Unknown Disturbance Isolation in Electro-Mecanical Systems: Application to PMSG-based wind turbine drivetrain

System diagnostic and disturbance rejection based control is vital nowadays for protecting systems against unwanted perturbations while reducing the fault occurrence rate. Diagnostic and control strategies require knowledge of the disturbance frequency signature in order to detect or cancel out its effect. This paper studies the estimation of stationary and nonstationary system disturbance by the means of an Extended Kalman Filter (EKF) or Unknown Input Observer (UIO) in order to isolate or estimate the disturbance in a system. An application to a wind tubine drivetrain with Permanent Magnet Synchronous Generator (PMSG) is presented in order to compare the results of the classic EKF apporach with the UIO.