基于异常数据清洗与混沌分析的变压器绝缘老化机理研究

Research on aging mechanism of transformer insulation based on abnormal data cleaning and chaos analysis

由于变压器绝缘老化过程非常复杂,对其还需从多种角度进行深入研究。文章通过研究变压器的油中溶解气体色谱数据,进一步探索变压器的绝缘老化特性。由于变压器油色谱数据易包含异常干扰数据,这将影响到后续对变压器绝缘老化机理的研究,因此采用一种基于迭代检验算法的异常数据清洗方法来有效识别并修正油色谱原始数据中的异常数据。然后将变压器绝缘老化过程分为10个阶段,从阶段1~阶段10,其老化程度不断加深。采用混沌分析方法提取出各绝缘老化阶段下,变压器油色谱数据的混沌特征量,深入研究了变压器油色谱数据的混沌特性与绝缘老化发展阶段的关系。

Due to the complexity of the transformer insulation aging process,it is necessary to conduct in-depth research from various angles.This paper further explores the insulation aging characteristics of transformers by studying the chromatographic data of dissolved gases in the oil of transformers.Since the chromatographic data of transformer oil is easy to contain abnormal interference data,this will affect the subsequent research on the aging mechanism of transformer insulation.Therefore,an anomaly data cleaning method based on iterative test algorithm is used to effectively identify and correct abnormal data in the original data of oil chromatography.Then,the transformer insulation aging process is divided into ten stages.From stage one to ten,the aging degree is deepened.The chaotic analysis method is used to extract the chaotic characteristics of the transformer oil chromatographic data under the insulation aging stage,and the transformer oil chromatography is deeply studied.The relationship between the chaotic characteristics of the data and the development stage of insulation aging.