基于区间-粒子群算法的油纸绝缘等效电路参数计算

Parameter calculation for equivalent circuit of oil-paper insulation based on interval-particle swarm optimization algorithm

回复电压法是一种研究变压器绝缘老化状态的非破坏性检测方法。现阶段,为了从理论上分析介质极化特性,许多学者通过建立等效电路模型来辅助分析油纸绝缘老化机理。因此,准确地辨识等效电路参数对后续研究油纸绝缘老化起到至关重要的作用。本文使用初始斜率特征量辨识介质响应等效电路参数的数学优化模型,其计算简单,操作方便。考虑到实际测试中充电电压的波动和测量误差会带来参数计算的误差,本文首次提出采用区间-粒子群算法对等效电路模型进行区间参数计算。区间-粒子群算法在高维多峰值函数中不仅有较好的执行效率,而且能够搜索到最优可行区间。最后通过仿真实例验证了等效电路模型的准确性和区间-粒子群算法进行参数计算的适用性。该研究结果为进一步评估油纸绝缘老化奠定了良好的基础。

Recovery voltage method （RVM） is an approach of non-destructive measurement for diagnosing insula- tion of power transformer. At the present stage, in order to analyze the polarization properties of the medium in the- ory, the scholars build equivalent circuit models to analyze oil-paper insulation aging mechanism. Therefore, the accuracy of equivalent circuit parameters plays a crucial role in the follow-up study of aging oil-paper insulation. The paper establishes a mathematical model with the data of initial slope of return voltage for solving parameter of dielectric response equivalent circuit, which is simple, and easy to operate. Considering charging voltage error and measurement error, the paper firstly adopts Interval-Particle Swarm Optimization Algorithm （I-PSO） to calculate model parameters. I-PSO has higher execution efficiency and better optimization performance in high-dimensional model. Finally, simulation experiment results have shown the accuracy of equivalent circuit and the effectiveness of this algorithm. The results of this study have laid a good foundation for further evaluation of oil paper insulation aging.