Model Predictive Control for Nonlinear System

Abstract

This project comprised of the usage of Model Predictive Control (MPC) for different tank level control systems. MPC is advance control algorithm mostly used in process industries where controlling of a different parameters are very important from process as well as safety point of view. MPC has extended in popularity because it provides a methodical approach to control the multivariable systems with constraints. Conventional PID and LQR (Linear Quadratic Regulator) are generally used for the SISO process and where there are no constraint in the process. So MPC is the solution which overcome the demerits of the PID controller and LQR algorithm. Here there are three different scheme of tank level systems where MPC has employed, like in single tank level control system as single input single output (SISO), coupled tank level control system as multiple input single output (MISO) and quadruple tank level control system as multiple input multiple output (MIMO) system. The work started with mathematical modelling of these systems and then after Model Predictive Control law for unconstrained and constrained cases with appropriate prediction horizon and control horizon been applied and analysed the predicted output response and its input control action. Comparative analysis of systems with manually tuned design parameters are also analysed in this project with the help of MATLAB and SIMULINK toolbox.