短时傅里叶变换结合DRSN的滚动轴承故障诊断研究

Fault diagnosis of rolling bearings based on short-time Fourier transform and DRSN

针对滚动轴承在复杂噪声环境下故障分类困难等问题,文章提出一种短时傅里叶变换(STFT)和深度残差收缩网络(DRSN)相结合的轴承故障诊断方法。首先利用短时傅里叶变换对滚动轴承原始振动信号进行时域频域处理得到信息更丰富的故障时频图样本,分为训练集和测试集;将软阈值模块引入到深度残差网络残差块中,其中的残差连接和软阈值模块能够滤除噪声并提取样本特征中的有效信息,输出到分类器上完成端对端的高准确率轴承故障分类。为验证所提方法的可行性,将该方法与其他模型作对比,实验结果表明,该方法在强噪声干扰下能表现出较高的分类性能,稳定性优于其他模型。

In response to problems such as failure classification of rolling bearing in a complex noise environment,a bearing fault diagnosis method combined with a short-term Fourier transformation(STFT) and deep residual shrinkage network(DRSN) is introduced in this paper.First,the short-time Fourier transform is used to process the original vibration signal of the rolling bearing in the time domain and frequency domain to obtain fault time-frequency diagram samples with richer information,which are divided into the training set and the test set.The soft-threshold module is introduced into the residual block of the deep residual network.The residual connection and the soft-threshold module in it can filter out the noise and extract the effective information in the sample features,and output it to the classifier to complete the end-to-end bearing fault classification with high accuracy.To verify the feasibility of the method,this method is compared with other models.The experimental results show that the method can show high classification performance under complex noise interference,and the stability is better than other models.