Offset-free state-space nonlinear predictive control for Wiener systems

Abstract This work is concerned with state space Multiple-Input Multiple-Output (MIMO) Wiener systems which consist of a linear dynamic block connected in series with a nonlinear steady-state (static) one. Model Predictive Control (MPC) algorithms with successive on-line model or trajectory linearisation for dynamic processes described by such Wiener systems are discussed. Advantages of the presented MPC algorithms are: (a) computational efficiency since quadratic optimisation problems are only solved on-line, nonlinear optimisation is not necessary, (b) very good quality of control, (c) offset-free control (no steady-state error in presence of disturbances) assured by a novel approach to disturbance modelling and state estimation, resulting in a simple design and a simple control structure. All features of the discussed algorithms are demonstrated and their performance is compared with that of the MPC algorithm with nonlinear optimisation as well as with the traditional offset-free state-space MPC approach.