Data Driven Rolling Bearing Fault Diagnosis Research

In the modern industrial production process, rolling bearing is one of the most commonly used parts,especially in rotating machinery. Rolling bearing fault occurs frequently. It brings huge influence to the production process, threatens the safety of the production, and causes quantities of economic losses. Therefore, research on rolling bearing fault diagnosis is very meaningful.The theoretical basis of this paper is power spectrum analysis and data driven method. The paper puts forward using PCA and FDA methods to the study of all kinds vibration signal that they have been power spectrum analysis. The main research works are as follows:1. In this paper, the rolling bearing fault mechanism and vibration characteristics are analyzed in detail,several major bearing fault form and the main fault diagnosis methods have discussed. The method of frequency domain analysis is used feature extraction to bearing vibration signal. And further screen the extracted characteristics by PCA; reduce the dimension of the signal feature vector. It can improve the accuracy and real-time of the fault diagnosis.2. When rolling bearing fault occurs, its vibration signal energy has changed in some frequency. Construct fault diagnosis model with normal data or fault data, use T square statistics and Q statistics method to detect different fault. It comparative analysis the advantages and disadvantages of different data models through the identification accuracy.3. In order to solve for different bearing fault classification problem, according to the characteristics that maximum between-class discrete degrees and minimize within-class discrete degrees, it can classify different position of the fault and the same position different sizes of the fault. In the simulation test classification accuracy is ideal. Simulation results show that the presented method can accurately distinguish bearing normal and fault state, and recognize the position and size of the fault, it can solve the problem of rolling bearing fault diagnosis well.