A nonlinear disturbance observer for multivariable systems and its application to magnetic bearing systems

This paper proposes a new nonlinear-disturbance observer for multivariable minimum-phase systems with arbitrary relative degrees. The model uncertainties and the system nonlinearities are treated as disturbances. The estimation of individual disturbances is independent of each other and the derivatives of the disturbances can be independently estimated. The proposed formulation is inspired by the variable structure-control method and adaptive algorithms where the a priori information concerning the upper bounds of the disturbances and their derivatives is not required. The nonlinear-disturbance observer is robust to the types of disturbances. Stability analysis shows that the estimation error decreases exponentially to a steady value, which is determined by the design parameters. To illustrate the method, the proposed design is applied to a vertical-shaft magnetic-bearing system where the rotational disturbances and their derivatives are estimated based on a linearized model of the rotational motion. Simulation results show the effectiveness of the proposed method.