A reduced-order robust observer using nonlinear parameter estimation for induction motors

This paper presents an original approach for the design of a robust observer for a fifth order model of induction motors (IM), defined by a linear system with polytopic representation and linear parameter variation (LPV), to reconstruct the non measurable state variables and mainly the rotor flux. This approach is based on the notion of robust detectability and the using of the linear matrix inequality (LMI) tools to calculate the gain of the polytopic reduced-order robust observer, and the identification online of the variable parameter (the rotor time constant) using a special nonlinear observer. The dynamical performances of the proposed observer is then analyzed and compared.