Dynamic Matrix Control (DMC) Parameters Tuning Method and Performance Evaluation Technique

Abstract

Model predictive control (MPC) was developed to meet control challenges of advanced process. Dynamic Matrix Control (DMC) was the first MPC algorithm developed by Shell Oil Company in 1970s. DMC is available in all industrial process control systems and on a number of control simulation platforms. Conventional PID based controller is implemented to control concentration of Continuous Stirred Tank Reactor (CSTR). Conventional PID controller can be electively used to regulate the output concentration around a small operating range that PID controller is designed for. But CSTR is a highly nonlinear system. Control performance degrades if operating point changes significantly. Linear Quadratic Regulator (LQR) controller is implemented to control concentration of Continuous Stirred Tank Reactor (CSTR). LQR performance is satisfactory in lower operating range. While for the higher concentration (Normally above 0.12 (mol/lit) ) not performing satisfactory. This work deals with DMC for an unconstrained and constrained Single Input Single Output (SISO) system. DMC has been formulated and simulated for the CSTR concentration control. DMC based CSTR is compared with other control strategies like PID and LQR. Different parameters like model length, prediction horizon, controller horizon and sampling time has been tune manually and by using different method and analyzed for the constraint and unconstraint problem. The further scope of work is to apply tuning method for delayed MIMO system and control performance can be evaluated and implemented.