19592

Application of an optimal MPC tuning strategy in control of a nonlinear reactor system

Favoritar este trabalho

This paper concerns the use of a particle swarm optimization-based MPC tuning method so as to compare the
performances between the conventional MPC and an infinite horizon MPC, when both are applied to a reactor system. More
specifically, the tuning method is carried out on a simulated CSTR system using linearized models along with
process/model mismatch, and so the optimal tuning parameters are also applied to the CSTR nonlinear model. The
simulated results show that the infinite horizon MPC remains stable in all simulated scenarios, whereas the
conventional MPC destabilizes when the nonlinear system is required to be controlled.