Homogenization-free reactor core analysiswith general first collision probabilities method

Abstract

The C5G7 MOX Benchmarkfor current codes has been proposed as a basis to test the ability of current transport codes to teat reactor core problems without spatial homogenization. This is a seven-group form of the C5G7 MOX fuel assembly problem specified by Cavarec et.al. There are four fuel assemblies, two contain UO 2 fuel elements and two contain MOX fuel elements. Seven group cross sections for different kinds of fuel (three enrichment of MOX and UO 2), the guide tubes, the fission chambers and moderator are given. Thus this benchmark is just a mathematical test that allows testing the accuracy of the neutron transport equation solution with different methods and codes. In this paper the General First Collision Probabilities Method (GFCPM) is used to analyze the two-dimensional configuration of this benchmark. A linear flux approximation is used in the reflector. Different calculation schemes in the reflector region have been used. The output results, K eff and the pin powers have been analyzed. The convergence of the results has been analyzed both as a function of the subdivision scheme of the reflector region and of the number of points in the calculation scheme for general first collision probabilities. Comparison has been carried out for K eff and pin powers both with the reference results (external convergence) and with the results of different approximations of GFCPM (internal convergence).