Design Of DSP Rolling Bearing Condition Monitoring And Fault Diagnosis System

Rolling bearings are widely used in rotating machines as general mechanical components. The rolling bearing running status directly affects the entire machine's performance. And the rolling bearing's real-time condition monitoring and fault diagnosis is an important approach to guarantee the operating safety for some key equipment.In this thesis, a rolling bearing condition monitoring and fault diagnosis system based on DSP is presented, which can be separated from the upper computer's control and run independently. The system can monitor rolling bearing's state information such as temperature, rotate speed and vibrations. The applied fault-diagnosis method combines wavelet packet analysis with BP Neural Network, and the operation status information and diagnosis results are shown on the LCD monitor.For the hardware of the system, TI's high performance 32-bit digital signal processor TMS320LF2407A is selected as core CPU. The AD conversions are extended up to 16 channels, and LCD circuit and signal conditioning circuits for acceleration signals, temperature signals and rotate speed signals are designed.The fault diagnosis algorithms applied to DSP is wavelet packet decomposition and BP Neural Network. Using this algorithm, the features of vibration signals can be accurately extracted. And then the energy of each band can be calculated and normalized to obtain fault feature vectors. Finally, BP neural network is adopted for fault identification. Through the algorithms, the intelligent fault diagnosis of rolling bearing can be realized, which can reduce the depence on professional testers and can facilitate industrial field application.Based on the hardware circuit design and fault diagnosis algorithm, each functional module programming and debugging are completed. To confirm the wavelet packet algorithm, firstly, the algorithm is realized on Matlab platform successfully, and then it is debugged on DSP using C language. This method can reduce the difficulty of software development.Through experiments and debugging, the DSP intelligent fault diagnosis system is verified, and real-time performance of the system is analyzed. The system can operate independently from the computer to meet the requirements of field monitoring and diagnosis.