RSICC PACKAGE MIS-003

Miscellaneous Software Package 003

VVER -BENCHMARKS

Collection of Neutronic VVER Reactor Benchmarks

Idaho National Engineering and Environmental Laboratory

Idaho Falls, Idaho, USA

Reference: INEEL/EXT-99-00818, IPPE (August 1999)

Abstract

A system of computational neutronic benchmarks has been developed. In this CD-ROM report, the data generated in the course of the project are reproduced in their integrity with minor corrections. The editing that was performed on the various documents comprising this report was primarily meant to facilitate the production of the CD-ROM and to enable electronic retrieval of the information. The files are electronically navigable.

The set of benchmarks is primarily comprised of interrelated benchmarks that share situation descriptions and data. The system covers the VVER-440 and VVER-1000 reactors extensively, and in a few cases the newer VVER-640 reactor design. Benchmarks within the system address the static, dynamic (with and without feedback), fuel depletion and fuel loading optimization, and normal as well as accident situations.

The development of new computer codes for the analysis of nuclear reactors requires comparisons of their results to reference results obtained for well-established computational benchmarks. This comparison allows the assessment of the computer code capability to produce accurate results. The level of fidelity of the models the codes can use is assessed in comparisons with experimental benchmarks. The comparisons with either kind of benchmark allow the identification of areas for improvement and refinement of the code. The importance of obtaining useful benchmarks for VVER reactors has long been recognized and a number of completed as well as on-going efforts address this need. Among these are the AER computational benchmarks, the TIC experimental benchmarks for assemblies, the ORNL sponsored benchmarks for the utilization of excess weapons plutonium, and the fuel cycle benchmarks presented in the IAEA TECDOC-847 report (IAEA 1995).

The goal of the project was the production of a comprehensive interrelated set of computational benchmarks for VVER reactors. Although the benchmarks are intended to address primarily neutronics issues, some of them incorporate thermal-hydraulic feedback. The new set of benchmarks is meant to complement previous VVER benchmarking efforts, and in particular the AER benchmarks. The set of benchmarks is extensive. It consists of 36 different problems for each of the two types of current VVER reactors. These include steady state as well as kinetic benchmarks (without and with feedback) and uranium as well as MOX fueled cores.

The benchmarks, though address hypothetical situations and though using simplified data values, are designed to provide a realistic model for the expected behavior of the reactors under consideration. The results they produce are representative of the known physics of the VVER reactors. The numerical values obtained are always plausible and follow expected trends. For each benchmark situation, two independent solutions have been obtained. These solutions are "reasonable" accuracy solutions that are obtained via the ordinary use of the relevant computer codes. Most were generated using production codes. The advantage of using such codes is that the trends derived from the benchmarks can be assessed correctly against those expected in real situations. A disadvantage is the limitation in fidelity stemming from the lack of modeling flexibility in some of the codes. Reference solutions that require the use of special codes or the extrapolation to infinitesimal space and time mesh, or the use of tighter than ordinary convergence criteria, were not produced. Such reference solutions are highly desirable and should be obtained in the future.

The set of benchmarks and solutions presented in this report, though extensive, could be expanded. Besides the production of reference solutions mentioned above, the production of additional reasonable accuracy solutions would allow the inter-comparison of computer code systems.

Two principal directions for future work are identified. These are the production of high-accuracy reference solutions and the extension of the benchmarks system to other applications.

The production of high-accuracy, high fidelity reference solutions for the Benchmark System is incomplete. Reference solutions are 10-100 times more accurate than ordinarily used numerical solutions. Obtaining such high precision solutions is a painstaking task that requires the use of high performance computers. Nevertheless, the availability of reference solutions is the only means for evaluating the numerical truncation errors from the discretization approximations and for assessing the convergence of iterative methods. The contribution of reference solutions is planned and is invited from other contributors. Collaborative efforts are particularly desired.

It is easily recognized that the benchmark set could be extended. In particular, it would be very interesting to generate 3D depletion benchmarks, and reactor vessel fluence benchmarks. A subject of special interest would be benchmarks with core geometry irregularities caused by deformation of the subassemblies. The generation of additional benchmarks is planned and is invited from other contributors.

Although much care was put into generating the information presented in this report, it is not expected to be free of errors. We apologize in advance to those who might be inconvenienced by them, and we thank those who will point them to us. We encourage and welcome criticism and the pointing of errors and inaccuracies by the readers. In addition, we welcome contributions by colleagues everywhere who would want their solutions and their benchmarks incorporated in a future version of this report. Please, send comments, error reports, independent solutions, and new benchmarks to A. M. Ougouag (oom@inel.gov) or to I. R. Suslov (suslov@ippe.rssi.ru).

Abderrafi M Ougouag

Idaho National Environmental and Engineering Laboratory, USA

Dr. Igor R. Suslov and O.G. Komlev

Institute for Physics and Power Engineering, Russia

KEYWORDS: BENCHMARK PROBLEM CROSS SECTIONS; DATABASE; THERMAL HYDRAULICS