Assessment of The Coarse-Mesh Finite-Difference Method With The Multicell Methodology in RFSP for ACR-1000®

Abstract

Though the WIMS-AECL/RFSP-IST suite of codes has been in widespread use for design, safety and operational analysis over the years in the CANDU® industry, it is necessary to assess its adequacy for future CANDU applications, especially when new fuels and new reactors are under development. Use of the coarse-mesh finite-difference method in RFSP 2-group neutron-diffusion calculations has drawn particular attention from the designers and regulators, especially for the application to ACR-1000®. This paper presents assessment results between WIMS-AECL/RFSP (with various lattice sub-meshes with and without using the multicell methodology) and MCNP5 full-core calculation for the ACR-1000. The assessment results show that the agreement in the channel power between RFSP and MCNP is significantly improved when the multicell methodology is used in the RFSP calculations. The assessment also shows that the coarse-mesh finite-difference method (FDM) gives consistently smaller differences than the fine-mesh FDM. The reason for the better performance of the coarse-mesh FDM over the fine-mesh FDM is also discussed in the paper.