基于遗传粒子群融合的射频电源阻抗匹配方法  

Genetic Particle Swarm Optimization-based Impedance Matching Method for RF Power Amplifiers

作　　者：蔡馨慧 易志强[1] 李光俊 CAI Xinhui;YI Zhiqiang;LI Guangjun(School of Communication Engineering,Hangzhou Dianzi University,Hangzhou 310018,China)

机构地区：[1]杭州电子科技大学通信工程学院,浙江杭州310018

出　　处：《杭州电子科技大学学报（自然科学版）》2024年第4期44-50,共7页Journal of Hangzhou Dianzi University：Natural Sciences

摘　　要：在实际工作中,射频电源的负载特性会发生快速变化,导致阻抗失配,使得电源输出功率无法始终处于额定值。为克服该现象,须对时变负载完成实时阻抗匹配。本文提出一种基于遗传粒子群融合的射频电源阻抗匹配方法。首先对代表阻抗匹配网络中可变元器件参数的粒子进行交叉、变异操作,增加粒子多样性;同时引入非线性递减的惯性权重和动态学习因子,根据粒子寻优情况调节搜索范围,只需少量迭代即可求得阻抗匹配网络参数的最优解,提高了寻优速度。仿真结果表明,该融合算法在阻抗匹配实时性及精度方面均优于传统粒子群算法和遗传算法。In practical work,the load characteristics of RF power sources change rapidly,which leads to impedance mismatch and the output power of the power source cannot always be at the rated value.To overcome this phenomenon,real-time impedance matching of time-varying loads is necessary.This paper proposes a genetic particle swarm optimization-based impedance matching method for RF power sources.Firstly,particles representing the variable component parameters in the impedance matching network are subjected to cross-over and mutation operations to increase particle diversity.At the same time,nonlinearly decreasing inertia weight and dynamic learning factors are introduced to adjust the search range according to the particle optimization situation,and the optimal solution of the impedance matching network parameters can be obtained with only a small number of iterations,improving the optimization speed.The simulation results show that the fusion algorithm is superior to traditional particle swarm algorithms and genetic algorithms in terms of impedance matching real-time performance and accuracy.

关 键 词：射频电源 阻抗匹配 遗传粒子群融合 动态学习因子 非线性递减惯性权重 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]