基于遗传算法的电力电容器宽频建模方法

摘要

A genetic algorithm⁃based wideband modeling method is proposed in this paper. First of all, a modeling idea for the limitations of the vector fitting⁃based modeling method is proposed. Secondly, the current experiment measures the wideband impedance of a capacitor by using vector network analyzer. Thirdly, in order to represent the wide impedance of the measured capacitor, this paper proposes a special model, of multiple parallel RLCG branches in which the number of the branches is variable. Next, the initial solution of genetic algorithm is calculat⁃ed by using the special points of the measured wideband impedance. Then the calculation of the optimal solution af⁃ter multiple iterations helps to establish the model of the measured capacitor and simulate the model in time domain by EMTP. The reasonable result proves that the model is stable and accurate.%提出一种基于遗传算法的电气设备宽频建模方法。首先，针对以往常用的基于矢量匹配法的建模方法的局限性，提出自己的建模思路；其次，利用矢量网络分析仪对某型号电容器的宽频阻抗进行测量；然后，提出特殊的等效电路模型来表征所测量设备的阻抗特性。用多条RLCG支路并联电路等效被测量设备的阻抗特性，且并联支路数可变；利用测量到的阻抗特性中的一些特殊点推导出遗传计算的初始解，经遗传算法多次迭代求出最优解得到所测量电容的宽频模型，利用EMTP进行时域仿真，证明所建模型为稳定模型。