Robust fault diagnosis for a class of nonlinear descriptor systems

The paper addresses the problem of designing a robust fault detection filter (RFDF) for a class of nonlinear descriptor systems whose nonlinear term are globally Lipschitz and linear term described by a linear parameter varying (LPV) form. The influence of perturbations on the residuals is minimized using the H_∞ norm while the sensitivity performance is guaranteed using the H_∞ index. Based on the Lyapunov stability theory, the design problem is formulated as a convex optimization problem and solved via linear matrix inequality (LMI) techniques. Performances of the proposed filter are shown through the application to an activated sludge process model.