Steady State Optimization and Model Predictive Control of a Tubular Reactor

In this paper, the steady state and dynamic optimization of a continuous non-isothermal tubular reactor is considered, where a first-order irreversible chemical reaction takes place. The reaction is exothermic and a coolant flow is continuously circulating through the reactor jacket. The control of reactor is based on its analytical model in the spatiotemporal space. The steady state optimization of reactor aims at maximizing the product concentration at the reactor output for given inlet reagent concentration and temperature. Then, the optimal stabilization of the outlet reagent concentration and temperature at their optimal set-point values is achieved by the design of a model predictive controller. The performance of the predictive controller is studied in the presence of parametric model uncertainty (about the heat transfer coefficient) and in the presence of disturbances in the coolant temperature. The results show that the optimal control system guarantees an offset-free stabilization of the output variables despite of model uncertainty and leads to high quality transients in the presence of disturbances.