摘要
针对旋转机械滚动轴承在恶劣工作环境中易于损坏,而目前故障诊断过于依赖人工特征提取的问题,提出了一种基于小波包分解和卷积神经网络(CNN)的滚动轴承故障诊断方法(WPDEC-CNN):通过小波包分解对振动时域信号进行处理,获得表征信号相似的小波系数,再将其进行预处理后输入CNN进行分类识别.试验结果表明,WPDEC-CNN的损失率低于BP神经网络和CNN,为0.1089;WPDEC-CNN的故障分类准确率均高于BP神经网络和CNN,达到97.3%,验证了所提故障诊断方法的有效性.
Aiming at the problems that the rolling bearing of rotating machinery is liable to be damaged in bad working environment,and the fault diagnosis at present relies too much on manual feature extraction,a rolling bearing fault diagnosis method based on wavelet packet decomposition and convolutional neural network(WPDEC-CNN)was proposed.This method used wavelet packet decomposition to process the vibration time-domain signal to obtain wavelet coefficients that characterized the signal similarly,and then preprocessed them and input them into CNN for classification and recognition.The experimental results showed that the loss rate of WPDEC-CNN was 0.1089,which was lower than that of the BP neural network and the CNN.The fault classification accuracy of WPDEC-CNN reached 97.3%,higher than that of the BP neural network and the CNN,which verified the effectiveness of the proposed fault diagnosis method.
作者
楼剑阳
南国防
宋传冲
LOU Jianyang;NAN Guofang;SONG Chuanchong(School of Energy and Power Engineering,University of Shanghai for Science and Technology,Shanghai 200093,China)
出处
《轻工学报》
CAS
北大核心
2021年第3期79-87,共9页
Journal of Light Industry
基金
国家自然科学基金项目(91852117)。
关键词
小波包分解
卷积神经网络
故障诊断
滚动轴承
wavelet packet decomposition
convolutional neural network
fault diagnosis
rolling bearing
