基于深度度量学习的轴承故障诊断方法

针对机械大数据因故障类内离散度和类间相似度较大而导致诊断精度低的问题,提出一种深度度量学习故障诊断方法,采用深度神经网络(Deep Neural Network, DNN)对故障特征进行自适应提取,并利用基于欧氏距离的边际Fisher分析(Marginal Fisher Analysis, MFA)方法进行了优选,在构建的深度度量网络(Deep Metric Network, DMN)顶层特征输出层添加BPNN(Back Propagation Neural Network, BPNN)分类器对网络参数进行微调,并实现故障的分类识别。通过对不同类型和严重程度的轴承故障进行了诊断分析,验证了该方法可以有效地对轴承故障进行高精度诊断,效果优于传统深度信念网络(Deep Belief Network, DBN)故障诊断方法以及常用时域统计特征结合支持向量机(Support Vector Machine, SVM)分类的故障诊断方法。

Bearing fault diagnosis method based on deep metric learning

As the intra-class scatter and inter-class similarity are big in bearing fault data, which constrains the diagnostic accuracy, a new method of deep metric learning for fault diagnosis is proposed. The deep neural network (DNN) is used to adaptively extract the fault features, and the marginal Fisher analysis method based on Euclidean distance is used to optimize the features, then, the BPNN classifier is added to the top-level feature output layer of the constructed deep metric network (DMN) to fine tune the parameters and realize fault classification and recognition.  Finally, diagnostic analysis on bearing fault experimental data of different fault types and different fault severity verified that the method can effectively diagnose bearing faults with high precision and the effect is better than traditional deep belief network (DBN) fault diagnosis method as well as the common time-domain statistical features combined with support vector machine (SVM) classification fault diagnosis method.