Prediction based optimal control of a quadruple tank process

Industrial processes are nonlinear, multivariate, constrained and interactive in nature. To ensure stable operation of the process under these circumstances, a tight control action should be employed. Conventional controllers like PID provides simplicity in design and ease of access but fails to provide good control action for the process which are nonlinear and operating under heavy constraints. In this paper Model predictive controller has been designed for a bench mark quadruple tank process and its performance is compared with a PI controller. It is found that Model predictive controller provides satisfactory results when compared to PI controllers.