基于IFOA-SVM的变压器故障诊断与定位研究  

Study of Transformer Fault Diagnosis and Location Based on IFOA-SVM

作　　者：申文强 王宗琳 樊尚旭 郭永吉[2] SHEN Wenqiang;WANG Zonglin;FAN Shangxu;GUO Yongji(State Grid Gansu Electric Power Extra High Voltage Company,Lanzhou 730070,China;College of Electrical and Information Engineering,Lanzhou University of Technology,Lanzhou 730000,China)

机构地区：[1]国网甘肃省电力公司超高压公司,甘肃兰州730070 [2]兰州理工大学电气工程与信息工程学院,甘肃兰州730000

出　　处：《电工技术》2023年第12期46-50,共5页Electric Engineering

摘　　要：为了充分发掘变压器油中溶解气体所蕴含的故障信息,实现变压器故障性质和故障位置的准确判断,利用邻域粗糙集(Neighborhood Rough Set, NRS)和强化型果蝇算法(Improved Fruit Fly Optimization Algorithm, IFOA)优化的支持向量机(Support Vector Machine, SVM)构建变压器故障诊断与定位多层分类模型。首先,利用邻域粗糙集按照条件属性重要度对变压器故障样本特征值进行约简。其次,为了提升变压器故障诊断与定位模型的分类精度,设计一种强化型果蝇算法对SVM的核函数参数和惩罚因子选取进行优化。利用Tent-logistic混沌映射产生的混沌序列生成果蝇种群的初始位置信息,减少随机过程带来的不可控性;利用动态自适应步长参数调节个体的搜索范围,增强FOA的寻优效率。仿真分析结果表明,基于改进模型的方法不仅可以实现变压器故障位置的判定,而且能提升变压器故障诊断的精度。In order to fully explore the fault information contained in the dissolved gas in transformer oil and realize the accurate judgment of transformer fault nature and fault location,a multi-layer classification model for transformer fault diagnosis and location is constructed using the support vector machine(SVM)optimized by neighborhood rough set(NRS)and enhanced fruit fly optimization algorithm(IFOA).Firstly,the neighborhood rough set was used to reduce the characteristic values of transformer fault samples according to the importance of conditional attributes.Secondly,in order to improve the classification accuracy of transformer fault diagnosis and positioning model,an enhanced fruit fly algorithm is designed to optimize the kernel function parameters and penalty factor selection of SVM.The initial position information of chaotic sequence fruit fly populations generated by Tent-logistic chaos mapping was used to reduce the uncontrollability caused by random processes.The dynamic adaptive step parameter was used to adjust the search range of individuals to enhance the optimization efficiency of FOA.Simulation analysis shows that the method based on the improved model can not only determine the fault location of transformers,but also improve the accuracy of transformer fault diagnosis.

关 键 词：变压器 故障诊断 故障定位 果蝇优化算法 支持向量机 

分 类 号：TM07[电气工程—电工理论与新技术]