Least-Squares Fitting Method for On-Line Flux mapping of CANDU-PHWR

Abstract

A least-squares fitting method is developed for advanced on-line flux mapping in the CANDU-PHWR system. The method solves both the core neutronics design equations and the detector response equations on the least-squares principle which leads one to normal equations. The fine-mesh finite difference two-group diffusion theory calculations by SCAN code for Wolsong-3 unit are conducted to obtain the simulated real flux distribution and detector signals. The least-squares flux monitoring calculations are compared with the flux distribution calculation by the SCAN code without detector signals. It is shown that the least-squares method produces the flux distribution in better agreement with reference distribution than the coarse mesh SCAN calculation without detector signals. Through the 500 full power day burnup-history simulations of Wolsong-4 unit for benchmark, the mapped detector signals are compared with real detector signals. Maximum root mean squares (RMS) difference between the mapped detector signals and real detector signals are shown to be about 0.04 % by least-squares method, while it is about 5.43 % by the current flux-synthesis method. It is concluded that the least-squares fitting method is very promising as the advanced flux mapping methodology for CANDU-PHWR.