MODEL PREDICTIVE CONTROL OF INDUSTRIAL SYSTEMS
Project no. 720 · Status: Kapanmış · 2016
Summary
Model Predictive Control (MPC) is a commonly used control technique in advanced process control. Numerous successful studies on linear system model-based MPC method have been reported. Nonlinear system model based studies are still going on. In the project, linear MPC designs have been performed by using state-space models of the systems. The performance of the designed MPC algorithm was examined with real-time application on a real industrial system as well as the simulation studies. The master's thesis of the researcher was carried out in parallel with the project. The performance of the unconstrained linear MPC system is tested on MATLAB by either simulating with the mathematical model of the system or applying the real system. In the constrained cases of MPC algorithm, the simulation studies are performed by defining different cost functions on the system outputs, control input signals and control input step signals. The simulation and application studies in the project shown that the MPC method can be effectively used for industrial control problems.