Topic 2.5 ANALYSIS OF PROTEUS LWHCR PHASE 11 EXPERIMENTS PERFORMED USING THE AARE MODULAR SYSTEM AND JEF BASED LIBRARIES

The capability of the AARE modular code system and JEF-1 based nuclear data libraries to analyse LWHCR lattices is investigated by calculating the wet and dry cells of the PROTEUS-LWHCR Phase II experiment. The results are compared to those obtained using . several celI codes. Main features of AARE, such as the self-sbelding of resonance cross sections in the whole energy range, the generation of adequate fission source spectra, the accurate calculation of migration areas, and the efficiency of the elastic removal correction are investigated. In particular, it is shown that AARE can predict the k, void coefficient well with a 1% deviation from experiment, whereas the other codes give larger deviations. INTRODUCTION Existing spectrum codes and related nuclear data libraries were mainly developed for design calculations of Light Water or Fast Breeder reactors. Therefore many simplifications were’ considered. For example, the resonance and fast energy ranges (in the case of LWR-codes), and the lowest resonance and thermal ranges (in the case of FBR-codes) are treated in a vay, 0 which is not always adequate for Light Water High Conversion Reactor (LWHCR) calculations. This can lead to conflicting calculational results especially in the determination of eigenvalues, 0 reaction rates, and void coefficients. Therefore the development of suitable methods and t.he generation of adequate nuclear data libraries is a key issue in assessing the feasibility of a LWHCR core. In this framework the new general purpose code system AARE (Advanced Analysis .for Reactor Engineering)* was created and JEF (Joint Evaluated Fide) based nuclear data libraries were generated.* In this paper the capability of the AARE modulk system and JEF based : ,,.,.: ;‘;: :: .&clear data libraries to analyse LWHCR lattices is investigated by calculating the PROTEUSLWHCR Phase II experimentP The results are compared to those previously published whare several cell codes, including WIMS-D, MICROX-2, KARBUS, and SPEKTRA were used. AARE, a PSI (Paul Scherrer Institute) update of the DANDE (applied nuclear DAta, core Neutronics DEpletion) system from Los Alamos, is a flexible code package which allows applications to a broad class of reactor types. It includes as well burn-up and sensitivity codes. Witbin AARE the coupling a+d reformatting programme TRAMIX, a PSI update of TRANSX-CTR, and the one-dimensional discrete-ordinates transport code ONEDANT are available. TRAMIX is able to shield resonance data in the whole energy range using the intermediate resonance :absorption shielding method of resolved resonances (IR method) by calculating isotope depend,& Goldstein-Cohen X factors and energy dependent Dancoff corrections, to compute accuraicly fission spectra from fission matrices, to correct elastic scattering matrices for differences between actual flux and library weighting flux, and to produce a variety of library formats which can be used in most transport and diffusion codes. Particular features of AARE, such as the self-shielding of resonance cross-sections in the whole energy range important for structural materials and actinides (including the low energy resonances of heavy actinides such as the capture resonance of *4zPu at 2.7eV), the shielding of oxygen resonances in the MeV range, the generation of adequate fission source spectra, the accurate calculation of migration areas with transport cross sections obtained using the: inflow formula, the choice of different group structures, and the efficiency of the elastic removal correction are investigated. Also analysed is the IR method by making a comparison with an accurate pointwise slowing down calculation in the resolved resonance range using the cell code MICROX-2. For the sake of consistency all cross-sections are taken from the JEF fle. The cross isection generation scheme NJOY (version 6/83) h as been employed to produce MATXS formatted nuclear data libraries in 70, 193, and 308 groups for use in the AARE system. Additionally, the editing module MICROR was utilized for creating the MICROX-2 libraries (FD, GG, GAR data files), and the can&an module WIMSR was tested and updated to generate with the management code WILMA a JEF based library for the code WIMS-D (WIMS-JEF library). The preprocessed WIMS81 data library was used for comparison. Eigenvalues, migration areas, reaction rate ratios of important actinides, and void coefficients are computed and compared with the series of experiments (Phase II) currently performed at the PROTEUS-LWHCR facility. PROTEUS includes a central test zone of mixed oxide fuel PuO2/UO2 character&d by a volumetric moderator/fuel ratio of 0.48. Cores 7-8 indicate, respectively, the fully watered and dry core. Both cores consist of a three region cell wi,th fuel pin, steel cladding, and moderator. Experimental errors on the standard types of reaction rate ratios were found to be between &1.5% and f2.0%. For the fission of *41Pu and the capture of 242Pu the errors are f3.0%3. RESULTS Table I summarizes some calculated-t~experimental (C/E) al Y ues for reaction rates a.:nd k,, and computed absolute values for migration areas (M’) obtained using the AARE system. The IR method was utilized, and cross sections were shielded in the whole energy range. A cell flux weighted fission spectrum for the whole cell (converged spectrum) was determined iteratively. The elastic removal correction was utilized in the case of 70 neutron group calculations. Throughout the whole paper F9 denotes the fission rate of ?‘u per atom, C8 the capture (i.e. (n,7)) rate of *%U per atom, F8 the fission rate of *%U per atom, F5 the fission rate of =‘U per atom, Fl the fission rate of 241Pu per atom, and C2 the capture rate of 242Pu per atom. Table I shows that all investigated energy structures and JEF data libraries give generally a good prediction of experiments. Furthermore, the migration area calculations agree well for both cores within 1% for all energy structures. The use of 70 neutron groups together with previously mentionned AARE options gives small deviations from the results carried out with finer energy structures. However, the prediction of the capture rate of 242Pu in Core 7 improves with increasing number of energy groups, and is accurate within 1.3% when using the 308 neutron group MATXS library. This is because TRAMIX is able to shield the main resonance of *42Pu