Research On Method For Fault Diagnosis Of Rolling Bearing In Heavy Noise

Coal is the main energy in china, and coal mine safety can not be overlooked.Electrical haulage shearer is the key equipment of modern mine, which is related tocoal mine safty and efficient production. As the domestic shearer often breakes down,it is important to study the shearer monitoring and fault diagnosis technologies.Rolling bearing is the key component of electrical haulage shearer, as well as the oneof the important sources of failure of coal mining machine. Whether or not can weaccurately determine bearing work status directly relate to the coal mining machinesecurity. Research of monitoring and diagnosis technology of rolling bearing is a topicthat has considerable theoretical and practical interest.for avoiding the occurrence ofmajor accident. In this paper, on the previous work, aiming at the respectiveshortcomings and limitations of EMD and stochastic resonance, mechanical faultfeature extraction method based on improved EMD and stochastic resonance isproposed. The main research work and conclusions are as follows:(1) Study on vibration characteristics of the rolling element bearing andtime-frequency characteristics of bearing when there is a defect in the inner ring, outerring or rolling element is gained. Misalignment fault diagnosis examples show thatthe fault characteristic frequency obtained is correct and it is feasible to extract faultfeatrues of rolling element bearings through the frequency spectrum.(2) On the basis of EMD basic theory, algorithm and its characteristics, study onend effect and mode mixture, and put forward the effective solutions to theseproblems: the support vector machine and window function are combined to eliminateEMD end effect; the problem of mode mixing is solved by using the EEMD method;Signal-to-noise ratio is used to chose the intrinsic mode functions which contain faultinformation. Simulation results and misalignment fault diagnosis examples show thatthe improved method can inhibit end effect effectively and eliminate mode mixing inrolling element bearings fault diagnosis.(3) The feature extraction of the bearing’s weak fault is not only very importantbut also very hard. Traditional methods mostly extract the weak fault feature throughnoise reduction, but the noise reduction weakens the useful feature unavoidably,which may cause misdiagnosis. An adaptive stochastic resonance (ASR) method isproposed to address the issue. The ASR method utilizes the optimization ability ofgenetic algorithm and adaptively realizes the optimal stochastic resonance system matching input signals. Simultaneously, the method combined with frequency-shiftedand re-scaling stochastic resonance (FRSR), enables to achieve the stochasticresonance under the conditions of great parameters. Using the ASR method, the noisemay be weakened and weak characteristics highlighted, and therefore the faults can bediagnosed accurately. Simulation results and misalignment fault diagnosis examplesshow that weak fault feature buried in strong noise are well extracted in case of smallnumber of sample points and small sample frequency.(4) The single EMD or stochastic resonance, however, fails to extract the faultfeatrues when the Signal-to-noise ratio of the bearing vibration signals is very low. Toaddress this problem, an Improved Ensemble Empirical Mode Decomposition(IEEMD) and adaptive ASR method is proposed in this paper. Experiments are madeto simulation of shearer rocker arm bearing working environment. The proposedmethod is used to process the collected signals and the results of feature extractionand fault diagnosis demonstrate its effectiveness.At the end of this thesis, the summarizations of the research and expectation ofthe related technology development are presented.