Mini-CSEWG-2015
→
US/Eastern
B (Berkner)
B
Berkner
Brookhaven National Laboratory
, ,
Description
Goal of the Mini-CSEWG-2015 meeting is to review evaluation work related to the next version of the ENDF/B library that was not discused during the CSEWG meeting in November 2014. Due to the focus on the last CSEWG on the CIELO evaluation work some topics important for ENDF/B were not covered. These include progress on non-CIELO evaluations, related validation, format changes, experimental activities and advancements in the covariance methodology.
Participants
Boris Pritychenko
Brad Sleaford
Bret Beck
Brian Epping
Brown David
Chuck Lee
David Heinrichs
Donald Smith
Dorothea Wiarda
Gustavo Nobre
Ian Thompson
Jeremy Conlin
Jesse Holmes
Jesse Holmes
John Bess
Luiz Leal
Mark Chadwick
Michael Dunn
Michael Rapp
Michael Zerkle
Mike Herman
Ramon Arcilla
Skip Kahler
Sonzogni Alejandro
Timothy Trumbull
Toshihiko Kawano
Yaron Danon
Support
Minutes of the mini-CSEWG Meeting
7-8 May 2015
Evaluation Session
Luiz Leal: Cu and W resonance evaluations
W isotopesWork on the RRR with Geel extends previous upper energy limit, allowing more accurate simulations involving self-shielding. A draft file has been produced and preliminary testing has been done
Cu isotopes
The Cu evaluation performs well with Leal's u235 added into JENDL, for the Zeus benchmark but we need to understand the performance in other uranium evaluation work. It appears that the copper work has changed things little itself. It does provide consistent covariance information.
Since these new evaluations enable very detailed angular distributions to be computed, we will need to define what is recommended for users regarding full use of these details in processed dat in simulations or use of averaged quantities.
Toshihiko Kawano: Summary of updated data at LANL
Ar40New inelastic scattering from the GEANIE detector combined with model calculations have informed a new evaluation.
Ni
The 58,60Ni(n,alpha) (n,p) and (n,2n) have been updated. Work has been done on the elastic scattering representation. It was found that a smoothed representation of the P1 angular shape was adequate.
Cu63,65
The (n,2n) was improved for 63Cu. For capture we see a problem in the continuity between the resonance and statistical regions. It appears that resonance work is need to understand the under prediction of capture from the resonance region. It appears that levels are missing in the resonance analysis, especially missing P waves. There are also problems above 100 keV.
Kawano also removed a large negative P1 resonance in the ORNl evaluation because of his concerns.
236mNp
A new evaluation has been submitted.
Photo nuclear reactions.
The 39K evaluation was presented.
The new actinide data being measured at TUNL could also impact our evaluations.
Other work is progressing on As74, Ne, and Re isotopes
Brad Sleaford: Capture gamma data
Capture gamma, especially for thermal reactions.The data came from the IAEA collaboration using the Budapest reactor.
Li6 and F19 examples were presented. MT102 file 12 was used.
Our ENDF evaluators will need to interact with Brad to figure out how best to integrated these data into ENDF files. This might include a small group including: Kawano and Brown.
There are issues to pay attention to regarding consistency with ENDF data on things like thermal cross sections.
David Brown: Atomic data: EPICS2014
Atomic data and an update to the 1989-2001 Livermore evaluations. Red Cullen has fixed bugs (eg average energy from bremsstrahlung). Comparisons were made to a recently published review of modern physics summary. Fudge, PREPRO and checking codes are used on these data.The MCNP released version had hand-editing owing to some NJOY issues (some sections missing?). The relaxation library is used by BNL for the decay sub library. There is a historical work by Storm and Israel from LANL that is used as a resource too. An IAEA collaboration is updating EADL for x Rays - Vivian is involved.
Skip and Morgan will look into creating a new MCNP library for this. Various people need to test this new set including Trimble, Sood's group, etc.
David Brown: ENDF fixes.
A review was given, and a few weeks is needed to fix many of these bugs.
David Brown: Other evaluations
Neutron target evaluation - add from Hale. New Cd data from IRMM. Mughabgab and Danon will check. Hf work from UK, IRMM and JEFF get checked by others incl RQ, Trumble, etc.
Alejandro Sonzogni: Decay data
Various fixes to decay sublibrary have been made. Nice comparisons were shown between calculated and measured actinide beta decay spectra following fission. The decay heat as a function of time is also plotted as a function light and heavy fp contributors. Asymmetric fission makes most of the delayed neutrons that play such an important role in reactors.
Dave Brown: ADVANCE advances
For automated QA testing of ENDF data. Dave Heinrichs
He uses the JENDL alpha induced library in a new capability in COG, an advance to the old LANL sources code. We could consider adopting the alpha library from Japan but it does have ENDF format deficiencies.Thermal scattering
Based on an WPEC subgroup collaboration.
He has propose fix for SiO2. NCSU.
The Argentine group has a D2O evaluation that Chalk River prefers over ENDF.
There are normalization issues in the ACER processing in NJOY for mixed moderators.
Holmes (Bettis lab) may be able to coordinate improvements for ENDF/B-VII.
Skip Kahler
CuWe are calculating high for Zeus. This includes both the new 235U data and copper data. The copper capture looked too low from ORNL - fixing this may help improve C/E.
W
The new oak ridge files perform with a similar quality to vii.1 (but this needs to be rerun using the angular information available) - perhaps with an increase in bias for hmf3 cases.
For the hst assembly the new argentine water scattering kernel makes little difference.
Mesurement Session
Yaron Danon: Update on Nuclear Data Research at RPI
Re185,187 are being measured by RPI, we should ensure that the new resonace parmaters find their way to ENDF evaluation
236U RPI measuremet shows that capture is lower, paper is publinshed
186W between 2-4 keV there was a gap, JENDL was good, ENDF/B was not very good but new ORNL evaluation might be OK. New evaluation is not on the NCSP schedule
56Fe capture measurement up to 2 MeV should be included in CIELO evaluation.
181Ta new capture measurement up to 1.5 MeV should inform ENDF that looks deficient.
16O experiment on water to reolve normalization issue. Differences between evvaluations are small. Normalization can be done within 1.5%. There is a deep 16O cross section betweeen 2-3 MeV that helps to estimate H contribution. H normalization is better that 1%. ENDF/B is fine!
Anton Tonchev: Fission Product Yields of U-235, U-238 and Pu-238 & Neutron Induced Reactions on Specific Nuclei
TUNL measurement confirms the energy trend indicated by Kawano.
Boris Pritychenko; EXFOR
Recovering of old data in ORNL that were not included in EXFOR.
Effort to reduce number of 'unobtainable' data in EXFOR for Area 1 focused on the CIELO isotopes has been completed.
Toshihiko Kawano: LANCE experiements
Preliminary results for 235U PFNS confirm current ENDF above 0.1 MeV
Covariance Session
Donald L. Smith, Chairman
Donald L. Smith, Chairman
There were only two talks presented in this session, as follows:
Jesse Holmes (Bettis), Ayman Hawari (NCSU), and Michael Zerkle (Bettis): Covariance Methodology for ENDF File 7 S(alpha,beta) and Thermal Neutron Inelastic Scattering Cross Section Data
Thermal neutron scattering from materials can be strongly dependent on the structures and molecular bonds of these materials and not just on the free-atom cross sections for the individual component atoms that form the materials. Scattering at these low energies is dominated by the interaction of the thermal neutrons with phonons (quantized vibrational modes) for the materials in question. Interaction of the neutrons with phonons is the only consequential method for energy exchange in neutron scattering with such materials. Consequently, the nature of scattering is strongly dependent on the material temperature as well as on its chemical composition and physical structure. The phonon energy spectrum ρ(ε) is the fundamental parameter determining the quantity S(α,β) which is recorded in the ENDF/B library to characterize thermal-neutron scattering. The thermal scattering cross-section can be expressed in terms of S(α,β). The work described here draws from theoretical developments in a field considerably removed from normal nuclear physics, i.e., from the field of materials science.
Heretofore, there has been no concerted effort to develop a formalism for deriving covariance information for thermal neutron scattering data. This paper introduces a formalism that can be used for this purpose. It utilizes Monte Carlo methods as part of the process to derive the required uncertainty information. In principle, implementing the approach discussed in this presentation can lead to extremely large covariance files. However, a procedure to reduce the amount of this information has been developed, and this was discussed in the talk. It involves approximating the true phonon spectrum by lower-resolution version that nevertheless captures the main features of the region of the phonon spectrum to which the thermal neutrons are expected to be sensitive at various material temperatures of interest for applications.
While this work establishes a rigorous approach for deriving S(α,β) covariance data, the job of developing formats within the ENDF/B system to deal with this information remains to be accomplished. This database management task should be addressed as soon as possible so as to enable S(α,β) covariance data generated in this manner to be incorporated into ENDF/B .
Donald Smith (ANL-Retired), Denise Neudecker (LANL), and R Capote (IAEA): Prompt Fission Neutron Spectrum Covariances: Impact of Scaling and Normalization
The emission of neutrons from fission is represented in the ENDF/B library as the product of two quantities stored in separate files. The total number of neutrons emitted, nu-bar, is very well known as a function of incident energy, at least for the major actinides. The energy distribution of these neutrons is represented as a less well-known, normalized (to unity) probability distribution. The work described in this talk focused on this distribution, also known as the prompt fission neutron spectrum (PFNS).It was pointed out that various experimental PFNS data sets used in an evaluation need to be scaled so they will be comparable. However a PFNS cannot be normalized unless the representation spans the entire energy range of neutron emission. Usually, this is only possible for an evaluated spectrum derived by consideration of both experimental data (with limited energy coverage) and model-derived data (which can span the entire energy range). PFNS data are scaled by multiplying all values of a particular data set by a single constant. There is no unique way to do this, but optimal scaling is reflected in the lowest possible chi-square for an evaluated solution. The scaled covariance matrix for a particular data set is produced by multiplying all elements of the unscaled covariance matrix by the square of the same constant as used to generate the scaled PFNS. Scaling a covariance matrix does not change its correlation pattern.
Normalization of a “complete” PFNS leads to a non-linear expression that sums over all groups to unity (in an energy-group representation). The corresponding covariance matrix can be “normalized” by a matrix transformation that leads to a result where all rows and columns sum to zero. A normalized covariance matrix leads to smaller uncertainties than the corresponding non-normalized matrix owing to the vanishing of all fully correlated uncertainties and uncertainty reduction even for partially correlated components. Model-calculated PFNS generally exhibit strong correlations over large energy ranges owing to the limited numbers of parameters involved. This can lead to “unrealistically low” uncertainties in evaluated results, especially in the region of highest neutron yield, even when experimental data are plentiful, which they often are in this region (the so-called “pivot point” region). Attempts are being made to utilize models with more parameters and weaker correlations in PFNS evaluations, but this effort remains a “work in progress”. A case can be made for expecting “reasonable” uncertainty for an evaluated PFNS near the pivot point to reflect uncertainties from the scaled experimental data (when considered collectively), and for finding ways to prevent the uncertainty in this region from being excessively reflective of the uncertainty of a utilized model-calculated PFNS component.
There are minutes attached to this event.
Show them.
