Search

The MUSE4 programme of ADS-related integral experiments has recently been completed at the MASURCA facility at Cadarache (France). Investigations have been carried out in a reference critical fast-spectrum configuration and related sub-critical core configurations driven by an accelerator. Among the various characteristics which need to be investigated are the peak power and the peak fast-neutron flux (damage in structural materials), the presence of the accelerator tube in the core creating new types of heterogeneities, the effects of which need to be modelled and quantified by appropriate calculational tools. In this context, different fission chamber traverses and various activation foil measurements in the MUSE4 programme have covered a wide range of threshold and non-threshold reaction rates. This paper presents the analyses of the corresponding experimental results from the reference critical MUSE4 configuration. Comparisons are made with calculations using both deterministic (ERANOS-2.0) and stochastic (MCNP-4C) codes in conjunction with different data libraries (JEF-2.2, ERALIB1, ENDF/B6). Significant discrepancies are observed, particularly for high-threshold reactions in the central lead diffusing region. Of the alternative calculational routes applied, the most appropriate ones for treatment of the ADS-type heterogeneities are found to be ERANOS-2.0 (with either JEF-2.2 or ERALIB1 data) and MCNP-4C in conjunction with its ENDF/B6 library.

International audience; The MUSE4 phase of the MUSE (Multiplication with an external source) programme has been launched recently at the MASURCA facility at Cadarache (France). This series of integral experiments relevant to accelerator driven systems (ADSs) consists in studying a reference critical fast-spectrum configuration, followed by three different subcritical modifications of the same, each driven by an external neutron source produced via an accelerator. The presence of the accelerator tube in the core creates new types of heterogeneities, the effects of which need to be modelled and quantified by appropriate calculational tools. A part of the unique ADS-relevant validation base being provided by MUSE4 are results from a wide range of activation foil measurements covering various reaction rates, both with and without thresholds. Each experimental campaign consists in irradiating the foils in different representative locations to characterise the new types of spectral variations. This paper presents experimental results obtained in the reference critical MUSE4 configuration. Comparisons are made with calculations using both deterministic (ERANOS-1.2) and stochastic (MCNP-4C) codes in conjunction with different data libraries (JEF-2.2, ERALIB1, ENDF/B6). Good agreement has been obtained between calculation and experiment in most cases. However, significant discrepancies have been found for the high-threshold reaction rates inside and around the central diffusing lead region, indicating the need for more detailed investigations in this context.

As part of a joint research programme between the Paul Scherrer Institute (PSI) and swissnuclear, with the co-operation of the Leibstadt nuclear power plant in Switzerland and fuel suppliers Westinghouse Sweden, measurements and calculations have been made of the axial and radial distributions of fission and 238U capture rates in the fuel rods of a Westinghouse SVEA-96 Optima2 boiling water reactor assembly. The measurements, made in the zero-energy research reactor PROTEUS at PSI, have been compared with calculations carried out using the Monte Carlo code MCNPX. The results reported are for the regions near the ends of the part-length fuel rods, which are a feature of SVEA-96 Optima2 assemblies. The sudden increase in moderation above the ends of the part-length rods leads to power peaking in the adjacent rods. Careful attention needs to be given to this phenomenon in the deployment of such fuel, the present paper providing experimental evidence for the ability of a stochastic code to predict such effects. [All rights reserved Elsevier].

Characterisation of the SVEA-96 Optima2 boiling water reactor assembly, in terms of the radial distributions of normalised total fission and 238U capture rates, is reported at its central elevation, i.e. at the 92-pin section, where the one-third part-length pins are replaced by water. Measurements performed in the PROTEUS facility are compared with MCNPX predictions. The calculation model included the measured locations of the SVEA-96 Optima2 assemblies and sub-assemblies, within the PROTEUS test zone. Predicted and experimental fission and 238U capture rates are found to agree, respectively, within 3.5% and 4% for all pins. Fission rates in the burnable-absorber UO2-Gd2O3 fuel pins have been predicted without bias using the ENDF/B-VI data library but show an average 1.4% under-prediction with the JEFF-3.1 data library. A slight overestimation of the total fission rate in the pins located at the periphery of the assemblies was observed and has been attributed to an inaccurate modelling of the pin positions. However, there was no systematic bias observed due to the absence of the one-third pins at the corners of the assembly.[All rights reserved Elsevier].

A detailed three-dimensional, continuous-energy MCNP4B model of the LWR-PROTEUS critical facility has been developed for the analysis of whole-reactor characteristics using ENDF/B-V, ENDF/B-VI and JEF-2.2 cross-section sets. The model has been applied to the determination of the critical loading, as well as the evaluation of reactivity worths for safety/shutdown rods, control rods, and individual driver-region fuel rods. The initially obtained results for the first configuration investigated (Core 1B) indicated that, for the same geometrical and materials specifications, the ENDF/B-V data library yields the closest critical prediction (discrepancy of 64040 pcm), followed by ENDF/B-VI (98040 pcm) and JEF-2.2 (134040 pcm). The differences in results between the three data libraries were studied by considering the contributions of individual materials to the neutron balance. 235U and 238Pu cross-sections from JEF-2.2, for example, explain an effect of ~400 pcm. Refinement of the materials specifications in the MCNP4B whole-reactor model, in particular the impurities assumed for the graphite driver of the driver and reflector regions, has been shown to reduce the final discrepancy of the ENDF/B-V based keff result to ~0.2%. The MCNP4B results for relative reactivity effects, such as control rod worths, are found to agree within experimental errors with the measured values

The recent experimental programme conducted in the PROTEUS research reactor at the Paul Scherrer Institute (PSI) has concerned detailed investigations of advanced light water reactor (LWR) fuels. More than fifteen different configurations of the multi-zone critical facility have been studied, each of them requiring accurate estimation of operational safety parameters, in particular the critical driver loadings, shutdown rod worths and the effective delayed neutron fraction eff. The current paper presents a full-scale 3D Monte Carlo model for the facility, set up using the MCNPX code, which has been employed for calculation of the operational characteristics for seven different LWR-PROTEUS configurations. Thereby, a variety of nuclear data libraries (viz. ENDF/B6v2, ENDF/B6v8, JEF2.2, JEFF3.0, JEFF3.1, JENDL3.2, and JENDL3.3) have been used, and predictions of keff and shutdown rod worths compared with experimental values. Even though certain library-specific trends have been observed, the keff predictions are generally very satisfactory, viz. with discrepancies of 0.5% between calculation (C) and experiment (E). The results also confirm the consistent determination of reactivity variations, the C/E values for the shutdown (safety) rod worths being always within 5% of unity. In addition, the MCNP modelling of the multi-zone reactor has yielded interesting results for the delayed neutron fraction (eff) in the different configurations, a breakdown being made possible in each case in terms of delayed neutron group, fissioning nuclide, and reactor region. [All rights reserved Elsevier].

In the context of the LWR-PROTEUS program, radial and azimuthal 235U fission (F5) and 238U capture (C8) rate distributions have been calculated for zero-burnup pins of a Westinghouse SVEA-96 Optima2 boiling water reactor fuel assembly using the stochastic MCNP4C and the deterministic CASMO-4 codes. The within-pin F5 distributions predicted by the two codes are in very good agreement; the C8 distributions are more pronounced, and there are significant discrepancies between the codes, both azimuthally and radially. The calculations have been compared with experimental results obtained from activation foil measurements in two pins of the assembly irradiated in the center of the PROTEUS test zone. The measurements confirm that the two codes can accurately predict the radial and azimuthal F5 distributions but that MCNP4C within-pin C8 distributions are much more accurate than those of CASMO-4