Numerical Verification of Robust Model Predictive Control to Electrically Heated Stirred Tank System
Ori Amao FOIMAE (1125125)
The aim of this thesis is to numerically verify the application of a min-max robust Model Predictive Control approach to an electrically heated stirred tank process. The robust min-max model predictive control method used in this thesis has been previ- ously shown to guarantee the stability of predicted states at the terminal region in the presence of disturbance inputs. In this thesis we check the performance of the electri- cally heated stirred tank system with a min-max robust MPC against several types of process noise. In particular we confirm the performance with respect to the robustness to model uncertainty introduced by state feedback and bounded disturbance input. We evaluate this performance criteria using four types of process noise namely sinusoidal, sawtooth, random normalized standard noise and constant disturbance input. The re- sults show that with the use of the min-max robust MPC, the electrically heated stirred tank process model considered was robust to bounded disturbance inputs and model uncertainty.