Particle Filter Based Approach for Wind Turbine High-Speed Shaft Bearing Health Prognosis

Mechanical failures in wind turbine generators cause many problems of stopping electricity production, increase the price of maintenance or the destruction of the wind turbine. Gearbox is the important component in wind turbine it composed by bearing and gearings, generally failure appear in the bearing then in gearings.A probabilistic framework based on Bayesian methods applied to vibration bearing signals is a suitable tool to health state estimation. Vibration signals are nonlinear and non-Gaussian process, then the use of particle filter as derived from Bayesian approaches is a suitable method for this kind of signals.The proposed method is applied on a real high speed shaft bearing vibration signals, wish give a good results.