An infinite model predictive controller for multi input nonlinear processes

This paper presents a novel algorithm of an infinite model predictive controller for controlling nonlinear multi-input multioutput (MIMO) processes. The new strategy uses a set of continuous nonlinear functions that captures the nonlinear characteristics of the MIMO plant over a wide operating range resulting in a more accurate prediction of the controlled variables. The method formulates a nonlinear dynamic matrix that is manipulated variable dependent during closed-loop control. The proposed algorithm was implemented on a nonlinear MIMO thermal system comprising of three temperature zones to be controlled with interacting effects. The experimental closed-loop responses of the proposed algorithm were compared to a multi-model dynamic matrix controller (DMC) with improved results for various setpoint trajectories. The MIMO process has nonlinear parameters such as process gain and time constant that are dependent on the size of the control actions. Good disturbance rejection was attained resulting in improved tracking of multi-setpoint profiles in comparison to multi-model DMC.