Upload
others
View
5
Download
0
Embed Size (px)
Citation preview
1Wonder 2009, Cadarache Sept. 29th - Oct. 2nd
Unresolved Resonance Range Cross Section, Probability Tables
and Self Shielding Factors
J-Ch. Sublet *, R.N. Blomquist, S. Goluoglu, R.E. MacFarlane
CEA , DEN, Cadarache, FranceArgonne National Laboratory, USA
Oak Ridge National Laboratory, USALos Alamos National Laboratory, USA
*UKAEA Fusion, Culham Science centre, Abingdon, United [email protected]
2Wonder 2009, Cadarache Sept. 29th - Oct. 2nd
PT’s impact on the ICSBEP benchmarks KeffCode Tripoli-4.5 Tripoli-4.5 Tripoli-4.5Library JEFF-3.1 JEFF-3.1.1
Experiment Calc. Calc. Calc.Kef f Unc. Kcalc S.D. Kcalc Δ (C-C) Δ TP
ICSBEP Name
IMF-007 Cyl. U Metal (10% 235U), thick 238U ReflectorBig Ten deta. 1.0045 70 0.99863 13 0.99878 0.99406
simp. 1.0045 70 0.99790 13 0.99770 0.99329Δ (C-E) -623 -626 -2 -1082 -456
t.z.h. 0.9948 130 0.98830 13 0.98838 0.98432Δ (C-E) -650 -642 8 -1048 -406
IMF-012 Cyl. U Metal (16% 235U), Al and Steel, Reflected by Depleted-UZPR(16%) c-1 1.0007 270 1.00261 13 1.00262 0.99968
Δ (C-E) 191 192 1 -102 -294IMF-10 Cyl U Metal (9% 235U), thick Depleted U Reflector ZPR-U9 c-1 0.9954 240 0.99181 13 0.99191 0.98631
Δ (C-E) -359 -349 10 -909 -560IMF-002 Nat. U Reflected Assembly of Enriched U Plates
c-1 1.0000 300 0.99216 10 0.99207 0.99231Δ (C-E) -784 -793 -9 -769 24
IMF-001 Bare Cyl. Conf. of Enriched and Natural UJemima c-2 1.0000 120 0.99837 12 0.99797 0.99868
c-3 1.0000 100 0.99741 12 0.99779 0.99847c-4 1.0000 100 0.99850 12 0.99821 0.99921
Average 0.99809 0.99799 0.99879Δ (C-E) -191 -201 -10 -121 80
JEFF-3.1.1-ur TP
Fast Range
Excellent way Excellent way to test the to test the influenceinfluenceof the URRof the URR
PT’s off
3Wonder 2009, Cadarache Sept. 29th - Oct. 2nd
PT’s impact on the ICSBEP benchmarks
• Probability table treatment in the URR is important
• Probability table impact the URR range and below
• Thermal benchmarking are not impacted whatever the fuel mixture and enrichment: LCT’s, Thermal solutions, etc…
• Fast benchmarking are impacted most, but only if U-238 and Pu239, Pu240 are present in large quantities, not with U-235
– U238 20-150 Kev, U235 2.25-25 Kev– Pu239 2.5-30 Kev, Pu240 5.7-40 Kev
• There is nothing to shield in the U-235 ur, this is not the case for many other isotopes, including absorbers and structural materials
4Wonder 2009, Cadarache Sept. 29th - Oct. 2nd
PT’s impact on the ICSBEP benchmarksWARNING: TRIPOLI through its usage of CALENDF PT’s
account for fluctuations for the U-238 inelastic “competition”
INT = 2wiggle
All otherINT= 22
Interpolation law
+2% in the ur-4% below
5Wonder 2009, Cadarache Sept. 29th - Oct. 2nd
PT’s impact on the ICSBEP benchmarks
MCNP .884520 .121459 0.99406 TART .884691 .121389 0.99402 TRIPOLI .884657 .121288 0.99409
With PT’s
Inelastic ur MF-2 Γx, CALENDF waysimpact = 50 pcm with TRIPOLI
Without PT’s
Measurement Absorption Leakeage 0.996+/-.003COG .876983 .125390 0.99763 MCNP .876444 .125880 0.99768 TART .876536 .125633 0.99783 TRIPOLI .875432 .126260 0.99831 VIM .876523 .125853 0.99763
MCNP .884520 .121459 0.99406 TART .884691 .121389 0.99402 TRIPOLI .884657 .121288 0.99409
PT’s impact380 pcm
6Wonder 2009, Cadarache Sept. 29th - Oct. 2nd
“Lubets” challenge, Lubitz - Sublet
UR parameters interpretations and formalism sensitivity
First, CALENDF had to be modify in order to allow a “user choice” of the formalism used in the unresolved range
Second, U-235, but also the U-238 and Pu-239 of ENDF/B-VII fission spectra have been modified to emit only, in respectively the unresolved resonance range of each evaluation:
Unresolved range 2.5e+3 – 2.5e+4 U-235Unresolved range 2.0e+4 – 1.5e+5 U-238Unresolved range 2.5e+3 – 3.0e+4 Pu239
7Wonder 2009, Cadarache Sept. 29th - Oct. 2nd
“Lubets” challengeMNBW = MNBW CALENDF interpreted PT’sSLBW = SLBW CALENDF interpreted PT’sRM = RM CALENDF interpreted PT’sinfd = no PT’s, smooth NJOY unresr infinitely dilute xs
A single geometry is used, a 30 cm sphere filled independently with two types of materials: H2O (ENDF/B-VII) + one isotope or the isotope alone. With 100 millions (1.0x108) histories and the Red- 616 group structure, the neutron source is also been sample in a flat spectra within the unresolved range.
Water sphere with U-235 Solid sphere of U-235Water sphere with Pu-239 Solid sphere of Pu239Water sphere with U-238 Solid sphere of U-238
Sodium sphere with U-238Carbon sphere with U-238
8Wonder 2009, Cadarache Sept. 29th - Oct. 2nd
Water sphere, U-235, Pu-239, U238
No impact
9Wonder 2009, Cadarache Sept. 29th - Oct. 2nd
Carbon and Sodium sphere , U-238
No impact
10Wonder 2009, Cadarache Sept. 29th - Oct. 2nd
Solid sphere, U-235, Pu-239, U238
Noticeable impact with U-238 and
Pu239
URR
11Wonder 2009, Cadarache Sept. 29th - Oct. 2nd
URR Benchmarking interpretation
With water as coolant, U-235, U-238 or Pu239 as fuel, it does not matter which formalism is thrown in the unresolved range, its effect, impact is negligible.
With Sodium or Carbon as moderator, U-238 as fuel, the interpretation formalism in the unresolved range do not matter either.
The slowing down of neutron within solid fuel is marginally influenced by the formalism use in the unresolved range. The more susceptible isotope been Pu-239, and may be other..
The slowing down of neutron within solid fuel is significantly influenced by a probability table treatment compared with none. Here again Pu-239 is most affected although U-238 and U-235 cannot be considered as insensible.
12Wonder 2009, Cadarache Sept. 29th - Oct. 2nd
ENDF-102 URR formalism
-Only the SLBW LRF=1 formalism for ur parameters is allowed
-Single level Breit-Wigner; no resonance-resonance interference one single-channel inelastic competitive reaction allowed
-LSSF flagLSSF=0, MF-3 contains partial background to be addedLSSF=1, MF-3 contains infd xs, MF-2 is used solely forthe calculation of self-shielding factor ssf
• InconsistencyΓx can be given if LRF=2, in U238 accounted for by CALENDF but notby NJOY, PREPRO, PURM or AUROX that read in the infd xs in MF-3disregarding the competition widths in MF-2.
13Wonder 2009, Cadarache Sept. 29th - Oct. 2nd
Direct cross section contributions- LSSF=1
One need to be sure that the direct inelastic (that is not accounted for in MF-2 assembly) has been added to the compound one
Direct and compound angular distribution are very different, direct is forward peaked while compound is more symmetrical
With LSSF=1 the “direct” component can be added as “background”, but the MF-4 (angular distribution) is the same !!
For Pu239 : direct component18.1% at 3.00E+05
TALYS U-238 Discrete Inelastic cross section - Level 1, Spin= 2.0 Parity= +, Direct component contributions to the total level Energy Direct 5.00E+04 0.80%5.50E+04 0.97%6.00E+04 1.15%8.00E+04 1.87%1.00E+05 2.55%1.20E+05 3.19%1.40E+05 3.74%1.60E+05 4.24%1.80E+05 4.69%2.00E+05 5.08%2.04E+05 5.15%2.06E+05 5.19%2.08E+05 5.22%2.10E+05 5.26%2.16E+05 5.36%2.20E+05 5.42%2.22E+05 5.46%2.24E+05 5.49%2.26E+05 5.52%2.30E+05 5.58%2.50E+05 5.87%2.60E+05 6.01%2.70E+05 6.14%2.80E+05 6.27%2.90E+05 6.39%3.00E+05 6.51%
14Wonder 2009, Cadarache Sept. 29th - Oct. 2nd
U-235 ur, nothing to self-shield !! new NJOY graphs
PTPT’’s derived SSF s derived SSF
Not Not BondarenkoBondarenko
15Wonder 2009, Cadarache Sept. 29th - Oct. 2nd
Pu-240 ur, PT derived ssf impact, new graphs
SSFSSF’’ss are are channelschannelsdependentdependentMT=1,2,102MT=1,2,10218 and 4 18 and 4 (4 only for (4 only for CALENDF PTCALENDF PT’’s) s)
16Wonder 2009, Cadarache Sept. 29th - Oct. 2nd
Unresolved resonance range computation
Evaluation UR range (eV) Points/decade INT LSSF Shape W-184 2.65E+03- 1.00E+05 3 / 2.5 5 lin-lin U-233 6.00E+02- 4.00E+04 17 2 1 rough U-238 2.00E+04- 1.49E+05 18 5 1 lin-lin, Gx Pu-238 2.00E+02- 1.00E+04 constant Pu-239 2.50E+03- 3.00E+04 48 2 rough Pu-240 5.70E+03- 4.00E+04 constant
New in NJOY-296 (not -259)PREPRO since many years
- Parameter against cross sectioninterpolation impact- INT = 5 not accounted forAUROX or PURM
17Wonder 2009, Cadarache Sept. 29th - Oct. 2nd
CALENDF PT’s U-238 self-shielded cross-section
Inelastic 7% (unique..)Capture 14%Elastic 11%
18Wonder 2009, Cadarache Sept. 29th - Oct. 2nd
CALENDF pointwise data – W-184
Peak and depress
Different SSF
Statistically generated resonances
19Wonder 2009, Cadarache Sept. 29th - Oct. 2nd
Effective cross sections comparisons
• Four different processing codes; NJOY-PURR, CALENDF, AUROX and PURM on the six evaluations
• Those isotopes have been chosen to encompass most cases encountered in the unresolved energy range of any ENDF/B-VII or JEFF-3.1.1 evaluations.
• The minimum energy encountered for an unresolved range is 1 eV and the maximum is 1 MeV.
• What was asked of the participants were all unresolved range cross-sections in the unresolved range of those six ENDF/B-VII evaluations, at 293.6 Kelvin, both infinitely dilute, and 1 barns, in simple 2E11.4 column format. However simple this may seem, it took quite a few iterations to finalize the series of graphs in this section
20Wonder 2009, Cadarache Sept. 29th - Oct. 2nd
Pu238 UR cross sections, INT = 5
21Wonder 2009, Cadarache Sept. 29th - Oct. 2nd
Pu239 UR cross sections, AUROX corrected
22Wonder 2009, Cadarache Sept. 29th - Oct. 2nd
Pu240 UR cross sections, PURM corrected
23Wonder 2009, Cadarache Sept. 29th - Oct. 2nd
U233 UR cross sections, 20% at 1 barn, LSSF=1
24Wonder 2009, Cadarache Sept. 29th - Oct. 2nd
U-238 UR cross sections, 1 to 2% at 1 barn, LSSF=1
25Wonder 2009, Cadarache Sept. 29th - Oct. 2nd
W-184 UR cross sections, 20% at 1 barn
26Wonder 2009, Cadarache Sept. 29th - Oct. 2nd
Conclusions
• The primary conclusion is that these four processing codes usually agree within a target accuracy of 1% for both infinitelydilute and 1-barn self-shielded effective cross sections (with the exception of PURM) when the parameters files data and ENDF-102 rules have been properly and consistentlyinterpreted by both the evaluators and those who processed the data. In our comparison, this only occurred on one fourth (1/4)of the evaluations.
• A secondary conclusion derives from the fact that processing codes have to palliate the data format deficiencies, either because the format rules have not been well defined, have been interpreted differently, or are inconsistent or unphysical.
• Check what your transport code is using, do not assume, what you see is NOT what you get, infinitely dilute versus shielded.
27Wonder 2009, Cadarache Sept. 29th - Oct. 2nd
Conclusions
The ENDF-102 rules should be revised (as evolutions not major modifications) along the following lines:
• CALENDF’s way: a format need to be defined (ENDF-102) but different specifications may also be applied• Allow for other formalisms in the UR : MLBW, RML.• Account for the effect of multiple fission channels.• Allow all competition channels to be open in this range, e.g., inelastic levels, direct components, charge particle emissions. This would make the sum-up and energy interpolation rules clear, but would not require that everybody could or should use them all.• Privileges, enforces LSSF=1 formalism (self-shielding from file 2, cross section in file 3) if you can be sure that the SSF’s can be correctly predicted else LSSF=0• Make the formats and specifications unambiguous.
28Wonder 2009, Cadarache Sept. 29th - Oct. 2nd
Conclusions
• CALENDF can already use SLBW, MLBW and RM in the URR on the actual evaluation
• CALENDF evolution: interpretation of intermediate structures in fission cross sections
• If in the U-238 evaluation one changes LSSF=1 to 0, does it recover File 3 cross sections, every time ??? If not what shouldbe done…
• Evaluator should comments clearly what they have done, or not done on the smooth cross section provided in MF-3 and MF-2 parameters (ΔE=100 ev histogram in the UR of Pu-239 MF-2, not clearly defined)
JCS says many thanks to Pierre Ribon