Performance Analysis of MPC for Level Control of Modified Quadruple Tank System

摘要

This work carries out a comprehensive performance analysis of Model Predictive Controller (MPC) for a modified version of benchmark Quadruple Tank System (QTS) which is a Multi input multi output (MIMO), highly nonlinear coupled system. The objective is to control the water level of Modified Quadruple Tank System (mQTS) with the optimization of objective cost function of system error while satisfying input amplitude constraints. Simulation studies have been carried out and results are fairly compared with PI and LQR controllers. Results proclaim that the centralized MPC with a sufficiently large prediction horizon is better choice as compared to decentralized PI and centralized LQR controller as it measures the loss of performance due to the linear nature of the prediction model.