Research On Vibration Diagnosis Method For Wind Power Spindle Bearing Based On Ensemble Local Mean Decomposition

With the gradually highlight of global energy shortages and environmental degradation problems,wind power as a typical representative of clean energy,has more and more attentions from all countries.When the wind turbine is running,the spindle bearings are subjected to both radial load,lateral load and overturning moment.Under the strong impact load of the wind,the spindle bearing is prone to failure.Spindle bearing,as the key part of wind turbines,its performance and reliability determine the overall operating state of wind turbine.Therefore,the research on the fault diagnosis technology of the spindle bearing is a great significance to the safe operation of the wind turbine.In this paper,the spindle bearing is taken as the research object,which fault diagnosis method is researched and analyzed.The main research work is as follows:(1)The characteristics of wind turbine generator structure and spindle bearing are described.Based on the Hertz contact theory,the roller-raceway contact model and the failure mechanism of the spindle bearing are analyzed,and the calculation formula of the typical fault characteristic frequency of the bearing is analyzed,which lays the foundation for the research of the fault diagnosis.(2)Firstly,the principle and algorithm of ensemble local mean decomposition(ELMD),fuzzy entropy and GK clustering are expounded.The endpoint effect problem of ELMD is solved by method of extreme value wave extension.The GK clustering is improved by improving the selection of the initial clustering center.Then,a fault diagnosis method which is based on ELMD fuzzy entropy and improved GK clustering(IGK)is proposed.The fault method is verified with the actual bearing fault data.(3)Aiming at the vibration signal acquisition of wind turbine spindle bearing,the test system is built.Firstly,the design of spindle bearing scale model is carried out based on the similarity theory.Secondly,the structure and working principle of test facility are introduced.Then,it selects the appropriate sensor,determines the location of the measurement points,and analyzes data collection,storage and transmission.Finally,the fault diagnosis system of spindle bearing which is based on LabVIEW is designed.(4)Through the analysis of the normal bearing,the outer ring fault and the rolling body vibration signal,it is proved that the method can well identify and classify the spindle bearing fault.The proposed fault diagnosis method is used to analyze the wind turbine spindle bearings.The bearing vibration signals are firstly decomposed into a series of PFs by the ELMD which represent different frequency bands respectively,and then the PFs including dominant information are selected to calculate their fuzzy entropy values.Finally,these fuzzy entropy values are treated as the fault feature vectors and input into IGK clustering to identify different bearing conditions.