Fault diagnosis of rolling bearing of wind turbine generator based on PSO-SEBD

As a key component of wind turbine drive system, health status monitoring of rolling hearing is very important to the safe and stahle operation of the whole unit. Here, aiming at the problem of rolling hearing fault diagnosis, on the basis of the method of extracting cyclo-stationarity of repetitive transients from envelope spectrum based on prior-unknown blind deeonvolution technique (SEBD), a SEBD rolling bearing fault diagnosis method based on particle swarm optimization (PSO) algorithm (PSO-SEBD) was proposed to realize adaptive selection of SEBD filter length. Firstly, taking the maximum characteristic frequency ratio ( CFH) as the fitness function, the PSO algorithm was used to optimize the filter length. Then, the optimal filter length was used for SEBD processing. Finally, the effective identification of bearing faults was realized according to envelope spectrum characteristics of signals processed with SEBD. The effectiveness of PSO-SEBD was verified by using it to analyze simulation signals and the published bearing fault data of University of Paderborn in Germany. By comparing PSO-SEBD with several common diagnosis methods and analyses in noise environment, it was shown that PSO-SEBD has better diagnosis performance and anti-noise ability. © 2023 Chinese Vibration Engineering Society. All rights reserved.