250 / 2021-04-15 00:05:23

Bearing fault diagnosis under variable conditions based on adaptive variational mode decomposition and generalized morphological fractal dimensions for wind turbines

Adaptive variational mode decomposition,Improved frequency band entropy,Generalized morphological fractal dimensions,Variable conditions,Wind turbine bearing,Faults diagnosis

Machine condition monitoring and fault diagnosis

The weak fault features extraction and fault diagnosis of rolling bearings under variable conditions and strong ambient noise is a hot topic. In this paper, a fault diagnosis method under variable conditions based on adaptive variational mode decomposition and generalized morphological fractal dimensions for wind turbine bearings is proposed. Firstly, to improve the decomposition effect of variational mode decomposition (VMD) algorithm, the adaptive learning particle swarm optimization (ALPSO) algorithm is introduced, and the minimum average envelope entropy is used as the fitness function of ALPSO to search the optimal influence parameters in VMD, then the adaptive variational mode decomposition (AVMD) is constructed. Secondly, to solve the problem that the bandwidth parameter of the band-pass filter is set as the empirical value, the envelope kurtosis maximum principle is used to optimize the bandwidth parameter, then the frequency band range (i.e. optimized bandwidth of the band-pass filter) at the minimum frequency band entropy (FBE) is designed as the characteristic frequency band, and improved frequency band entropy (IFBE) is obtained. Thirdly, IFBE is adopted to select the sensitive intrinsic mode function (IMF) from the multiple IMFs decomposed by AVMD, and the envelope power spectrum analysis is carried out to extract the fault characteristic frequency. Finally, generalized morphological fractal dimensions (GMFD) is extracted for different bearing faults diagnosis under variable conditions. Experimental result of wind turbine gearbox bearing vibration signals shows that the AVMD-GMFD method can effectively extract the weak fault features of bearings and realize accurate diagnosis of different fault types under variable conditions.