Research And Optimization Of Point-wise Nuclear Data Treatment In Monte Carlo Code

Design studies of Generation IV reactor systems require the capable computational tools for arbitrary geometries, various spectrum and complicated material composition. Particle transport codes based on Monte Carlo method can perform this kind of tasks. However, Monte Carlo method still confronts many challenges, among which the most significant problem is time consuming.By quantitatively investigating the time distribution of Monte Carlo method when performing core analysis, we found that calculation of point-wise cross section mostly accounts for more than half of the total CPU time in both criticality problem and depletion problem involving reaction rate tallies. Most Monte Carlo codes use the point-wise nuclear data format. The main work of this paper is to improve the efficiency of utilizing this kind of data.We made a scrutiny into the feature of cross-section utilization in criticality calculation, and presented'Step Method'. The key idea is to select the best time and appropriate energy grid in cross-section calculation of all levels, according to the using frequency. Step Method can increase the efficiency of fast-spectrum cases to 2~3 times, and speed up the thermal-spectrum ones about 15%~50%.The individual energy grids of nuclides require large number of energy searches in cross-section calculation. To figure out the problem,'Hash-table Method'is chosen to reduce the searching times by elaborately designed hash functions. As compared with traditional binary search, Hash-table Method achieves a speedup of 3~9 times. Combining the Hash-table and Step Method, the depletion calculation with large number of reaction rate tallies sees the efficiency gains of about 20%~40%.Step Method and Hash-table Method is now implemented in the Monte Carlo code RMC, developed by Department of Engineering Physics, Tsinghua University.