MIMO filtered positional generalized predictive controller design for handling offset

The purpose of this paper is to present an alternative synthesis for the Generalized Predictive Controller (GPC) applied to a multiple-input multiple-output (MIMO) model, using both a positional plant model and cost function to deal with reference tracking, disturbance rejection and model-plant mismatch (MPM) with minimum control energy and closed-loop robustness. This new control design called Multivariable Filtered Positional GPC (MFP-GPC) uses an integral polynomial weighting filter for the setpoint and output of the plant, thereby extending the applicability of the predictive controllers. Additionally, robustness aspects are incorporated into the control design of the weighting polynomials, and this implementation involves the filter tuning parameters using a multi-objective optimization algorithm. Numerical simulations for a hydroelectric power and a methanol/water distillation plants show the effectiveness of the proposed control methodology.