反应堆瞬态的蒙卡模拟中自适应样本数方法研究

Neutron transport is a key research topic in reactor physics, in which Monte Carlo simulation is widely used. So far Monte Carlo method has been broadly adopted in static calculations, but dynamic simulation of reactor transient process is still in its initial stage. This project attempts to study dynamic Monte Carlo simulation of neutron space-time kinetics during a transient..1) Obtain the continuous distribution of discrete survivors by kernel density estimator to reduce correlation and optimize the sampling..2) Apply Shannon entropy and relative entropy in phase space to get a quantitative convergent criterion of Monte Carlo calculation..On this basis, we can implement adaptive sampling function in dynamic Monte Carlo program and speed up the transient simulation significantly...This project will lay the foundation for multi-physics coupling of dynamic Monte Carlo with thermal-hydraulic feedback, and will be applicable in many other fields, such as fast-burst reactors and inertial confinement fusion, so it has important scientific value and practical significance.

中子输运是反应堆物理的核心研究课题，而蒙卡模拟是解决这类问题的一种重要方法。目前蒙卡方法的应用集中定态问题方面，针对瞬态分析等动态问题的研究刚刚起步，本项目拟对其进行深入研究：.1）由于时间截断导致的残存粒子源是通常的定态蒙卡计算未涉及的，通过核密度估计方法将残存粒子源的离散分布转化为连续分布，可以减小样本间的相关性。.2）将香农熵、相对熵从坐标空间推广到相空间，应用于每一时间步内蒙卡计算的收敛性考察，给出定量判据。.在上述工作基础上，能够实现瞬态模拟的自适应样本数计算功能，对程序起到明显的加速作用。..本项目不仅为后续开展中子输运和热工水力多物理耦合的瞬态数值仿真奠定了基础，相关工作还能够应用于快中子脉冲堆、惯性约束聚变等领域，具有重要的科学价值和应用前景。