Research On Vibration Fault Diagnosis Method Of Wind Turbine Gearbox

With the development of the economy,human dependence on energy is increasing day by day,and the massive use of wind energy can effectively alleviate the energy crisis and environmental pollution.As the capacity of the wind turbine assembly machine increases year by year,the wind turbine generator fails frequently,and the gear box is an important mechanical component of the fan transmission system,and belongs to a component with high failure frequency.Once the gearbox fails,the downtime is long and the maintenance cost is high,which has a huge impact on the safe and reliable operation of the wind turbine.Therefore,the diagnosis of gearbox faults is carried out,and the fault diagnosis method of gearbox is mainly studied.The gears and bearings of the main parts of the gearbox are taken as research objects,mainly the filtering noise reduction method based on fault vibration signals,fault feature extraction and intelligent fault diagnosis methods to conduct specific research.This paper mainly does the following three aspects:(1)For the traditional wavelet denoising and wavelet packet denoising methods,there are some defects such as artificially setting the threshold,and the method of combination of morphological filter and noise reduction is adopted.The MATLAB simulation analysis and experimental results show that the method introduced in this paper has higher noise reduction ratio and better noise reduction effect.The envelope demodulation analysis of the noise-reducing gear and bearing vibration signals is carried out,and the fault type of the gearbox can be preliminarily determined by extracting the fault characteristic frequency.(2)For the defect that the particle swarm optimization(PSO)often falls into local optimal solution,a improved model based on black random hole model-modified simulated annealing particle swarm optimization(RBH-ISAPSO)is proposed.After comparing six standardtest functions and PSO and IPSO algorithms,the superiority of RBH-ISAPSO algorithm is verified and applied to the parameter optimization of LSSVM.The fault classification model of RBH-ISAPSO-LSSVM is established.The UCI standard dataset is compared with the LSSVM and CV-LSSVM classification models to verify the effectiveness of the model.Finally,it is applied to the gearbox fault diagnosis and achieved ideal results.(3)In view of the non-linear and non-stationary characteristics of the vibration signal of gear box in practice and the problem of mode mixing in the empirical mode decomposition(EMD)method,the new method of empirical mode decomposition-least squares support vector machine(EEMD-LSSVM)is proposed.Firstly,PCA is used to reduce the noise of the gearbox vibration signal after denoising,and then the EEMD decomposition is used to extract the energy entropy of the first 8 IMF components as a set of feature vectors,which are input to the LSSVM classifier for training and testing.The experimental results show that the method can effectively identify the fault type of the gearbox.