Simulation of Model Predictive Control using Dynamic Matrix Control algorithm

Model Predictive Control has emerged as a very powerful technology in the area of process control for three decades. The objective of this work was to develop Dynamic Matrix Control Algorithm, one of the most widely used Model Predictive Control Algorithms using MATLAB and simulate it for a real world Single Input Single Output system. This thesis focuses on the impacts and importance of the tuning parameters of Dynamic Matrix Control along with an overview of the general Model Predictive Control strategy. The tuning of the Dynamic Matrix Controller was done by trial and error based on the knowledge of the simulated system under consideration and the control strategy. The Control Signal computed was then implemented on the system to study its effect on the system output using a discrete transfer function model. The results of the tuned controller were observed to be similar to the other tuning methods discussed in the literature.