基于深度学习与支持向量机的滚动轴承故障诊断研究

Research on Rolling Bearing Fault Diagnosis based on Deep Learning and Support Vector Machine

针对滚动轴承运行环境复杂,传统故障诊断方法难以从强非线性信号中提取有效故障特征,且无法充分利用信号自身特征的问题,提出CNN-LSTM-SVM故障诊断方法。以滚动轴承加速度寿命实验数据为研究对象,基于卷积神经网络(Convolutional Neural Network,CNN)与长短期记忆网络(Long Short Term Memory,LSTM)技术提取信号特征并结合支持向量机(Support Vector Machine,SVM)完成故障分类。结果显示:该方法具有良好外推性能,在变演变阶段下的平均准确率达到95.92%,与现有方法相比,至少高出11.34%,且在噪声环境下的诊断准确率均高于现有方法,稳定性更佳,体现良好的鲁棒性与泛化性。

Considering the complex operating environment of rolling bearings,the traditional fault diagnosis methods are difficult to extract effective fault features from strong nonlinear signals and cannot make full use of the characteristics of the signals themselves.The signal features were extracted based on convolutional neural network(CNN)and long short term memory(LSTM)technology,and the classification was completed with support vector machine(SVM),so as to propose the fault diagnosis method of CNN-LSTM-SVM.Taking the experimental data of rolling bearing acceleration life as the research object,the CNN-LSTM-SVM method was used to analyze and diagnose the bearing failure.The results show that the proposed method has good extrapolation,and its average accuracy is 95.92%in the phase of transformation and evolution,which is at least 11.34%higher than that of the existing methods.In addition,the diagnostic accuracy of the proposed method is higher than that of the existing methods in the noise environment,and the stability is better,showing good robustness and generalization.