Abstract
We address the distributed hierarchical optimization of industrial production complexes where the individual plants exchange resources via networks. Due to the site-wide couplings a centralized or a distributed hierarchical optimization is needed to achieve the best overall performance of the site and to balance the networks of the shared resources. We discuss market-like algorithms that set prices of the shared resources in order to influence the individual optimizers so that the overall operation converges to the site-wide optimum. A novel algorithm for price adjustment based on the quadratic approximation of the responses of the individual optimizers is presented. It shows convergence to the site-wide optimum with significantly less iterations in comparison to the standard subgradient-based method for a set of case studies, including a petrochemical complex.
Zusammenfassung
Dieser Beitrag behandelt die verteilt-hierarchische Optimierung industrieller Produktionskomplexe, in denen die einzelnen Anlagen durch über Netzwerke ausgetauschte Ressourcen eng miteinander verkoppelt sind. Durch die Kopplungen entspricht das standortweite Optimum nicht den individuellen Anlagenoptima. Es wird daher eine zentralisierte oder verteilt-hierarchische Optimierung benötigt, um das optimale Gesamtergebnis zu erreichen. Wir diskutieren marktähnliche Algorithmen, die Preise für Ressourcen festlegen und damit die individuellen Optimierer steuern, um das Standortoptimum zu finden. Ein neuer Algorithmus für die Preisanpassung wird vorgestellt, beruhend auf der quadratischen Approximation der Antworten der Teilsysteme auf die Preissignale. Der Algorithmus benötigt in verschiedenen Fallstudien signifikant weniger Iterationen im Vergleich zur Subgradientenmethode.
About the authors
Simon Wenzel, M.Sc. (born 1988) received the M.Sc. degree in Chemical Engineering – Specialisation Process Systems Engineering, at TU Dortmund University in 2014. Since then he has been a research associate with the Process Dynamics and Operations Group at TU Dortmund University, working on his PhD in the field of distributed coordination and market-like algorithms.
Technische Universität Dortmund, Fakultät Bio- und Chemieingenieurwesen, Lehrstuhl für Systemdynamik und Prozessführung, Emil-Figge-Straße 70, 44227 Dortmund, Germany
Dr. Radoslav Paulen (born 1984) graduated at Slovak University of Technology in Bratislava (Slovakia) in 2008. He received his PhD from the same university in the field of Process Control in 2012. Since 2012, he has been with the Process Dynamics and Operations Group at TU Dortmund University. His research interests include distributed optimization and control, model-based optimizing control, and guaranteed parameter and state estimation.
Technische Universität Dortmund, Fakultät Bio- und Chemieingenieurwesen, Lehrstuhl für Systemdynamik und Prozessführung, Emil-Figge-Straße 70, 44227 Dortmund, Germany
Goran Stojanovski, PhD (born 1984) received the Dipl.-Ing. and his PhD degree from the Ss. Cyril and Methodius University in Skopje, Macedonia in 2007 and 2013 respectively. In 2014, he joined the Process Dynamics and Operations Group at TU Dortmund University, working in the field of distributed coordination for plant-wide optimization. In April 2015, he was appointed to his present position as an Assistant Professor of Automation and Control Systems with the Faculty of Electrical Engineering and Information Technologies at Ss. Cyril and Methodius University in Skopje.
Ss. Cyril and Methodius University in Skopje, Faculty of Electrical Engineering and Information Technologies, Rugjer Boshkovik 18, PO Box 574, 1000 Skopje, Republic of Macedonia
Dr.-Ing. Stefan Krämer (born 1972) studied Chemical Engineering at the University of Newcastle upon Tyne (GB) and at TU Dortmund University. He joined INEOS in 2004 as an Advanced Control Engineer and ran a team of APC and DCS engineers from 2009 to 2013. As Site Energy Manager his responsibilities now include operating the site wide energy management system and coordinating site energy optimization and energy efficiency projects. He teaches Batch Process Operation as a Guest Lecturer at TU Dortmund University. He is active in the NAMUR working group on Energy Efficiency.
INEOS Köln GmbH, Alte Str. 201, 50769 Köln, Germany
Benedikt Beisheim, M.Sc. (born 1987) studied Chemical Engineering at the RWTH Aachen University and received the M.Sc. degree in 2014. Since 2014 he has been with INEOS in Köln, working in the energy management department. He is responsible for energy optimization projects and is involved in the EU-funded projects DYMASOS and MORE.
INEOS Köln GmbH, Alte Str. 201, 50769 Köln, Germany
Prof. Dr.-Ing. Sebastian Engell (born 1954) received the Dipl.-Ing. degree in Electrical Engineering from the Ruhr-Universität Bochum, Germany, in 1978, and the Dr.-Ing. degree from the Universität Duisburg, Germany, in 1981. In 1990, he was appointed to his present position as a Full Professor of Process Dynamics and Operations with the Department of Biochemical and Chemical Engineering at TU Dortmund University. His areas of research are robust optimizing control, in particular of chemical and biochemical processes, dynamic modeling, coordination in systems of systems, and planning and scheduling under uncertainty. He is recipient of an ERC Advanced Investigator Grant and the current President of the European Control Association.
Technische Universität Dortmund, Fakultät Bio- und Chemieingenieurwesen, Lehrstuhl für Systemdynamik und Prozessführung, Emil-Figge-Straße 70, 44227 Dortmund, Germany
Acknowledgement
The authors gratefully acknowledge the support of the European Commission under the grant agreement number 611281 (FP7 ICT project DYMASOS, www.dymasos.eu).
©2016 Simon Wenzel et al., published by de Gruyter.
This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 3.0 License.
