基于多元模糊联系度模型的变压器油纸绝缘老化评价

Aging State Evaluation of Oil⁃paper Insulation Transformer Based on Multivariate Fuzzy Relation Degree Model

针对变压器油纸绝缘评价指标存在的不确定性和评价标准存在的模糊性,提出结合集对分析与模糊集组合方法,将多元模糊联系度模型应用于变压器油纸绝缘老化状态评价中。首先,基于去极化电流谱线解析提取了3个新的绝缘老化特征量,结合变压器的实测数据进行统计分析并建立变压器9项老化评价指标体系,基于模糊C⁃均值聚类(FCM)算法,确定3等级评价标准;然后,应用同异反分析构建各评价指标与评价等级标准的联系度表达式,引入主客观相结合的组合权重赋权法并对差异度信息进行数据挖掘,确定组合联系度;最后,按最大组合联系度对应的等级作为集对分析结果,实例结果验证了多元模糊联系度模型能有效、准确地评价变压器油纸绝缘老化状态,具有推广应用价值。

In view of the uncertainties existing in the transformer oil⁃paper insulation evaluation index and the ambi⁃guity of the evaluation criteria,a combination of set⁃pair analysis and fuzzy set combination method is proposed to ap⁃ply the multivariate fuzzy connection degree model to the evaluation of transformer oil⁃paper insulation status.First⁃ly,three new features of insulation aging quantities are extracted based on the analysis of depolarization current spec⁃trum,and combined the transformer’s measured data for statistical analysis to establish the nine aging evaluation in⁃dexes system for transformers,and determine the three levels of evaluation criteria based on fuzzy C⁃means clustering(FCM)algorithm.Then,the same⁃and⁃backward analysis is used to construct the connection degree expressions of each evaluation index and the evaluation grade standard,a combination weighting method with subjective weight and objective weight is also introduced and the data mining method is used for difference degree information to deter⁃mine the combination contact degree.Finally,according to the level of the maximum combination degree of connec⁃tion as a set pair analysis results,the example results verify that the multivariate fuzzy connection degree model can effectively and accurately evaluate the transformer oil⁃paper insulation state and has promotion and application value.