Improving the Performance of A Continuous Stirred Tank Reactor using Moving Horizon State Estimation and Model Predictive Control

Since chemical reactors are utilized to produce specific and valuable products, concentration of products should be regulated at a specified level. As a disturbance input, a change in the inlet concentrations can vary the product concentration. So, in order to regulate the product concentration, the inlet concentrations and the product concentration should be measured. However, measurement of concentration encounters some problems such as high cost and time delay. For compensation of these failures, estimation of concentration has been proposed. In this work, the inlet concentration and the product concentration of a continuous stirred-tank reactor (CSTR) are estimated based on the moving horizon state estimation (MHSE), and the product concentration is regulated based on the model predictive control (MPC). Simulation results indicate that the proposed strategy improves the performance of the CSTR compared with the method in which the inlet concentration is not estimated.