Transmutation in light water reactors

Transmutation in light water reactors
Janne Wallenius
Reactor Physics, KTH
Alignment with course objectives
After todays seminar and home assignment you will be able to:
Assess nuclear aspects of reactor safety when introducing
plutonium, americium and curium into the fuel of pressurised
water reactors.
Plutonium recycling in PWRs
Where was MOX fuel first used for commercial power generation?
Which countries do presently use MOX fuels?
What is the typical plutonium fraction in MOX fuels?
Why do we need more plutonium than U-235?
What is the impact of the presence of Pu-240?
Plutonium cross sections: capture
Cross sections [b]
105
104
The spectrum averaged cross section for
fission of U-235 in a core with UOX fuel
~ 100 barn.
sc (240Pu)
sa (10B)
103
The spectrum averaged cross section for
fission of Pu-239 in a core with MOX fuel
~ 40 barn
102
101
sf (235U)
Boron worth is reduced
0
10
E [eV]
0.1
1
10
100
Shut-down margin is reduced.
Safety coefficients in UOX & MOX fuel
Coefficient @ BOC
UOX
MOX
Moderator temperature αM [pcm/K]
­−17
­−33
Fuel Doppler αD [pcm/K]
-2.5
-3.4
Boron concentration αB [pcm/ppm]
-7.0
-2.5
MOX fuel responds better to over-power transients than UOX fuel.
Shut down (hot to cold state transition) increases reactivity more than in UOX cores
Exercise: Calculate increase in reactivity when going from hot to cold state in a PWR
with UOX/MOX fuel. How much boron must be injected into the coolant to achieve 1000
pcm sub-criticality, assuming no scram?
Average coolant temperature decrease: 20 K
Average fuel temperature decrease : 300 K
Xenon worth: 1600 pcm
Reactivity management in PWRs
The typical k-infinity of a fresh UOX assembly is ~ 1.3. Why?
How does one compensate for burnup reactivity loss?
How does the reactivity loss of a MOX assembly compare?
What is the typical k-infinity of a fresh MOX assembly?
What is the total reactivity loss of a MOX assembly for 5% burnup?
PWR reactivity management
k-infinity
MOX fuel with 9% Pu recently was approved for
use in France, permitting a burnup of 50 GWd/t.
1.3
1.2
UOX (4.2%
For MOX fuel, Δ∆k ≈⋲ 5300 pcm/cycle = 2100 ppm
natural boron.
235
U)
1.1
1
Solubility limit: 2500 ppm.
MOX (9.1% Pu)
400 ppm boron injection not enough for cold
shutdown without scram!
0.9
Burnup [GWd/t]
10
20
30
40
50
Permissible MOX content in standard PWR: 30%.
100 % MOX cores
Why does mixing of UOX and MOX assemblies leads to high power
peaking at the interface?
How is that adressed in the design of MOX assemblies?
How can one re-design a reactor to permit use of 100% MOX
assemblies?
Break!
Multi-recycling of plutonium
How large is the reduction of Pu in MOX fuel at 50 GWd/t burnup?
How much of the reduction is due to conversion to higher actinides?
What is the impact on the radiotoxicity of spent MOX fuel?
How does the plutonium vector change during burnup?
If one would recycle this plutonium, what is the impact on reactivity?
How does one compensate for this?
Impact on safety
1.0
Fission probability
sf
sf + s c
0.8
Multirecyling leads to increased fraction of
fertile plutonium.
0.6
0.4
Increase of Pu concentration at BOC
necessary to manage reactivity
0.2
E [eV]
1
10
2
10
3
4
10
10
5
10
6
10
Increase in coolant void worth
Void worth [pcm]
More than12% Pu: positive void worth
10000
5000
No more than 2 recycles feasible with
standard MOX design
0
5000
How can we address this issue?
10000
Recycle #
1
2
3
4
CORAIL assembly for multirecycling of Pu
A CORAIL fuel bundle consists of 84 MOX
pins and 180 UOX pins.
Void worth remains negative during
multirecycling of Pu in existing PWRs.
Pu-burnup in MOX pins slightly exceeds
production in UOX pins.
80% of power park has to operate on
CORAIL assemblies to stabilise Pu
inventory.
9% Pu
4.9% 235U Thimble CR-tube
25% of the plutonium is converted to
americium.
MOX fuel with support of enriched uranium:
MOX-UE
By using enriched uranium for fabrication of the MOX fuel, the
degradation of neutron spectrum becomes less severe.
A MOX fuel with 3-4% 235U and 12% Pu meets shutdown criteria for
multi-recycled plutonium.
30% of fuel MOX-UE assemblies in the power park sufficient to
stabilise plutonium inventory.
All fuel pins must be fabricated in MOX fuel factories
Recycling of Am in PWRs
Fission probability in a PWR
Fission probability of major Am nuclides
virtually equal to zero in PWRs
237
Np
238
Pu
More neutrons are emitted in fast fission
of Am (3.3 prompt neutrons) than in
fission of Pu.
239
Pu
240
Pu
241
Pu
242
Pu
Void worth increases rapidly with
americium concentration
241
Am
243
Am
244
Cm
245
Cm
0.2
0.4
0.6
0.8
1
Still, Pu and Am inventories may be
stabilised by use of MOX-UE assemblies
with 1% Am, if 40% of the power park
uses this fuel.
Transmutation of americium
242mAm
~70% of 241Am is converted to 242Cm
15%
n + 241Am
242Cm
85%
83%
~15% becomes 242mAm
Into which nuclides do 242Cm & 242mAm
decay?
242gAm
17%
242
Pu
242Pu –> 243Am -> 244Cm -> 245Cm
Recycling of Cm in PWRs
Multi-recycling of Cm in PWRs will increase equilibrium
concentration of 252Cf by 2-3 orders of magnitude!
Neutron source of fresh fuel increases similarly.
Cm should be recycled in fast neutron spectra
Decay heat production in 244Cm ~ 3 W/g
Neutron production in
244Cm ~ 107 n/s/g
Storage of separated Cm would require forced cooling for
~100 years!
If one does not separate Cm, the decay heat in a single BWR
assembly (at equilibrium) approaches 6-7 kW, which is on
the limit for management in air.
What shall we do?
What did we learn so far?
Assignment!
Simulation of transmutation of plutonium in a PWR
Use Serpent to calculate cross sections for transmutation (capture and fission) in a
simplified pin cell model of a PWR with either MOX fuel or MOX-UE fuel.
Write a mathematica file that simulates three irradiation campaigns of the MOX fuel,
including reprocessing where mixing with fresh UOX assemblies is made.
Calculate the Doppler coefficient, coolant temperature coefficient and coolant void worth
at beginning of first cycle as well as at end of the third campaign.
Some boundary conditions are given in the text book. Take special care when calculating
the new Pu composition when mixing spent MOX with spent UOX assemblies.
Use geometry from NEA:s MOX benchmark
Explain why the calculated safety parameters change with burnup
Mixing MOX with UOX
Mixing MOX with UOX is not that trivial!
Recall:
Spent UOX fuel contains 1.2% plutonium
Fresh MOX fuel contains ~ 9-12% plutonium
How many UOX assemblies are required for producing 1 MOX assembly?
Spent MOX fuel will contain ~ 7-9% plutonium (25% reduction of Pu inventory)
How many spent UOX fuel assemblies must be added to one spent MOX assembly in order
to produce a ”2nd recycle” MOX assembly?
What is the fraction of ”UOX-Pu” relative to the ”MOX-Pu” in the 2nd recycle MOX fuel?
Monte Carlo in non-multiplying media
Starts by simulating a birth by sampling the neutron energy (hence velocity)
from the fission (or delayed neutron) spectrum
Samples the direction
Calculates the macroscopic cross section ∑tot for scattering in a material with
constant composition in a region of space
Probability of not colliding witin distance d: Pnc(d) = Exp[-∑tot d]
Probability of collision: Pc(d) =1 - Exp[-∑tot d]
Calculate the probability of absorbtion versus scattering: Pabs= ∑abs/∑tot
Reduce the ”weight” of the neutron by Pabs
Sample new energy and direction of neutron
Monte Carlo in multiplying media
In multiplying media, an absorption can lead to emission of more than one
neutron
E.g. fission (2-10 neutrons) & (n,2n) reactions
At the absorption event, the probability for fission is Pf = σf/σabs
The number of fission neutrons ν is sampled from a probability distribution
The simulation may start ν neutrons with a weight change of Pf
and restart
Statistical analysis provides a statistical estimate for the probability of fission
neutrons to induce a new fission ~ keff!
Monte Carlo error estimates
Error estimates converge as the square root of the sample size!
In order to reduce your error estimate by a factor of ten, 100 more histories
have to be sampled!
Nuclear Monte Carlo codes typically provide 1σ error estimates for keff.
One out of three error bars will not cover the converged result!!
Fitting keff as function of fuel temperature with a 2nd order polynomial, one
may expect the curve to go through 2/3 of the 1σ error bars.
SERPENT
Monte Carlo based burnup code
Written by Jaakko Leppänen at VTT (Finland)
3-4 times faster than MCNP for the same statistical accuracy
Requires > 10 times more memory, typically ~ 1GB
KTH license, you may install and use the code on your laptop as
long as you are KTH student. Cross section data require > 5 GB of
free storage
Manual and other information about the code:
http://montecarlo.vtt.fi
Serpent input data: geometry
% PWR pin cell input
set title "PWR pin lattice with UO2 fuel"
pin 1
fuel 0.4095
% fuel outer radius
clad 0.4750
% cladding outer radius
water
% coolant outside of clad
% Geometry
surf 1 sqc 0.0 0.0 0.665
% square pin cell surface
cell 1 0 fill
% square pin cell filled with pin 1
1
-1
cell 2 0 outside
1
Serpent input data: materials
mat fuel -10.0
92235.09c 0.037
92238.09c 0.963
8016.09c 2.000
%
%
%
%
mat clad -6.56
40000.06c -0.984
50000.06c -0.016
% Zircalloy cladding density
% Composition including alloying with tin
mat water -0.757
1001.06c
2.0
8016.06c
1.0
% Water moderator density in PWR
mat boron
5010.06c
5011.06c
% Natural boron
1.0
0.2
0.8
Uranium oxide fuel
U-235 (3.7% enrichment)
U-238
Two oxygen atoms per metal atom
set abs boron -600e-6 water % Mix 600 ppm boron absorber with water
Serpent input data: Execution
set
set
set
set
acelib "jeff31.xsdata"
bc 3
pop 1000 100 20
power 180
%
%
%
%
Cross section library pointer
Reflecting boundary condition
Run 1000 neutrons in 100 cycles
Power normalised to 180 W/cm
Serpent input data: Analysis
% Definition of "detectors" to register fluxes and reaction rates
det 1
dm fuel
% Detector number 1
% Flux in the fuel
det
dm
dr
dr
dr
dr
dr
2
fuel
18 Pu238
18 Pu239
18 Pu240
18 Pu241
18 Pu242
% Detector number 2
% Flux in the fuel
% folded with fission reaction rates
det 3
dm fuel
dr 102 Pu238
dr 102 Pu239
dr 102 Pu240
dr 102 Pu241
dr 102 Pu242
% Detector number 3
% Flux in the fuel
% folded with capture reaction rates
Output &
calculation of microscopic cross section
General output in file INPUT_res.m
Detector output in file INPUT_det0.m
microscopic cross section = reaction rate / flux
Typical flux in LWR?
Typical flux in FR?