Research On Source Biasing And Mesh-based Weight Window Methods For Monte Carlo Particles Transport Simulation

Monte Carlo method is widely used in particle transport calculation for the advantages of high simulation accuracy and flexible geometry representation. However, the slow convergence rate of Monte Carlo method is the primary problem that limits its application in reactor analysis, and variance reduction is the major method to solve the slow convergence problem. Source biasing and mesh-based weight window methods are effective variance reduction methods to solve shielding calculation problems in nuclear systems. This thesis perform many studies for the efficient application of source biasing and mesh-based weight window, which were implemented in the Monte Carlo particle transport simulation, and integrated in Super Monte Carlo Calculation Program for Nuclear and Radiation Process (SuperMC) to carry out its variance reduction methods and enhance its advantages of dealing deep penetration problem in shielding calculations. The studies also make a base for more advanced reduction variance in SuperMC, which will be importance for advanced nuclear system shielding design.Firstly, source biasing sampling method combining energy, position and direction, and mesh-based weight window under Cartesian and cylindrical coordinate systems were implemented in SuperMC framework. With a series of benchmarking problems, the testing results show the calculation of the module has a good agreement with those of MCNP developed by Los Alamos National Laboratory, USA.Secondly, in order to apply source biasing and mesh-based weight window methods easily and efficiently, the source biasing parameters and mesh-based weight window division are studied. Two regulations are obtained with large number of testing and analyzing (1) source position biasing parameter should be inversely proportional to the volume of source cell and the distance of source to tally with a certain range, (2) along the direction of interest, rough mesh division in vacuum or low density regions and fine division in high density regions can improve the calculation efficiency.Finally, a comprehensive test was performed with the neutron flux calculation for China Lead-based Research Reactor (CLEAR-I) reactor vessel. The computational efficiency of source biasing method, mesh-based weight window method were improved with up to 13.58 times by setting the parameters with the two regulations, which further demonstrate the effectiveness.