电力变压器故障的客观熵权识别及诊断方法

发布时间：2018-06-08 07:05

本文选题：系统工程 + 信息模式识别　；参考：《电力系统自动化》2017年12期

【摘要】：为了有效管理和监测电力变压器的健康状态,在对变压器油中溶解气体数据进行分析的基础上,建立了一种基于客观熵权的电力变压器故障信息模式识别及诊断模型。首先,在定义包含电力变压器故障模式全局信息的矩阵范式基础上,引入信息熵权理论构建故障特征信息的客观熵权精确量化模型;然后,基于距离和投影原则构建了故障模式判别准则函数,并通过准则函数对模式进行排序,运用综合排序结果进行故障测试模式分类,得到用于判断故障类型的基准类心向量;最后,运用基于类心欧氏距离的方式判别故障测试样本所属的类别,实现变压器故障的客观熵权识别及诊断。利用从某电力公司采集到的120组电力变压器油中溶解气体样本进行实例验证,结果表明,所提出的方法能克服传统的三比值故障诊断方法存在无编码以及边界编码模糊致误判的问题。
[Abstract]:In order to effectively manage and monitor the health status of power transformer, a fault information pattern recognition and diagnosis model of power transformer based on objective entropy weight is established based on the analysis of dissolved gas data in transformer oil. Firstly, on the basis of defining the matrix normal form that contains the global information of power transformer fault mode, the information entropy weight theory is introduced to construct the objective entropy weight accurate quantization model of fault characteristic information. Based on the distance and projection principle, the fault pattern discriminant criterion function is constructed, and the pattern is sorted by the criterion function, and the classification of the fault test mode is carried out by using the comprehensive sorting result, and the benchmark class heart vector is obtained to judge the fault type. Finally, based on the Euclidean distance, the classification of fault test samples is determined, and the objective entropy weight identification and diagnosis of transformer faults are realized. An example of 120 sets of dissolved gas samples from power transformer oil collected from a power company shows that, The proposed method can overcome the problems of no coding and boundary coding fuzzy error in the traditional three ratio fault diagnosis method.
【作者单位】：重庆交通大学信息科学与工程学院;广东省石化装备故障诊断重点实验室(广东石油化工学院);
【基金】：国家自然科学基金资助项目(61004118) 教育部留学归国人员科研启动基金资助项目(2015-49)~~
【分类号】：TM41

【相似文献】