基于改进型多目标粒子群算法的光伏并网逆变器LCLLC滤波器参数优化方法

Parameter Optimization Method for LCLLC Filter of Photovoltaic Grid Connected Inverter Based on Improved Multi-objective Particle Swarm Optimization Algorithm

LCLLC型滤波器同时具有LCL和LLCL型滤波器的优点,即高频谐波衰减效果好,又可以滤除开关频率处的谐波,其性能主要取决于参数的设计。然而,LCLLC型滤波器参数设计复杂,传统设计方法过于保守且成本较高,单目标算法和传统的多目标优化算法存在权重选择困难、收敛性差和优化效果不佳的问题。通过对LCLLC滤波器的谐波进行分析,提出一种多目标参数优化方法,将纹波电流最小化、谐波电流衰减比最小化、总成本最小化作为优化目标,利用提出的三方竞争纵横交叉多目标粒子群算法,得到LCLLC型滤波器的最优参数组合,并利用Matlab/Simulink软件对光伏并网系统中的LCLLC型滤波器进行仿真。结果表明,设计方案正确、有效。

The LCLLC filter has the advantages of both LCL and LLCL filters,with good high-frequency harmonic attenuation effect and the ability to filter out harmonics at the switching frequency,and its performance mainly depends on the design of parameters.For the LCLLC filter,the parameter design is complex,and the traditional design methods are too conservative with high cost.Single objective algorithm and traditional multi-objective optimization algorithm have problems such as difficulty in weight selection,poor convergence and optimization effect.An analysis was conducted on the harmonics of LCLLC filters,and a multi-objective parameter optimization method was proposed.The optimization objectives were to minimize ripple current,harmonic current attenuation ratio,and the total cost.The optimal parameter combination of the LCLLC type filters was obtained using the proposed triple competitive crossbar multi-objective particle swarm optimization algorithm.Finally,the LCLLC type filter in the photovoltaic grid connected system was simulated using Matlab/Simulink software.The results demonstrate the correctness and effectiveness of the designed scheme.