NSGA-II Model-Based Dielectric Frequency Response Parameters for Aging and Moisture Analysis of Transformer Insulation
Zheshi Ding, Xianhao Fan, Boshu Song, Hui Hwang Goh, Yiyi Zhang, Jiefeng Liu
Abstract
The complex dielectric frequency response (DFR) behavior is established during the frequency domain spectroscopy (FDS) test, the methods for extracting the feature parameters to distinguish the diverse response process, as well as achieve the state analysis, still deserve variable attention. To address this, a combination of non-dominated sorting genetic algorithm (NSGA-II) and extended Cole-Cole model is presented to extract the required feature parameters. In the current work, the FDS of insulation samples with various states is first collected. Then, the decoupling analysis is performed on the extended Cole-Cole model to distinguish the diverse DFR behavior. The NSGA-II model is employed to solve the extended Cole-Cole model to obtain the required feature parameters. Finally, the insulation state analysis practiced by lab and field trials proves the feasibility and applicability of the extracted feature parameters. Regarding this, this work reports an intelligent optimization model to be responsible for a more accurate and reliable state analysis of transformer insulation via the extracted feature parameters.