An energy-group structure optimization from seven to four for PWR-core pin-by-pin calculation

Abstract

The pursuit of computing accuracy and reliability for PWR-core calculation encourages the pin-cell-homogenization-based pin-by-pin analysis to replace the traditional assembly-homogenization-based two-step method. Generally, the pin-by-pin whole-core calculation employs a multi-group (generally, more than 7 groups) low-order neutron-transport solver. The refinement of the spatial, angular, and energy meshes makes the computational resources required significantly increased. In this paper, a further condensation of the energy-group structure has been done as an optimization for the PWR-core pin-by-pin calculation to balance the computing cost and accuracy. Based on the homogenization theory, the influence of the group condensation on the core calculation accuracy is analyzed from the perspective of the fluxes under different boundary conditions. A spectra ration factor based on the fine-group flux difference is employed for the construction of the fitness function, which was later used for numerical optimization. Finally, a suitable 4-group structure condensed based on the 69-group heterogonous flux has been obtained by using an optimization algorithm (PSO in this paper) and verified in 2D multi-assembly, 2D whole-core, and 3D load-following problems. It has been demonstrated that the application of the 4-group structure can achieve the same level of precision compared with the existing 7-group structure but greatly improves the computational efficiency and reduces the storage by 50%.