基于细菌觅食改进SVM的变电站智能检测方法  

Intelligent detection method of substation based on big data technology

作　　者：金海川 陈佳雪 张磊 王东方 JIN Haichuan;CHEN Jiaxue;ZHANG Lei;WANG Dongfang(State Grid Wuzhong Power Supply Company,Wuzhong,Ningxia 751100,China)

机构地区：[1]国网吴忠供电公司,宁夏吴忠751100

出　　处：《自动化与仪器仪表》2021年第11期52-55,共4页Automation & Instrumentation

基　　金：青年科学基金项目,编号:51505055。

摘　　要：针对传统SVM算法在变电站变压器故障诊断中存在准确率不高的问题,在大数据环境下,结合细菌觅食算法和SVM算法,提出一种基于细菌觅食算法改进SVM的电力变压器的故障诊断模型。对此,在分析细菌觅食算法和SVM算法原理基础上,提出电力变压器故障的SVM模型;然后利用细菌觅食算法对SVM算法的参数C和γ进行寻优,并将得到的最优参数作为训练测试集的模型,对变压器故障测试数据进行测试诊断;最后,以Matlab2016a作为仿真工具,以某电力公司采集到的数据作为来源,在数据预处理基础上,分别对比PSO、GA与BFA优化SVM的优化结果,以及未经优化的故障诊断结果。结果表明,本研究提出的BFA改进模型的准确率达90%以上,高于传统的SVM和其他三种优化方法。Aiming at the problem that the traditional SVM algorithm is not accurate in transformer fault diagnosis of substation,in the big data environment,combining the bacterial foraging algorithm and SVM algorithm,a power transformer fault diagnosis model based on the bacterial foraging algorithm and the improved SVM is proposed.On the basis of analyzing the principle of bacterial foraging algorithm and SVM algorithm,the SVM model of power transformer fault is proposed.Then,the bacterial foraging algorithm is used to optimize the parameter C of SVM algorithm,and the optimal parameters obtained are taken as the model of the training test set to test and diagnose the transformer fault test data.Finally,Matlab2016a was used as the simulation tool,and the data collected by a power company was used as the source.On the basis of data preprocessing,the optimization results of PSO,GA and BFA optimization SVM were compared,as well as the fault diagnosis results without optimization.The results show that the accuracy of the improved BFA model proposed in this study is more than 90%,which is higher than the traditional SVM and the other three optimization methods.

关 键 词：细菌觅食算法 智能检测 变电站 

分 类 号：TP29[自动化与计算机技术—检测技术与自动化装置] TM631[自动化与计算机技术—控制科学与工程]