Abstract
Traditional model predictive control with a process model represented by an impulse or a step response cannot be applied to integral processes, which are typical in chemical plants. In this paper a parametric input-output model was used to develop a model predictive control algorithm for the integral process. Two control algorithms with different prediction models were studied with respect to closed-loop properties. They make a difference in the robust stability of the control system in the case of under-estimated process gain.
