Application of model predictive control with Kalman filter for continuous hydro-desulphurisation process

Abstract

Improving sulphur treating of Hydrodesulphurisation process is one of processes, which could reduce the cost of production. Hydro-desulphurisation process is an important process for continuously removing sulphur contents by catalyst from mixed oil, which is part of crude oil that contains high sulphur contents and has high boiling points. The reaction of interest is an exothermic reaction, which is nonlinear with respect to be operating temperature. This results to difficulty in sulphur content in product control at its required specification. Meanwhile, minimum sulphur treating would help in minimum energy use, extending catalyst live and the most important issue is to reduce the risk of unstable temperature from rate of heat removal constraint. In the past, normal linear control technique like PID is not good enough to meet the above requirement. To improve the control performance, model predictive control is an alternative idea of advanced process control, which has been developed to solve the nonlinear process. Formulating model predictive controller, Kalman filter estimator is in corporate with model predictive algorithm for estimating unknown states and parameters. The result from simulation was found that a model predictive controller provided better control response than that of a generic model controller. In addition, estimates of sulphur content in product periodically updated by laboratory results could enhance the performance of Kalman filter in the presence of plant/model mismatches of sulphur in feed and reaction rate.