Fault Diagnosis Of Motor Rolling Bearing Based On Current And Vibration Signal

As one of the important electrical equipment, motors are widely used in manufacture, industry, transportation, and other areas. Rolling bearing is an important part of motor, and the quality of rolling bearings has direct influences on the motor, and even the performance of whole working system. In the motor system, bearings make the rotational movement and hold axial loads, the poor working conditions lead to bearing failure accounts for about40percent of motor failure. Therefore, accurate real-time monitoring and fault detection is of great significance, but also bearing fault diagnosis technology research highlight.The paper analyzes the failure mechanism and several common type of fault from the structure of the bearing, taking the fault characteristic frequency as the main criteria for bearing fault diagnosis based on the current signal, vibration signal and the fusion of both. The main contents are as following:(1) Analyse the relationship between the current signal characteristics and bearing vibration, and the amplitude recovery method based on the current signal is carried out to filter out the fundamental component of the current signal, thus highlighting the fault harmonic.(2) Time domain and frequency domain analysis are used for vibration signal. By selecting IIR butterworth filter and using of Hilbert transform for envelope, resonance demodulation based on vibration signal is achieved for bearing fault diagnosis, which can effectively diagnose the bearing fault occurrence and its location.(3) It emphasizes the signal cross-correlation analysis theory. Fusing harmonic spectrum of current signal and envelope spectrum of vibration signal, realizes the multi-source information fusion bearing fault diagnosis.(4) Built a test platform, including current, vibration sensor and PXI series devices produced by NI. Various signal processing programs are written by Lab VIEW. Experiment with artificial damaged bearing, and compare with the normal bearing. The results prove that the methods mentioned above are feasible.Based on the theoretical analysis, the research team designs an information collection interface circuit using STM32F3processor and sent the collected information to host PC. A software is programmed and by collecting data when the motor is running on-line monitoring of the motor bearings is achieved.