摘要
深度学习以其强大的特征提取能力展现了它在故障诊断领域的绝对优势。为此,提出了一种基于EMD和SSAE的滚动轴承故障诊断方法。首先采用EMD方法分析滚动轴承振动信号,并用得到的IMF构造Hankel矩阵,获得能反映信号特征的奇异值;其次将奇异值划分为训练集与测试集样本,建立基于SSAE方法的故障诊断模型;最后训练与测试搭建的深度神经网络,得到诊断准确率。所提方法不需要大量的故障诊断先验知识,无需对信号去噪处理,简化了滚动轴承故障诊断的特征提取过程,具有较高的故障诊断准确率。
Deep learning has demonstrated its absolute superiority in the field of fault diagnosis with its strong feature extraction ability in recent years. In this paper, a fault diagnosis method based on EMD and SSAE is proposed. Firstly, EMD method is used to analyze the vibration signal of the rolling bearing, and the Hankel matrices is constructed by the obtained IMFs, then the singular values representing the characteristics of the signal is obtained. Secondly, the samples are divided into a training set and a test set, and a fault diagnosis model based on SSAE method is established. Finally, after training and testing the deep neural network, the diagnostic accuracy can be obtained. The proposed method doesn't need a lot of prior knowledge of fault diagnosis and even signal denoising processing, simplifying the feature extraction of rolling bearing fault diagnosis, which also has a higher accuracy of fault diagnosis.
作者
王奉涛
邓刚
王洪涛
于晓光
韩清凯
李宏坤
WANG Feng-tao;DENG Gang;WANG Hong-tao;YU Xiao-guang;HAN Qing-kai;LI Hong-kun(School of Mechanical Engineering, Dalian University of Technology, Dalian 116024, China;School of Mechanical Engineering and Automation, University of Science and Technology Liaoning,Anshan 114051, China)
出处
《振动工程学报》
EI
CSCD
北大核心
2019年第2期368-376,共9页
Journal of Vibration Engineering
基金
国家自然科学基金资助项目(51375067
51775257)
关键词
故障诊断
滚动轴承
经验模态分解
自动编码器
奇异值分解
fault diagnosis
rolling bearing
empirical mode decomposition
autoencoder
singular value decomposition
