Residual life prediction of wind turbine bearings based on adaptive particle filtering and nonlinear Wiener process
SONG Debo
QI Wenzhe
QI Jinping
Abstract：[Objective]Aiming at the problem that wind power bearings are interfered by nonlinear factors and non-Gaussian noise during the monitoring process,and the degradation showing stage characteristics,a model combining adaptive particle filtering with a multi-stage nonlinear Wiener process was constructed.[Methods]Firstly,the Wiener process was introduced into the state-space model of particle filtering to enhance the nonlinear expression ability of the model.Meanwhile,the sampling process was adjusted using an adaptive sampling strategy to effectively avoid particle degradation,thus improving the accuracy of the state estimation and correcting the bearing degradation index.Secondly,a degradation model of bearings was established based on the multi-stage nonlinear Wiener process,and the drift coefficients were randomized by taking into account the variability of different individual bearings.The segmentation points of the degradation stages of bearings were determined by the cumulative sum algorithm,and the model parameters were estimated using the maximum likelihood estimation algorithm.Finally,verification and analysis were conducted on the test bench data and the monitoring data of the free-end bearings of the wind turbine,and the remaining life prediction results obtained by the proposed method were compared with those obtained from the degradation index without filtering correction.[Results]The results show that the proposed method has higher prediction accuracy.
Keywords：Wind turbineRolling bearingAdaptive particle filteringMulti-stage nonlinear Wiener processResidual life prediction
Publication Date：2026-05-31
Online Publishing Date：2026-09-12(First online date of this platform, not the publication date of the document)
Pages：8( 1-8 )