Analysis of Sneak-7A critical benchmark using 3-D deterministic transport and sensitivity analysis methods

Some high-quality reactor physics benchmark experiments are being re-evaluated with today's state-of-the-art methods, particularly using that of detailed 3-dimensional models. One experiment analysed in the framework of the International Reactor Physics Benchmark Experiments (IRPhE) project is SNEAK-7A. This assembly is characterised by a Pu-fuelled fast critical assembly in the Karlsruhe Fast Critical Facility for the purpose of testing cross section data and calculational methods. As the detailed information on the SNEAK-7A benchmark experiment becomes available, the purpose of this paper is to model this experiment as closely as possible to the configuration as it existed in the critical facility. The experimental keff was determined to be 1.0010, which is 29.6 cents supercritical. The realistic modelling of the SNEAK-7A assembly was performed using the DANTSYS code capability for X-Y-Z geometry. The calculated core eigenvalue from THREEDANT is 1.00975. With corrections applied for core plate cell heterogeneity and mesh sizes, the best-estimate core criticality with JEF-2.2-based cross-sections turns out to be 1.01137. While the plate heterogeneity effect from flux redistribution was at first estimated to be as large as 387 pcm from plate cell calculations, it proves to be 142 pcm when the core-wide heterogeneity effects are accounted for. In order to figure out the over-prediction of core eigenvalue, spectral indices are investigated, which led to projecting that the 238U capture cross-sections are being underestimated. This fact is confirmed in the comparison of the central material worth of 238U with the measured value. When the sensitivity of core eigenvalue to the cross section is used, the newly estimated core eigenvalue is 1.00175, which is very close to the measured core eigenvalue, when the 238U capture cross-section was assumed to increase by 5% implied from the comparison of spectral indices. Once the details in the old critical experiments are available, they can be valuable resources for verifying the quality of modern neutronic calculation capability. With minimal approximations involved, the results from the re-evaluations allow to draw firm conclusions on the overall quality of experiments or calculations, as it is shown in the article. (authors)