Analysis of the MZA/MZB benchmarks with modern nuclear data sets

• Published in: Annals of nuclear energy Vol. 62; pp. 504 - 525
• Main Author: van Rooijen, W.F.G.
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.12.2013
• Subjects: ERANOS; MOZART; Nuclear data covariance; Nuclear data processing; Nuclear reactor benchmark; Uncertainty quantification; Nuclear reactor benchmark; Uncertainty quantification; Nuclear data processing; Nuclear data covariance; MOZART; ERANOS
• ISSN: 0306-4549; 1873-2100
• DOI: 10.1016/j.anucene.2013.07.013

•ERANOS libraries are produced based on four modern nuclear data sets.•The MOZART MZA/MZB benchmarks are analyzed with these li- braries.•Results are generally acceptable in an academic context, but for highly accurate applications data adjustment is required.•Some discrepancies between the calculations and the benchmark results remain and cannot be readily explained.•Successful generation of ECCO libraries and covariance data for ERA- NOS. For fast reactor design and analysis, our laboratory uses, amongst others, the ERANOS code system. Unfortunately, the publicly available version of ERANOS does not have the most recent nuclear data. Therefore, it was decided to implement an integrated processing system to generate cross sections libraries for the ECCO cell code, as well as covariance data. Cross sections are generated from the original ENDF files. For our purposes, it is important to ascertain that the ECCO cross section libraries are of adequate quality to allow design and analysis of advanced fast reactors in an academic context. In this paper, we present an analysis of the MZA/MZB benchmarks with nuclear data from JENDL-4.0, JEFF-3.1.2 and ENDF/B-VII.1. Results are that reactivity is generally well predicted, with an uncertainty of about 1% due to covariances of the nuclear data. Reaction rate ratios are satisfactorily calculated, as well as the flux spectrum and reaction rate traverses. Some problems remain: the magnitude of the void effect is not satisfactorily calculated, and reaction rate traverses are not always satisfactorily calculated. On the whole, the ECCO libraries are sufficient for design and analysis tasks in an academic context. For high-precision calculations, such as required for licensing tasks and detailed design calculations, data adjustment is still necessary as the “native” covariance data in the ENDF files is not accurate enough.