基于机器学习的变压器油色谱分析监测系统研究

打开文本图片集

中图分类号：TM41；TQ646.1 文献标志码：A 文章编号：1001-5922（2025）06-0180-04

Research on transformer oil chromatographic analysis and monitoring system based on machine learning

WANG Haiyan，BO Bo，LI Meng，CAO Jingli，LI Siyuan，LIANG Jianing （State Grid Jibei Electric Power Co.，Ltd.，Chengde Power Supply Company，Chengde ，Hebei China）

Abstract：To improve the recognition accuracy of abnormal data of transformer in digital power supply station，a de⁃ viation identification and system construction of transformer oil chromatographic monitoring device in digital power supply station based on gas correlation analysis is proposed. Firstly，according to the time series change of dissolved gas in transformer oil chromatographic monitoring device，the sliding window algorithm is used to fit the data. Then， in order to realize data symbolization，the k- means algorithm is used to cluster the gas data. Finally，the Apriori al⁃ gorithm is used to analyze the gas correlation，and the data is segmented to refine the data interval，so as to realize the deviation identification and positioning of the transformer oil chromatographic monitoring device. The results of the example analysis show that the proposed method can effectively track the group deviation of the transformer oil chromatographic monitoring device in the digital power supply station，and has certain validity and accuracy.

Key words：transformer；oil chromatogram；correlation analysis；deviation identification

变压器油色谱监测装置是数字供电所重要的监测管理设备，对数字供电所日常运行具有重要意义。（剩余5159字）