摘要
对电机轴承进行有效的故障诊断,不仅可以保证设备的平稳高效运行,而且可以及时发现和排除运行故障,防止重大事故发生。本文提出一种基于小波包分解(WP)和堆叠自动编码机(SAE)的故障诊断方法,充分利用了WP时频信号分析和自动编码器无监督地从数据中学习信号特征的能力。仿真结果表明:所提出的方法在简化网络结构的基础上,提高了故障识别率。
Effective fault diagnosis of motor bearings can not only ensure the smooth and efficient operation of equipment,but also identify and eliminate operation faults in time to prevent major accidents.In this arti⁃cle,a fault diagnosis method based on wavelet packet decomposition(WP)and stacking automatic encoder(SAE)is proposed,which makes full use of WP time-frequency signal analysis and the ability of automatic encoder to learn signal characteristics from data unsupervised.The simulation results show that the proposed method improves the fault identification rate on the basis of simplifying the network structure.
作者
张军亮
刘欣
贺文博
胡树云
胡振亚
Zhang Junliang;Liu Xin;He Wenbo;Hu Shuyun;Hu Zhenya(CHN Energy Jinjie Energy Co.,Ltd.,Shenmu,Shaanxi 719300;Hangzhou Hollysys Automation Co.,Ltd.,Hangzhou,Zhejiang,310018)
出处
《能源科技》
2021年第5期52-55,共4页
Energy Science and Technology
关键词
故障诊断
电机轴承
时频信号分析
深度学习
自动编码器
fault diagnosis
motor bearing
time-frequency signal analysis
deep learning
automatic encoder
