Research On Noise Reduction Of Wind Turbine Gearbox Fault Signal Based On Variable Step Size LMS Algorithm

With the development of science and technology and the improvement of economic level,people’s demand for clean energy is increasing.Compared with other clean energy,wind energy is favored for its advantages of easy exploitation and sustainable utilization,and has become one of the most promising clean energy in the 21 st century.The current main way to use wind energy is to convert wind energy into electrical energy through wind turbines.Because wind turbines are usually in a relatively tough environment,it is very easy to cause failures of its various components,and the gearbox is the component with the highest failure probability.Generally,the collected gearbox fault signals also contain many other noise signals,which will affect the diagnosis of the gearbox fault signals.Therefore,it is necessary to denoise the collected gearbox fault signals.Based on the Least Mean Square(LMS)algorithm,this thesis has done a series of researches on the noise reduction of gearbox fault signals.The main work is as follows:(1)Aiming at the problem of large steady-state errors of the adaptive filtering LMS algorithm,an improved variable step size LMS algorithm based on normalization is proposed.The improved LMS algorithm uses the 2-norm of the input signal and the related signal error to calculate the step size factor,which reduces the steady-state errors of the system and improves the performance of the variable step size LMS algorithm.The simulation experiment results show that the improved variable step size LMS algorithm has lower steady-state errors compared with the traditional LMS algorithm.(2)Aiming at the problem that the parameters of the variable step size LMS algorithm cannot be automatically optimized,a normalized variable step size LMS algorithm based on Beetle Antennae Search(BAS)algorithm optimization is proposed.This paper optimizes the parameters in the improved variable step size LMS algorithm by using BAS algorithm,and then uses the optimized algorithm to denoise the gearbox fault signal.Simulation and actual experiment results show that the proposed algorithm realizes the automatic optimization of the parameters of the variable step size LMS algorithm,reduces the steady-state errors of the system by about 20%,and improves the stability and noise reduction performance of the algorithm.(3)In order to further improve the noise reduction performance of the variable step size LMS algorithm,an optimized variable step size LMS algorithm and the Variational Mode Decomposition(VMD)algorithm for joint noise reduction method is proposed.Firstly,the fault signal is decomposed by VMD;secondly,three Intrinsic Mode Functions(IMF)with the largest kurtosis value are selected for signal reconstruction;Thirdly,the reconstructed signal is denoised by the improved variable step size LMS algorithm.Simulation and actual experimental results show that,compared with the optimized variable step size LMS algorithm and VMD,the proposed method has better noise reduction performance on gearbox fault signals,which in turn helps to improve the accuracy of gearbox fault signal diagnosis.