A novel hybrid approach for damage identification of wind turbine bearing under variable speed condition

Abstract

Rolling bearings are one of the important components in wind turbine, whose damage may cause a series of concatenate accidents and even sudden shutdown. For the first time, the effects of center coefficient and weighting coefficient on adaptive chirp mode decomposition (ACMD) performances are investigated via fractional Gaussian noise numerical simulation experiments. Moreover, the new initial parameter guided ACMD method is creatively developed based on the energy distribution entropy fluctuation spectrum strategy and the two stage convenient search strategy. Targeting at the damage identification problem of wind turbine bearing under variable speed condition, a hybrid approach is further proposed by fusing the initial parameter guided ACMD method with the computed order tracking (COT) algorithm. Analysis results of experimental signals and engineering case demonstrate this novel hybrid approach can well identify weak bearing damage, the characteristic extraction ability and the damage identification precise of this approach are better than those of different contrastive methods.