Big Vibration Data Diagnosis of Bearing Fault Base on Feature Representation of Autoencoder and Optimal LSSVM-CRO Classifier Model

In this paper, based on deep learning method for the high-dimensional feature representation of vibration signal and optimal machine learning model, a new diagnosis technique is proposed for identifying the big vibration data of multi-level fault of roller bearing. Firstly, a deep learning network based on stacked autoencoders (SAE) with hidden layers is exploited for vibration feature representation (VFR) of roller bearing data, named as VFR-SAE, in which the unsupervised learning algorithm is used to reveal the significant properties in the data such as nonlinear, non-station properties. The represented features can provide good discriminability for fault diagnosis task. Secondly, an optimal classifier model based on least square support vector machine (LSSVM) classifier and chemical reaction optimization (CRO) algorithm, named as LSSVM-CRO, is formed to perform supervised fine-turning and classification. In this work, the transfer learning performance can be especially tuned in this diagnosis technique. That is, the features of target vibration data will be extracted by the learning of feature representation which depends on the weight matrix of hidden layers of VFR-SAE method. The experimental results by analyzing the roller bearing vibration signals with multi-status of fault have demonstrated that the VFR-SAE based feature extraction in conjunction with the LSSVM-CRO classifier model can achieve higher accuracies than the other popular classifier models.