A scenario-based framework for the integration of scheduling and control under multiple uncertainties

Increasingly dynamic market conditions require chemical processes to be operated in a more agile manner. Improved process response times can be achieved via tighter cooperation between the decision making-layers. In this study, we present a scenario-based stochastic formulation for the integration of scheduling and control that accommodates multiple uncertainty types, including model, demand and cost uncertainty. We account for the associated PI control system, including input saturation, in the problem formulation, enabling it to predict the closed-loop process dynamics. The resulting closed-loop problem is solved at the dynamic real-time optimization (DRTO) level, in a moving horizon fashion, to compute optimal scheduling decisions, and set-points that are provided to the tracking PI controllers. We test the efficacy of the proposed formulation to handle uncertainties in closed-loop process operation via case studies. We also investigate the effect of the DRTO execution frequency on the process performance. The performance benefits of the proposed formulation over a nominal DRTO implementation are demonstrated.