基于LS-QPSO的单相LCL逆变器参数辨识研究

Single-phase LCL Inverter Parameter Identification Research Based on LS-QPSO

电子电路参数辨识方法常采用递推最小二乘(RLS)算法,但RLS算法往往有较长的暂态过程,待估参数之间的耦合也会降低辨识精度。这里通过将预测时刻顺延的方法改进算法的优化目标,从而提高了算法的灵敏度和抗干扰能力。通过对量子粒子群优化(QPSO)算法进行深入研究,将优化目标与QPSO算法相结合提出最小二乘(LS)-QPSO算法,算法具有良好的并行性,建立了单极倍频调制方式的基于LCL滤波器的单相并网逆变器的近似离散化混杂系统模型,并将两种算法应用于电路参数辨识,在实际应用中验证了所提算法的低暂态过程和高精确度的辨识能力。

The recursive least square(RLS)algorithm is often used in the electronic circuit parameter identification method,but the RLS algorithm often has a long transient process.The coupling between the parameters to be estimated also reduces the identification accuracy.The optimization goal of the algorithm is improved by the method of delaying the prediction time,which improves the sensitivity and anti-interference ability of the algorithm.Through the indepth study of quantum particle swarm optimization(QPSO)algorithm,the optimization target and QPSO algorithm are combined to propose least square(LS)-QPSO algorithm.The algorithm has good parallelism.The single-phase grid-connected inverter based on LCL filter with unipolar frequency modulation is established.The discretized hybrid system model is applied,and the two algorithms are applied to the circuit parameter identification.The low transient process and high accuracy identification ability of the proposed algorithm are verified in practical applications.