Abstract:

A new fault detection scheme for continuous-time nonlinear dynamic system is studied. The nonlinear system is modelled by Takagi-Sugeno fuzzy model. A fuzzy observer-based approach is presented to detect the fault occurred in the dynamic system. The generated residual signal is robust with respect to undesirable effects of unknown inputs and modelling errors but sensitive to faults. By applying H∞ optimization techniques, a sufficient condition to solvability of the formulated problem is established in terms of Linear Matrix Inequalities. The algorithm for residual evaluation function and threshold computation are also given. Finally, an example is given to illustrate the effectiveness of the proposed design techniques.