Fault Diagnosis Of Wind Turbine Gearbox Based On Signal Resonance Sparse Decomposition

Operating on terrible conditions,the failure rate of wind turbine is high with the variable load,resulting in low utilization as well as utilization hours of the wind turbine and high operation and maintenance costs.In order to ensure the safety of the operation and improve the economics and market competitiveness of the wind power,it is necessary to carry out condition monitoring and fault diagnosis for the wind turbine.The wind turbine condition monitoring system(CMS)realizes fault diagnosis by monitoring the vibration of the transmission chain components.Most methods of vibration signal analysis are based on orthogonal linear transformation such as Fourier transform and wavelet transform,making the fault information decomposed into too many basis functions,which make against to the extraction of fault features.If the basis function is redundant,the signal can be represented with a series of linear combinations of basis functions with the best matching basis function selected adaptively according to the characteristics of the signal.Due to the high redundancy of the dictionary,sparse decomposition can represent complex vibration signals,and the signal can be represented with fewer basis functions,avoiding the fault features scattered into too many basis functions.In order to find the fault characteristics of equipment in early stage,this paper introduces the sparse decomposition method into the vibration signal processing of wind turbines.According to the different morphology between the fault component and the stationary meshing component in the vibration signal of the gearbox vibration,the sparse decomposition method based on the tunable Q-factor wavelet transform(TQWT)is used to extract the fault component in the signal.and the effect of the proposed method was proved through the simulation signal and the actual signal.The main research contents of this paper are as follows:(1)According to the structure and transmission characteristics of the planetary gearbox in the wind turbine,the characteristics of the vibration signals of various faults of the gearbox are studied,the corresponding fault frequency is calculated,and the spectrum characteristics are studied by simulation signals.(2)The basic theory of signal sparse decomposition and several basic analysis methods of sparse decomposition are introduced.The characteristics and limitations of traditional sparse dictionary are analyzed.The mechanism and related characteristics of adjustable quality factor wavelet are also studied.(3)According to the theory of resonance sparse decomposition,the sparse decomposition algorithm based on Li norm decompose the signal with low decomposition precision,and the outcome is unstable.to solve this problem,an improved resonance sparse decomposition method based on non-convex sparse metric function and noise optimization is proposed.Combining this method with multi-scale envelope analysis,a multi-fault diagnosis method for wind turbines is formed with the program achieved.The method is validated by the simulation signal and the actual wind turbine vibration signal,which proved the effectiveness of the method.