Dsp-based Rotating Equipment Failure Monitor

With the rapid development of modern manufacture and the advancement of science and technology, modern devices are becoming more and more complex, degree of automation is higher. But on the other hand, the expenses for repairing and maintenance are ascending. The domestic market of instrument for inspecting is occupied by foreign enterprises. So it is important for us to have independent property right and kernel technology of high quality malfunction-detecting instrument. In this paper, on the base of investigation, the scheme is discussed for existing vibration signal analysis and mechanical malfunction diagnosis. The following work is completed: 1.Hardware design for the instrument In this paper, TMS320VC33 is choosen as core processor for the malfunction detector and a multi-channel information gathering and processing system is built. This system can also communicate with PC. In the analog circuits, software -controlled amplifier and anti-alias filter are used. So a good input signal for ADC is provided. The DSP system can communicate with PC through PCI bus.,thus the requirement for real time processing is satisfied. 2.Software development for the instrument In this part, the work mainly includes the control of DSP system, the implementation of FFT, data processing after the event, such as graph drawing, data management, data storage and interfacing to office automation software etc. 3. Application of wavelet transform for axletree crack detection On the base of original technology of frequency analysis, wavelet transform is applied to mechanical malfunction diagnosis. Traditional frequency analysis is based on Fourier transform. As Fourier transform can only get good resolution in frequency domain, it is not fit for non-stationary signal analysis. But wavelet transform can get good resolution in both time and frequency domain, when it is applied to mechanical malfunction diagnosis, we can get good orientation of time. In this paper the application of wavelet transform for axletree crack detection is discussed.