中文摘要：

针对传统迭代法求解特定谐波消除脉宽调制策略（SHEPWM）开关角方程组需要合适的初值和难收敛的问题,提出一种改进的混合粒子群优化（HPSO）算法对其进行求解.该算法对粒子群（PSO）算法的权重系数和学习因子进行了改进,并且提出了一种温度系数线性递减的模拟退火（SA）算法与粒子群算法结合,有效弥补了传统粒子群（CPSO）算法求解开关角收敛速度慢和精度低的缺点.仿真分析表明,该算法消除了对初值的依赖,提升了算法寻优的能力,从而提高了求解速度与精度,并且通过实验验证了该算法的可行性.

英文摘要：

The requirement of an initial value and the speed of convergence are the main challenges for the traditional iteration method in selective harmonic elimination pulse width modulation（SHEPWM）.As to these problems,an improved hybrid particle swarm optimization（HPSO）algorithm was proposed in this paper.In the proposed strategy,the inertia weight and learning factors of PSO algorithm was adjusted.Meanwhile,a new kind of simulated annealing（SA）algorithm,whose temperature coefficient decreases with the increase of the number of iterations,is put forward to combine with PSO,making a favorable improvement in terms of the convergence speed and accuracy compared with the conventional PSO（CPSO）algorithm.Finally,the viability and performance of the proposed strategy are shown through simulation and experimental results in a laboratory prototype.