Search

The dynamics of the most of the industrial plants (for example electric power system, electromechanical system, electro-hydraulic system and so on) are highly nonlinear and, moreover, include actuator dynamics which increase the relative degree of the complete system. To stabilize the plant dynamics it is naturally to applied some feedback linearization (FL) technique: block control [18], backstepping [14] or input-output linearization [11], since the model of these plants can be presented in the nonlinear block controllable form or (the same) strict-feedback one. All these control techniques require to calculate the time derivatives of the plant dynamics vector fields (Lie derivatives), results in a computationally expensive control algorithm, and moreover, the closed-loop system is susceptible to plant parameter variations and disturbances. To simplify the control algorithm the actuator fast dynamics are usually skipped, and to overcome the robust problem the sliding mode (SM) control [25] in combination with FL technique [18], [17] can be can be applied. However, the presence of the actuator unmodeled fast dynamics can destroy the desired behavior of the SM control systems causing lost of robustness and accuracy and provoking the chattering effect [25], [8]. Therefore, the problem of control design for the systems with unmodeled actuator dynamics becomes to be a big challenge.