Adaptive Internal Model Control and its Application to a Batch Polymerization Reactor

Abstract

In this paper, the internal model control (IMC) and the adaptive parameter identification technique by the model refernece adaptive system (MRAS) are newly combined, and the combined one is called the adaptive IMC (AIMC). The proposed AIMC inherits the advantageous points from IMC, e.g., easiness to evaluate the robustness of the modelling error and to design a filter of an AIMC taking into account of its robustness. Moreover, parameters of the internal model are adjusted by MRAS in the proposed scheme. Then, the system can be controlled perfectly if the order of the system is appropriately assumed and if the system is a minimum phase one. One more advantageous point which is most significant one in practical application of the AIMC is; if the model is taken as 1st order system, the AIMC can be reduced to an auto-tuning proportional-plus-integral (PI) feedback controller. The proposed AIMC is applied to the control of a batch polymerization reactor. Temperature of the reactor is controlled so as to follow the desired temperature profile by the AIMC. The system equation of the batch reactor is time-dependent, and parameter adaptation will be required for getting a sophisticated controller. For this problem, the proposed AIMC can be utilized successfully and is shown to be superior compared with a traditional PI controller by the computer simulation.