中子截面Doppler展宽算法及并行计算方法

Parallel computing for Doppler broadening of neutron cross sections

为了提高现有中子截面Doppler展宽的计算效率,在反应堆物理蒙卡计算中实现在线Doppler展宽,提出了基于Gauss积分的连续能量中子截面Gauss展宽算法,以及与之相关的并行化计算方法,形成了全新的快速Doppler展宽方法,并编写了基于该方法的连续能量中子截面Doppler展宽程序。使用多个算例对该方法及程序进行了验证,并与传统方法进行了比较。验证结果表明,该方法在保证计算结果与传统Doppler展宽方法一致的前提下,计算速度相比传统方法提高了一个量级,完全可满足反应堆蒙卡计算中温度相关问题对截面快速Doppler展宽的需求。

A Gauss integration algorithm developed for Doppler broadening of continuous-energy neutron cross sections improves the processing efficiency for online Doppler broadening in reactor physics Monte Carlo calculations using parallel computing. The code was validated using several models. The results show that this method gives a speedup of roughly 10 fold compared with the traditional broadening algorithm with the same accuracy, which is needed for fast Doppler broadening in temperature-dependent Monte Carlo calculations in reactor physics.