中文摘要：

在反应堆pin-by—pin精细建模及蒙特卡罗模拟计算研究中，由于不同栅元的功率密度差异较大，导致蒙特卡罗方法临界计算的样本在不同栅元之间的分配不均衡，由此引起栅元内的各种计数的统计误差差异较大．为使大部分栅元内计数的统计误差降至一个合理的水平，单纯增加总样本已不是一个高效的解决方法．通过在特定临界计算迭代算法的基础上改进并实现均匀裂变源算法的思想，对大亚湾压水堆pin—by—pin模型取得了具有较高效率的数值结果．本工作为具有自主知识产权的蒙特卡罗粒子输运模拟软件JMCT最终达到反应堆pin—by—pin模型（包括一系列国际基准模型）的模拟性能要求提供了一个有效的工具．

英文摘要：

Because of a very non-uniform power distribution in core region, a very non-uniform distribution of relative un- certainties exists for tallies in Monte Carlo criticality calculations of pin-by-pin reactor model. To make a large part of cells obtain small enough relative uncertainties with reasonable time costs, increasing the totM sample scale is not a good choice. By realizing a modified uniform-fission-site algorithm on the basis of source iteration algorithm of parallel Monte Carlo transport code JMCT, we obtain higher efficiencies for tallies in the calculations of pin-by-pin model of the Dayawang reactor plant. This work supplies a useful tool for matching the goal of simulating the benchmark pin-by-pin reactor models with a pre-described standard（the so called Kord-Smith challenge）.