基于改进梅尔倒谱系数的GIS机械故障诊断方法  

Mechanical Fault Diagnosis Method of GIS Based on Improved MFCC

作　　者：徐明月 李喆[1] 孙汉文 盛戈皞[1] 江秀臣[1] XU Mingyue;LI Zhe;SUN Hanwen;SHENG Gehao;JIANG Xiuchen(Department of Electrical Engineering,Shanghai Jiao Tong University,Shanghai 200240,China)

机构地区：[1]上海交通大学电气工程系,上海200240

出　　处：《高压电器》2020年第9期122-128,共7页High Voltage Apparatus

摘　　要：机械故障是GIS常见的故障,若不及时发现会造成分合闸失误等重大安全隐患。文中提出了一种用于GIS机械故障在线监测的基于改进梅尔倒谱系数诊断方法。首先对预处理后的声音信号提取MFCC;为适应GIS运行声音能量变化平缓的特点,对MFCC进行优化得到改进特征;引入SVM构建基于声学的GIS机械故障诊断模型,并采用袋装算法对SVM模型进行集成。本研究通过在真型GIS上模拟机械故障,获取真实的故障声音信号进行训练和测试。实验结果表明,改进MFCC相较于传统MFCC在GIS故障声音识别系统中有着更高的识别精度。并且对比传统MFCC特征,改进的特征在噪声条件下也有更好的表现,尤其在信噪比低时,F1分数提升幅度可以达到30%左右。Mechanical fault is a common fault of gas insulated switchgear(GIS).If it is not found in time,it will cause major safety hazards such as opening and closing fault.In this paper,a diagnosis method based on improved mel-frequency cepstrum coefficients(MFCC) for on-line monitoring of GIS is proposed.Firstly,MFCC features are extracted from the preprocessed sound signals.Due to the energy of GIS sound changing gently,the mid-term MFCC features are obtained by optimizing MFCC.Support vector machine(SVM) is introduced to build an acoustic-based mechanical fault diagnosis model of GIS,and the Bagging algorithm is used to integrate the SVM models.In this study,mechanical faults are simulated on real GIS to obtain real fault sound signals for training and testing.The experimental results show that,compared with MFCC,mid-term MFCC has better recognition results in sound recognition system of GIS.Compared with the traditional MFCC features,the improved features also have better performance under noise conditions,especially when the signal-to-noise ratio is low,the F1 score can be increased by about 30%.

关 键 词：气体绝缘组合电器(GIS) 机械故障 故障诊断 梅尔倒谱系数 说话人识别 

分 类 号：TM595[电气工程—电器]