Doubly fed induction generator fault diagnosis using unknown input Takagi-Sugeno observer

This paper presents a new approach to detect and isolate the current sensor faults, a doubly fed induction generator (DFIG) for a wind turbine application. A method using an unknown input of multiple observers described via Takagi-Sugeno (T-S) multiple models. A bank of multiple observers generates a set of residuals for detection and isolation of sensor faults which can affect a TS model. The stability and the performance of the multiple models are formulated in terms of Linear Matrix Inequalities (LMIs). The LMIs can be efficiently solved using convex optimization techniques, where the convergence conditions of the state estimation errors are expressed in LMI formulation using the Lyapunov method.