Abstract: The temperature inside a Continuous Stirred Tank Reactor (CSTR) is difficult to control when chemical reaction takes place. The coolant circulates on the outer jacket of the reactor and extracts the heat energy liberated during the exothermic reaction. The temperature inside the reactor is controlled by manipulating the flow rate of coolant. This study compares the performances of control methodologies like Proportional Integral Derivative (PID) Control, Non Linear Auto Regressive Moving Average (NARMA) model control, Neural Network Predictive (NNP) control and Model Predictive (MP) control. A novel method of control is obtained by incorporating PID method in MP control i.e., the proposed method of control is MP-PID. A significant amount of reduction in time for the control action is obtained for the proposed methods. The time domain specifications on the response for the CSTR Model with the above controllers are tabulated and analyzed.
R. Madhu Sudhanan and P. Poongodi, 2016. Soft Computing Control Methodologies for Temperature Process System and its Performance Evaluation. Asian Journal of Information Technology, 15: 4213-4222.