The Application Of Mathematical Morphology In The Gear Fault Diagnosis

Gear is a kind of indispensable common component in mechanical equipment, andare widely used in the equipment in the field of modern industry. According to statistics,the gear failure accounted for a large proportion in the mechanical equipment failure, soit is of great significance to gear fault diagnosis.The fault characteristic signals are extracted to the gear is the core and key of thegear fault diagnosis, and the gear fault type and injury can be identified from theextraction of feature signals. On the basis of learning the common types and causes ofthe gear failure, for the nonlinear non-stationary and multi-component modulationcharacteristics of the gear fault vibration signals, the mathematical morphology isintroduced into the gear fault feature extraction in this paper, and combining with thecollection of empirical mode decomposition (EEMD) and the local mean decomposition(LMD) method, the fault feature information of the gear is effectively extracted. In thispaper, the main research content:1. The mathematical morphology method is used to extract the gear fault featuresignals, and the proposed method is demonstrated effectively by the simulationsignal and the simulation of gear fault experimental data. The average filtercan be used to reduce the noise signals, and difference filter can be used forthe extraction of feature signal.2. On the basis of mathematical morphology, EEMD morphology method isproposed. It can be proved that this method is effective through simulationsignals and simulation of gear fault experiment.3. For the problem of decomposing slowly of the EEMD, LMD will beintroduced to the noise reduction of the gear vibration signals in this paper,and combined with mathematical morphology, and LMD morphology methodis proposed to extract the gear fault feature.4. Three methods are compared in this paper, through comprehensive analysisand the results show that the LMD morphology method is superior to other two methods and more suitable for the gear fault feature extraction.Three kinds of fault feature methods in this paper are extracted. And this methodsare compared from the computation efficiency and treatment effect, it can shows thatLMD morphology algorithm is more suitable for application in real-time on-linemonitoring system for large equipment. But this method needs to be applied to more andmore real-time signals, in order to verify its effectiveness and provide a stable andeffective method for equipment real-time online monitoring.