Veranstaltung

Abstract: deregulation and dynamic market conditions have created significant opportunities for the engagement of industrial entities normally regarded as electricity consumers in the operation of the power grid. Chemical plants (e.g., air separation, ammonia production, chlor-alkali) are particularly promising candidates given their large power demand. Engaging in such activities (either demand response or providing ancillary services) calls for a close coordination between business (planning, scheduling) and process control decisions of a chemical plant. The integration of  business decisions with dynamic information from the control layer is a difficult task owing to the broad range of time scales involved in making the respective decisions, and the corresponding need to balance long-term prediction with real-time execution. In this presentation, I will focus on recent developments that allow for a closer coordination between production scheduling and supervisory control systems. I will introduce a new modeling framework, based on capturing the input-output behavior of the chemical process and its control system (whether multi-loop or MPC) in a low-dimensional model, which is then used in scheduling calculations. I will discuss ways to exploit historical process operating data (which are available in most plants, practically “for free”) in building such models. I will also show that the integrated scheduling-control problem can be cast as a mixed integer linear program that can be solved efficiently for problems of practical interest. An industry-based case study concerning the demand-response operation of an air separation plant will be discussed, along with other industrial applications.

The seminar can also be followed remotely by joining the online Zoom meeting (connection possible 15 minutes before the talk).
In case of problem, you can contact our IT support (37679 - it.vs@epfl.ch )

Link zur Website: https://memento.epfl.ch/event/demand-response-operation-of-chemical-plants-mod-2/

Weitere Informationen