Quality Improvement of the EXFOR

database: year 1

WPEC subgroup 30

Arjan Koning

NRG Petten, the Netherlands

SG-30, JAEA, Tokai, June 4, 2008

Contents

* Objectives

* Problems

* Progress in 2007-2008

* Plans and possibilities for 2008-2009

* Conclusions

SG-30, JAEA, Tokai, June 4, 2008

Objective

* Make EXFOR an easy accessible and correct

database, available in computational format.

This enables:

– More efficient data evaluation

– Easy and extensive validation of model codes

– More feedback from users to EXFOR

maintainers.

SG-30, JAEA, Tokai, June 4, 2008

Possible issues for improvement

* Completeness (briefly mentioned here)

* Database management and EXFOR format

(not considered here, apart from format

harmonization)

* Data Retrieval (completeness)

* Quality (correctness)

SG-30, JAEA, Tokai, June 4, 2008

SG-30, JAEA, Tokai, June 4, 2008

Why improvement?

* All other aspects of nuclear data evaluation are well

automated:

1. Robust nuclear model codes

2. Plotting software

3. Checking, processing software

4. `Scriptwise’ nuclear data evaluation

* Exp. data retrieval should not become the delaying factor

* Evaluator needs access to all data

* Covariance evaluation requires access to all data

* Experimentalist deserves appropriate credit.

SG-30, JAEA, Tokai, June 4, 2008

Approach

* Use format converters and checking routines

– Record how many % is converted

– E.g. check for negative cross sections

– E.g. check for xs > 4 barns for E > 0.1 MeV

* Some errors are obvious just by plotting the

values

* Use the power of nuclear model codes:

– Chi-square checks

– Visual inspection

SG-30, JAEA, Tokai, June 4, 2008

SG-30, JAEA, Tokai, June 4, 2008

Progress in 2007-2008

* April 20 2007: SG-30 approved at WPEC

meeting (compromise: remove quality flagging)

* June 2007: Initial emails sent:

– A lot of moral support

– Mailing list created: 150 emails so far

– About 30-40 members on mailing list

* June/July 2007: XC4 format by V. Zerkin. Bilateral

communication with A. Koning

SG-30, JAEA, Tokai, June 4, 2008

Progress in 2007-2008

* July 2007: Correction/investigation of AK’s first list

of errors by S. Dunaeva and O. Schwerer.

* September 2007:

– Statistical tests by Emmeric Dupont: large list of

problems

– Preparing JANIS for SG-30 by NEA

– Error reports by Konobeyev, Mashnik, Duarte,

Leray

* October 10-11 2007: first SG-30 meeting at IAEA

SG-30, JAEA, Tokai, June 4, 2008

Progress in 2007-2008

* November 2007-May 2008:

– Statistical tests by Emmeric Dupont

– High-energy (proton) problems

investigated and reported by David, Leray

and Otsuka

– Computational database C4 with JANIS by

NEA

SG-30, JAEA, Tokai, June 4, 2008

Progress in 2007-2008

– May 2008:

–XC4 database by Viktor Zerkin (IAEA):

–Claim: 91% of X4 to C4 conversion

–80% of this translated into directory-structured (reaction-by-reaction) database by Koning

–Conversion now includes cross sections, differential and double-diff spectra, angular distributions, nu-bar, fission yields, resonance parameters, ratios.

SG-30, JAEA, Tokai, June 4, 2008

Progress in 2007-2008

* May 2008: – The renormalization issue again comes to surface

(not part of SG-30)

– List of problems from analyses by Koning and

Dupont.– Viktor Zerkin produces list of statistics of EXFOR

(contents) and CINDA-NSR-EXFOR comparison

SG-30, JAEA, Tokai, June 4, 2008

Important info

* SG30 mailing list: sg30@nea.fr

* SG30 website at NEA:

www.nea.fr/html/science/wpec/SG30/

* IAEA repository of entire EXFOR and XC4

databases: www-nds.iaea.org/x4toc4-master/

SG-30, JAEA, Tokai, June 4, 2008

SG-30 meeting JAEA, June 4 2008

* Exfor coverage:

– Detailed insight in missing PRC, NPA, NSE, etc. papers (Mengoni). 1970-

2005: 90% coverage, 700 papers to be done.– Proton-induced residual cross section coverage (Yoshida, Kato): 200

papers to be done.

* Model comparison:

– Large-scale TALYS-EXFOR comparison by Konobeyev for protons. Koning

will do the same for neutrons.

* Renormalization and quality flagging:

– Could be included in EXFOR relational database

– Short term: keep it independent from EXFOR

SG-30, JAEA, Tokai, June 4, 2008

Plans for 2008-2009

* Finalize XC4 conversion by IAEA (X4toC4)

and NEA (JANIS) independently

* Input by USA announced: Sigma (BNL),

ORNL database, LLNL python package x4i

* Comparison with model codes

* Specific challenges:

– Unification of high-energy reactions

– Covariances

SG-30, JAEA, Tokai, June 4, 2008

See www.talys.eu

SG-30, JAEA, Tokai, June 4, 2008

SG-30, JAEA, Tokai, June 4, 2008

SG-30, JAEA, Tokai, June 4, 2008

Deliverables

* EXFOR database in computational format. Annual, or more frequent, releases in increasing quality and completeness (also for “mother” database)

– Responsible: Data Centers

– Input: Users

* Final report:

– Status in 2007

– Description of improvements and tools

– Final status in 2009

SG-30, JAEA, Tokai, June 4, 2008

Milestones

* 6 months:

– Collect and compare all software that processes

EXFOR(X4toC4, JANIS, etc.) and merge this into

one strategy for conversion into computational

format. Done

– Correction of EXFOR with first lists of errors

(Forrest, Koning, etc.) Done

* 12 months:

– Computational library #1: All cross sections Done

SG-30, JAEA, Tokai, June 4, 2008

Milestones

* 24 months:

– Computational library #2: All cross sections

+ angular distributions + single- and

double-differential energy spectra +

everything else

SG-30, JAEA, Tokai, June 4, 2008

Conclusions

* SG30 is well underway.

* Almost the entire EXFOR database is available in

computational format (XC4)

* Testing with independent tools will continue

* Significant lists of problems have been produced:

The “success” of the SG30 approach may put quite a

large burden on the EXFOR maintainers at the

NRDC. The positive impact on nuclear data

evaluation justifies this.