Development of a deterministic truncation of Monte Carlo solution for a pin-resolved nuclear reactor analysis

Abstract

An improved deterministic truncation of Monte Carlo (iDTMC) solution method has been developed for acceleration and variance reduction of the Monte Carlo (MC) neutronic analysis. For the iDTMC method, an original DTMC method was modified to eliminate numerical instability and potential bias and to guarantee stable and consistent solutions. A new calculation strategy has been applied by taking advantage of the partial-current-based coarse mesh finite difference (p-CMFD) and partial-current-based fine mesh finite difference (p-FMFD) methods in a combined way. In inactive MC cycles, fast convergence of source distribution and stable deterministic calculation are attained by the p-CMFD method, and both neutron multiplication factor and detailed pin-resolved power distribution are obtained by the p-FMFD method in active MC cycles. In addition, some assistant methods have been also devised to enhance the efficiency and robustness of the iDTMC method: a one-node p-CMFD method for fast deterministic p-FMFD calculation, strategies for CMFD-accelerated inactive cycles for stable and efficient calculation, and a real variance estimation method based on a sampling scheme with the FMDF parameters for reliable and acceptable stochastic error evaluation. Theoretical backgrounds of the auxiliary methods are also discussed in detail. The iDTMC method is tested and evaluated in a small modular reactor and a big commercial PWR, i.e. the APR1400 reactor. The essential principles of the iDTMC method are investigated and demonstrated in detail by analyzing characteristics of the FMFD parameters for the two benchmarks. The iDTMC solution is compared with that of the conventional CMFD method and a reference solution. Moreover, numerical performance is evaluated in terms of computing time and figure-of-merit.