Study On The Machinery Fault Diagnosis Under Low Frequency Vibration

Machinery fault diagnosis has a very important significance in the modem production. And it is the main method to extract information and parameters, which represent vibration status, from vibration signals. Based on researching a large number documents and papers and analyzing present situation of vibration signal analysis and machinery fault diagnosis technology, the design scheme has been proposed and proved. Design contents are decided as follows: In the hardware design, modeling of systematic design and integration pattern of analogue, digital and graph are presented, and the mode of 憃n-line analysis?and 憃ff- line diagnosis?is ascertained. And in the design of analogue parts, programmable gain amplifying and programmable filtering are applied. Programmable amplifying improves some drawbacks, which are produced by traditional hand switch to adjust gain. Programmable filtering can select cutoff frequency automatically according to inputted frequency parameter and can filter high-frequency interference. Dot matrix graphic liquid crystal display component can show graphs clearly. In the software design, based on the design principles of 慺rom top to end? each aspects are considered comprehensively, then monitoring programs. analyzing algorithm programs, displaying programs and communicating programs are designed and their algorithms flow chart are given. A typical expert system of machinery fault diagnosis is formed. On the basis of frequency analysis, wavelet analysis is used in the machinery fault diagnosis. Traditional frequency spectrum is an analyzing method based on FFT. Fourier analysis has only good resolution in the frequency domain, so it cannot be used to analyzing unstable signals, whereas wavelet transformation has good features in time and frequency domain, and has good localization analyzing capability of signals. When wavelet transformation is applied in the machinery fault diagnosis, it can provide time localization and frequency extraction. Specific circuits are designed and qualitative analysis experiments are finished, and experimental results are given through designing, adjusting the systematic circuits. Experimental results show that the system can achieve programmable selection gain and filtering frequency. From the experimental waveform, we can see that systematic -II- Abstract precision, stability and anti-interference ability are alleviated highly.