中文摘要：

提出了一种利用频域电磁场快速计算微波器件微放电阈值的粒子模拟方法.首先通过CST微波工作室频域求解器获得微波器件中频域电磁场分布,在微放电过程模拟时将其转换到时域,再采用Boris算法求解电磁场中的电子运动,然后判断电子是否与三角面片边界相交,进行二次电子发射处理.变化输入功率,经过系列粒子模拟后,根据电子数目随时间的变化曲线确定微放电阈值.采用该方法分别对平行平板和同轴传输线微波器件的微放电阈值进行模拟计算,并与CST粒子工作室的模拟结果进行对比.结果表明,两者获得的阈值基本一致,但本方法的计算效率提高了1—2个数量级.

英文摘要：

In order to compute the multipactor thresholds of microwave devices with high efficiency and precision, a novel fast particle-in-cell（PIC） method is proposed, which takes advantage of the frequency-domain（FD） electromagnetic field solver of CST Microwave Studio（MWS）. At the initial stage of multipactor（when there are not many electrons in the device）, the self-consistent field generated by the electrons is much smaller than the applied electromagnetic field.Therefore it can be ignored in calculating the multipactor threshold and this will significantly reduce the computation burden. During simulations of multipactor process, the FD field pre-calculated by CST MWS is converted into timedomain（TD） scaling with the square root of the input power. Then the electron motion is investigated by Boris algorithm. When the electrons hit the boundaries of the simulation region, where triangular facets from CST are used for discretization, the secondary electrons will be emitted. After a series of simulations with variable input powers, the multipactor threshold is determined according to time evolution of the electron number. The multipactor thresholds in a parallel plate and a coaxial transmission line are investigated, and used as relevant verifications. Compared with the CST Particle Studio（PS）, the fast method obtains almost the same thresholds, while the computational efficiency is improved by more than one order of magnitude. Since the self-consistent field generated by the electrons is ignored in the fast method and it is considered in CST PS, the results validate that the self-consistent field can be ignored in calculating the multipactor threshold. Finally, taking for example a parallel plate transmission line and a stepped impedance transformer, we study the effect of the number of initial macro-particles on the calculation precision. When the initial particles are so few that they can hardly reflect the randomness of the multipactor process, a higher calculated value will be resulted in. Wi