Ddynamic matrix control

Abstract

Dynamic Matrix Control (DMC) was the first Model Predictive Control (MPC) algorithm introduced in early 1980s. These are proven methods that give good performance and are able to operate for long periods without almost any significant intervention. Model predictive control is also the only technique that is able to consider model restrictions. Today, DMC is available in almost all commercial industrial distributed control systems and process simulation software packages.

This project thesis provides a brief overview of Dynamic Matrix Control which is the backbone of Model Predictive Control. A brief history of early industrial MPC applications is given followed by some of its industrial uses. Then some basic structure of model predictive control is discussed. Then follows the three main integral parts of any Model Predictive Control algorithm which are the process model, the cost function and the optimization technique. Various process models like state space model, step response model and impulse response model are discussed followed by cost functions. Quadratic and absolute value cost functions are explained. The receding horizon technique is then explained which simplifies the optimization of the process models. A brief idea about the DMC tuning is given. Finally the simulation outputs under the MATLAB window are provided for the sake of conformity with the theoretical approaches.