Scenario-integrated Optimization of Semi-batch Reactor Operation under Safety Constraints

Abstract:
Safety is of paramount importance in the operation of exothermic batch and semi-batch reactors. Frequently, failure scenarios such as a malfunction of the existing cooling system must be explicitly considered during the determination of their operational strategy. This thesis presents a methodology for the incorporation of these scenarios into optimization based recipe design. Using a hybrid (discrete-continuous) system description, two types of problem formulations are distinguished. In single-level scenario-integrated optimization problems, the economic performance is the only existing objective in all operational phases. In contrast, bilevel scenario-integrated optimization problems deal with cases where the scenario structure prevents economic optimization after failures and where the enforcement of (safety) constraints remains as dominant objective. For both approaches, tailored numerical solution techniques are developed. The derived methodology can be applied off-line and on-line. This is illustrated using a literature example as well as an industrial polymerization reactor.

Keywords:
Discontinuous reaction processes, batch reactors, safety, dynamic optimization, optimization under uncertainty, hybrid systems, model predictive control.