SN共轭函数用于蒙卡粒子输运自动减方差的应用研究..docVIP

SN共轭函数用于蒙卡粒子输运自动减方 差的应用研究 刘聪张斌張亮邂君萧陈义学 华北电力大学核科学与工程学院中广核研宄院有限 公司 蒙特卡罗粒子输运方法在求解中子深穿透问题时需要多种减方差技巧提高计算 效率，减方差参数需要用户凭借经验反复试算，获取可靠有效的减方差参数比 较繁琐。离散纵标法相比蒙卡方法具有效率高、可自动获得全局解的优势，但计 算精度受多群截面、离散误差等多方面因素影响。蒙卡-确定论耦合计算可结合 彼此优点，利用S、共轭函数估计粒子对探测器的相对贡献，自动生成源偏倚与 权重窗参数用于蒙卡减方差，提高蒙卡计算效率。基于么输运计算平台ARES编 制了三维共轭输运计算模块，根据一致性共轭驱动重要性抽样方法自动生成减 方差参数，用于加速MCNP5计算。数值结果表明，自动生成的减方差参数可有效 提高蒙卡计算效率，并保证结果无偏。自动减方差技术利用S-N共轭函数可更经 济准确的估计粒子重要性，避免手动估算减方差参数的复杂工作，对于复杂屏 敝问题的蒙卡计算具有较好的应用前景。 关键词： 离散纵标法;蒙特卡罗方法;共轭输运;权重窗;源偏倚; 刘聪（1992一），男，河北涿州，博士研究生，屮子光子输运数值计 算方法研宄。 :congliulOll@ncepu. edu.cn 基金:国家自然科学基金11575061） Application of SN Adjoint Function on Automated Variance Reduction for Monte Carlo Particle Transport Calculation LIU Cong ZHANG Bin ZHANG Liang ZHENG Jun-Xiao CHEN Yi—Xue School of Nuclear Science and Engineering，North China Electric Power University; China Nuclear Power Technology Research Institute; Abstract： Monte Carlo particles transport method requires various variance reduction techniques for solving neutron deep-penetration problems. However, to obtain usable and highly-efficient variance reduction parameters is relatively difficult, which needs users intuition and manual guesswork. The discrete ordinates method, by contrast, is more efficient and capable of acquiring global solutions naturally, the calculation accuracy of which is affected by mutigroup cross sections, discretization errors and other factors. Coupled MC-SN calculation can combine each other’s advantages, using SN adjoint function to evaluate particles’ relative contribution to a detector and generating source biasing and weight window parameters automatically in order to improve the efficiency of MC simulation. Three-dimensional adjoint transport calculation module was integrated into the S、transport code 八RES and automated variance reduction parameters could be generated based on consistent adjoint driven importance sampling method to accelerate the calculation of MCXP5. The numerical results show that, automated variance r