Robust actuator and sensor fault estimation for Takagi-Sugeno fuzzy systems under ellipsoidal bounding
Marcin Witczak, Marcin Pazera, Norbert Kukurowski, Teódulo Iván Bravo Cruz, Didier Theilliol · 2020
The paper undertakes the problem of designing a simultaneous sensor and actuator fault estimation scheme for Takagi-Sugeno fuzzy systems. Thus, the paper starts with the development of such a fault estimation scheme capable of estimating these faults simultaneously. Apart from estimating the faults, the proposed scheme provides the so-called uncertainty intervals, which overbound an unknown state and faults. These intervals can be applied for both reliable fault diagnosis as well as assessment of the estimation quality. Indeed, the smaller the intervals the better the estimation. To settle the above problem, an assumption is imposed, which yields that the external disturbances are overbounded by an ellipsoid. This permits employing the quadratic boundedness approach both both the estimator convergence analysis as well as determination of lower and upper bounds of uncertainty intervals. Finally, the performance of the proposed approach is examined by exploiting the laboratory three-tank system. In particular, the effectiveness of the proposed scheme is tested against a set of simultaneous actuator and sensor faults, respectively.