Fuzzy Kalman Filter Based Simultaneous State and Fault Parameter Estimation Scheme with an Application to a Continuous Stirred Tank Reactor

In this paper a multi-model approach based fault detection and identification has been proposed. Takagi-Sugeno dynamic model has been used in this paper to describe the nonlinear dynamic system using locally linearized linear models. The augmented local linear models function as state and fault parameter estimator and the overall state and fault parameter estimation is a non-linear combination of individual local observer outputs. The proposed FDI scheme is tested via simulation on the CSTR process. The performances of fuzzy Kalman filter and extended Kalman filter have been compared