ICF等离子体全粒子碰撞动力学的一维空间确定论模拟

Inertial Confinement Fusion (ICF) induced plasma contains various complex physic processes. Full scale numerical simulations usually resort to approximations, such as the local thermal equilibrium approximation, in the purpose of reducing computation complexities. Considering that these adopted approximations sometimes could hide the undergoing physics, it is necessary to develop full-kinetic numerical simulation capacities in lower dimensions. The underlying kinetic mechanism of some specific physics and the applicability of associated approximations will be better understood as a result. Deterministic kinetic simulation, in contrast to the Particle-In-Cell (PIC) method, benefits from higher efficiency and absence of stochastic fluctuations. Nevertheless, it suffers from the complicated treatment of Fokker-Planck collision terms of charged particles and limited capability for anisotropy consideration. Based on a verified work, this research plan proposes calculating “collision-induced probability transfer table” that contains complete information for two-body particle collisions with high accuracy, and use it to replace traditional “on-the-fly” treatment of collision terms. For other issues including the high scattering dominance ratio problem in particle transport, the corresponding solutions are also outlined in this research plan. At last, the one-dimensional deterministic simulation code for ICF plasma kinetics with complete consideration of particle collisions will be established, and its effectiveness is demonstrated by investigation of several key kinetic processes.

惯性约束聚变（ICF）等离子体包含非常复杂的物理过程，全尺寸数值模拟往往采用诸如局域热动平衡等近似条件来降低计算复杂度。鉴于这些近似条件有时会掩盖真实的物理过程，有必要开发低维空间全动力学描述的数值模拟能力。以此深入理解特定物理过程的动力学机理，以及相关近似条件的适用性。确定论动力学模拟相比含粒子碰撞方法（PIC），具有高效率和无统计涨落的优势。然而，确定论处理带电粒子Fokker-Planck碰撞项的模型复杂，且难以适应强各向异性问题。为此本研究基于经过验证的工作，提出高精度计算包含粒子碰撞完备信息的碰撞转移概率密度表，替换传统的碰撞项“在线”处理方式。对于粒子输运计算中高散射占优比等其他问题，本研究也相应的提出了解决思路。最终形成ICF等离子体全粒子碰撞动力学一维空间确定论模拟程序，并通过考察若干关键动力学过程展示其有效性。

基于 Fokker-Planck 碰撞项的 Rosenbluth 势函数形式，运用数值积分方法获取粒子碰撞转换系数表实现了碰撞项高精度处理。程序编码主要采用 python 语言，包括运行逻辑，数据结构定义与计算结果后处理等；缩减了代码量且提升了可读性。计算量大的任务则由 numba 在线编译以及 Fortran 语言实现（f2py 调用），保证了计算效率。初步用于多粒子系统算例求解，数值结果表明算法具备保持碰撞项能量守恒、熵增等性质。