. Analysis of the decay ratio and of the amplification factor C of the void response
Decay Ratio at BOC C at BOC
4 1
Abs Abs
Diff Diff
2 0
C (1/K.m )*1E5
DR (%)

0 −1

−2 −2

−4 −3
full UOxUMOxm38 p38 m39 p39 m40 p40 m41 p41 m42 p42 full MOx full UOxUMOxm38 p38 m39 p39 m40 p40 m41 p41 m42 p42 full MOx

Decay Ratio at EOC C at EOC

4 1
Abs Abs
Diff Diff
3
0
C (1/K.m )*1E5

2
DR (%)

−1
1

−2
0

−1 −3
full UOxUMOxm38 p38 m39 p39 m40 p40 m41 p41 m42 p42 full MOx full UOxUMOxm38 p38 m39 p39 m40 p40 m41 p41 m42 p42 full MOx

The following Table 5 summarises the results of the stability calculations for the mixed
UOX/MOX core (with the nominal Pu isotopic vector) and the full MOX core. Only the comparison
with the full UOX core is presented since the simple model used in this study does not allow an
absolute determination of the Decay Ratio, i.e. only a qualitative analysis is possible.
One notices that the mixed and full MOX cores are much less stable at BOC than at EOC in
comparison with the full UOX core. This is mostly due to the fuel depletion and the production of Pu
isotopes by the fresh UOX assemblies in the full UOX cores, so that the differences in the core
characteristics die out with burn-up. Although the MOX cores (mixed UOX/MOX and full MOX
cores) are more than ten times less stable at BOC, full power, and full core flow than the full UOX
core, it is difficult to give an absolute value for the DR since the absolute DRs are too low to be
realistic. Furthermore, it is not granted either that the same ratio can be applied to any part of the
power-flow map (this study only investigated the full power and full core flow case). Nevertheless, it
is very likely that the same tendency can be noticed for other points of the power-flow map, i.e. that
the MOX cores are much less stable than the UOX cores. Therefore, other points in the power-flow
map (where the stability of the full UOX core is known to be very low) might have a DR very close to
instability or might even show limit-cycle oscillations for MOX cores.

812
 Table 5. Stability of the mixed UOX/MOX and full MOX cores (at full power and full core flow)
Ratio between the UOX/MOX core Ratio between the full MOX core
DR and the full UOX core DR DR and the full UOX core DR
(absolute DRs) (absolute DRs)
BOC 14.11 (2.54%/0.18%) 11.17 (2.01%/0.18%)
EOC 1.44 (2.73%/1.90%) 2.08 (3.95%/1.90%)

Conclusions

The main goal of this study was to determine the possibility of loading MOX fuel in BWRs. In
order to be able to compare the core characteristics, it was necessary to develop three equilibrium core
models: a full UOX core, a mixed UOX/MOX core (approximately 2/3 of UOX assemblies and 1/3 of
MOX assemblies), and a full MOX core.

The assembly/core calculations showed that it seems to be possible to load MOX assemblies in
BWRs (either in a mixed UOX/MOX pattern, or a full MOX pattern), since most of the core
characteristics are comparable with the ones of a full UOX core. Nevertheless, three main problems
arise: the shutdown margin at BOC is lower than 1%, the value requested by the safety authorities; the
cores with MOX fuel are less stable; finally and most importantly, the radiotoxicity of the discharged
MOX fuel is significantly larger than of UOX fuel.

One way of solving the two first problems could be to modify the Pu isotopic vector: a variation
of the 239Pu and/or the 240Pu content affects very much both the SDM and the DR. Unfortunately, these
variations are not compatible: if one increases the 239Pu content for instance, both the SDM and the DR
decrease, whereas the opposite behaviour occurs with 240Pu. Consequently, one needs to modify both
the Pu quality for the stability and also the control rod efficiency for the SDM, i.e. a new control rod
design is necessary. The same conclusion regarding the SDM was also drawn in Ref. [12] (where it
was demonstrated that increasing the B-10 content of the control rods allowed improving the SDM).
This may be a major obstacle to the loading of MOX fuel, unless the SDM safety limit is lowered.
Considering the SDM, the stability, and the radiotoxicity, MOX fuel assemblies in BWRs raise major
concerns that have to be solved prior to their extensive use in commercial nuclear reactors.

Acknowledgements

This study was supported by CEA Cadarache, France (CEA/DEN/DER). Vattenfall AB is

acknowledged for the financial support given for using the Studsvik Scandpower codes. Thanks are
also due to Mr. A. Mourogov for his comments regarding this paper.

REFERENCES

[1] L.J. Covington (1996), Generic CMS BWR Equilibrium Model – Revision 1. Studsvik Report
SOA-96/06, Studsvik of America.
[2] C. Demazière (2000), Analysis of the Reactivity Coefficients and the Stability of a BWR Loaded
with MOX Fuel. In: Proceedings of PHYSOR2000, ANS International Topical Meeting on
Advances in Reactor Physics and Mathematics and Computation into the Next Millennium,
May 7-12, Pittsburgh, Pennsylvania, USA.

813
[3] H.W.Jr. Graves (1979), Nuclear Fuel Management. John Wiley & Sons, United States of
America.
[4] M. Edenius, K. Ekberg and B.H. Forssén (1995), CASMO-4 – A Fuel Assembly Burn-up
Program – User’s Manual. Studsvik Report SOA-95/1, Studsvik of America.
[5] J.A. Umbarger (1995), TABLES-3 – Library Preparation Code for SIMULATE-3 – User’s
Manual. Studsvik Report SOA-95/16, Studsvik of America.
[6] J.A. Umbarger and A.S. di Giovine (1995), SIMULATE-3 – Advanced Three-Dimensional Two-
Group Reactor Analysis Code – User’s Manual. Studsvik Report SOA-95/15, Studsvik of
America.
[7] C. Demazière (1999), Boiling Water Reactors with Uranium-Plutonium Mixed Oxide Fuel –
Report III: A Survey of the Methodology of the Studsvik of America CMS Codes. Studsvik Report
SSP-99/431S, Studsvik of America.
[8] M. Salvatores, I. Slessarev and A. Tchistiakov (1998), The Transmutation of Long-lived Fission
Products by Neutron Irradiation. Nuclear Science and Engineering, 130, pp. 309-319.
[9] D. Henning, A Study on Boiling Water Reactor Stability Behaviour. Nuclear Technology, 126,
pp. 10-31.
[10] J.K. H. Karlsson (1999), Development and Application of Reactor Noise Diagnostics. PhD thesis,
Chalmers report CTH-RF-141, Chalmers University of Technology, Göteborg, Sweden.
[11] J. March-Leuba (1986), A Reduced Order Model of BWR Linear Dynamics. Nuclear Technology,
75, pp. 15-22.
[12] J.L. François, C. Martin del Campo and J. Hernández (2002), Design of an Over-moderated Fuel
and a Full MOX Core for Plutonium Consumption in Boiling Water Reactors. Annals of Nuclear
Energy, 29, pp. 1953-1965.

Nomenclature

BOC Beginning of Cycle

BOL Beginning of Life
BWR Boiling Water Reactor
CCC Control Cell Core
DR Decay Ratio
EFPD Effective Full Power Day
EOC End of Cycle
EOL End of Life
FP Fission Products
MOX Mixed Oxide
MTC Moderator Temperature Coefficient
PWR Pressurised Water Reactor
SDM Shutdown Margin
UOX Uranium Oxide

814