基于WPD和ACO-SVM的配电变压器绕组故障辨识方法  

Distribution Transformer Winding Fault Identification Based on Wavelet Packet Decomposition and ACO-SVM

作　　者：欧庆炀 陈志英 刘必兴 张修伦 陈国炎 OU Qingyang;CHEN Zhiying;LIU Bixing;ZHANG Xiulun;CHEN Guoyan(School of Electrical Engineering and Automation,Xiamen University of Technology,Xiamen 361024,China;School of Information Science and Engineering,Huaqiao University,Xiamen 361021,China)

机构地区：[1]厦门理工学院电气工程与自动化学院,福建厦门361024 [2]华侨大学信息科学与工程学院,福建厦门361021

出　　处：《厦门理工学院学报》2025年第1期33-41,共9页Journal of Xiamen University of Technology

摘　　要：为有效提高配电变压器绕组故障辨识准确率,提出一种基于小波包分解(WPD)与ACO-SVM模型的配电变压器绕组故障辨识方法。该方法先采用小波包分解配电变压器振动信号,提取信号的频段能量占比、峰峰值等6个特征值,然后采用蚁群算法(ACO)优化支持向量机模型(SVM)参数选择,最后利用ACO-SVM模型对特征值进行绕组故障分类辨识。对某500 VA油浸式配电变压器样机不同绕组故障状态下振动信号的采集和辨识结果表明,该方法可正确提取配电变压器振动信号特征值,对绕组故障状态的辨识准确率为96.2%,相比于ELM、GRNN、PNN方法准确率分别提高10.2%、7.8%、9.6%。In order to improve the accuracy of winding fault identification of distribution transformer,a fault identification method of distribution transformer winding based on wavelet packet decomposition(WPD)and ACO-SVM model is proposed in this paper.In this method,WPD is used to extract the vibration signals of distribution transformers,and six characteristic values including frequency band energy ratio and peak-to-peak value of the signals extracted.Next,ant colony algorithm(ACO)is used to optimize the parameter selection of support vector machine(SVM),and the ACO-SVM model is then used to classify winding faults.Acquisition and identification of vibration signals of a 500VA oil-immersed distribution transformer prototype under different winding fault states is tested.The results show that the proposed method can correctly extract the vibration signal characteristic values and realize the effective identification of winding fault states with an accuracy of 96.2%,surpassing by 10.2%,7.8%and 9.6%respectively compared with ELM,GRNN and PNN classification methods.

关 键 词：配电变压器 故障辨识 振动信号分析 小波包分解 支持向量机模型 蚁群算法 

分 类 号：TM411.2[电气工程—电器]