Gearbox fault diagnosis method based on heterogeneous information feature fusion

Abstract

In this paper, a fault diagnosis method based on heterogeneous information feature fusion is proposed to overcome the limitation of the single fault signal of wind turbine and the characteristics of fault characteristics. This method uses the collected vibration signal, speed signal, temperature signal, pressure signal and electrical signal as the original source, respectively extracting the kurtosis, wavelet packet frequency, speed, gearbox inlet temperature, fuel tank temperature, heater temperature, bearing temperature, gearbox pump pressure, inlet pressure, and power as the eigenvalue. Considering the correlation between eigenvalues, the principal component analysis is used to reduce the fusion of the original eigenvalues, and the feature quantity is obtained. The fusion feature is patterned by neural network optimized by genetic algorithm. The simulation results show that the proposed method has higher diagnostic accuracy than the similar information feature fusion method.