Research On Fault Feature Extraction Methods Of Wind Turbine Based On Local Mean Decompostion

Fault feature extraction of machines is a hot research direction in fault diagnosis.At present,wind power is still in a relative important stage of development.But the operating conditions of the wind turbine are poor,leading to the high failure rate and operation cost.So to reduce the losses bringing by falut,and reduce the cost of operation,it is very necessary to monitor and diagnose the condition of vibration.But due to the influence of other parts of the unit and the restrictions of its hardware conditions,in addition to influence the rotor speed changes,signal acquired is usually affected by strong noise,strong interference and strong coupling,and usually has non-stationary characteristics.In this context,it may not be very accurate to use classical methods like spectral analysis to extract feature and diagnose fault.Local mean decomposition is a relatively new time-frequency analysis method,which can be applied to the fault feature extraction of the driving chain of the wind turbine.The local mean decomposition method can be used to reduce the impact caused by the electromagnetic in the vibration signal,and can accurately extract the fault characteristics in the driving chain including the bearing and gear.In this paper,the fault feature extraction method based on local mean decomposition is studied.The main contents of this paper are as follows:(1)Explore the basic theory of empirical mode decomposition,draw forth the local mean decomposition on the basis and explore its basic theory.Simulation signals are used to verify the effectiveness of the local mean decomposition method,and comparison is made between the two methods.The mutual information theory is used to remove the false component in the local mean decomposition.(2)Combine the local mean decomposition and Hilbert envelope analysis and apply them to the fault characteristics of the driving chain of the wind turbine,which has a good result.(3)Explore the basic theory and algorithm of Teager-Kaiser Energy Operator,combine it with local mean decomposition and apply them to the fault characteristics of the drving chain of the wind turbine.(4)Aiming at the electromagnetic noise often appearing in the vibration signal,methods based on the empirical mode decomposition and local mean decomposition are used to eliminate the noise,which are verified by an example.