Fault Detection with Unscented Kalman Filter Applied to Nonlinear Differential-Algebraic Systems

This paper contemplates the unscented Kalman filter (UKF) algorithm for nonlinear semi-explicit index 1 differential-algebraic equation (DAE) systems. Incipient, the formulation of the UKF algorithm proposed by Mandela et al. in [1] concerning this system class is outlined and an revised formulation for the unscented Kalman filter for nonlinear semi-explicit index 1 differential-algebraic equation systems is described. Further, a simple and robust residual definition based on the state estimation with the introduced UKF is given aiming at the fault detection. Finally, simulation results illustrate the efficiency of the proposed fault detection approach applied to a hydraulic system.