Gao Lixin  Zhang Jianyu  Cui Lingli

(Key Laboratory of Advanced Manufacturing Technology, Beijing University of  Technology, Beijing 100022)

Ding Qingxin

(Electromechanical Project Institute, China University of Petroleum, Beijing 100022)

Abstract: The sudden fault on the low speed and heavy duty equipments are very difficult to recognize and it can bring about great loss. Though the equipments can be maintained by the vibration monitoring technology, the fault information on the low speed shaft can’t be easily picked through the frequency spectrum analysis. The necessary information on the low frequency range can be obtained by wavelet analysis on the monitoring data. The sultation coefficient of wavelet decomposition on the vibration signal in some time range can be regarded as the characteristic value to judge the fault. Furthermore, the cofficients can capture the developing process of the fault. For the same series monitoring data, the refinable spectrum analysis and FFT analysis compounding wavelet decomposition are both used to complete the comparative research. The achievements show that the refinable spectrum method can’t predict the fault reason and location. On the other hand, after the composite signal processing through wavelet analysis and FFT technique, the vibration amplitudes variation of the characteristic frequency of the hidden fault are very intense. As a result, it is very helpful to precisely recognize the early fault on the low speed and heavy duty equipments.

Key words: Low-speed heavy-duty  Spectrum analysis  Wavelet analysis

CLC No: TH17

北京市科委(H030330050110)和北京工业大学博士科研启动基金(KZ0107200382，00138)资助项目. Received 20050601, received in revised form 20051010