电力信息与通信技术2025，Vol.23Issue(7)：38-45,8.DOI:10.16543/j.2095-641x.electric.power.ict.2025.07.05
变压器油色谱诊断中多种改进比值的综合方法
An Integrated Approach to Multiple Improved Ratios in Transformer Oil Chromatographic Diagnostics
摘要
Abstract
Power transformers are indispensable key equipment in the power system,and accurate and reliable fault diagnosis is required to ensure the safe and stable operation of the power grid.There may be contradictions in the diagnostic results between different ratio methods,while data-driven diagnostic methods have a high demand for data quality and poor explainability.For this reason,this paper proposes a comprehensive method of multiple improved ratio methods,combining historical cases to maximize the advantages of various types of ratios according to specific situations.Firstly,based on multiple improved ratio methods,fuzzy processing is introduced to calculate the respective fault type affiliation and expand it;secondly,neural networks are used to analyze the gas content data and the historical case text of the equipment,and the weights of the ratio methods are calculated;finally,the final fault probability is calculated based on the weights and the fault mode affiliation of each type of ratio method.According to the experimental validation,it is shown that the fault diagnosis method proposed in this paper can effectively provide the equipment diagnosis results,and improve the F1 value by 8.21%to 24.25%compared with the traditional fault diagnosis method.关键词
油色谱/神经网络/故障诊断/自然语言处理Key words
oil chromatography/neural networks/fault diagnosis/natural language processing分类
信息技术与安全科学引用本文复制引用
张贲,王超,刘健,宫本辉..变压器油色谱诊断中多种改进比值的综合方法[J].电力信息与通信技术,2025,23(7):38-45,8.基金项目
国家电网有限公司科技项目资助"关联缺陷故障模式的特高压变压器健康评估和检修决策关键技术研究"(520101230003). （520101230003）
