Overview

This research explores the integration of machine learning tools, such as recurrent neural networks and ensemble learning, into nonlinear model predictive control (MPC) systems to enhance control performance, process operational safety, and cybersecurity for chemical processes.

The study presents a machine-learning-based predictive control framework, novel MPC structures for nonlinear processes, and a two-tier detector-controller architecture. These advancements aim to improve process economics, safety, and robustness against cyber-attacks, demonstrating their applicability in industrial applications like deposition processes and additive manufacturing.

Speaker

Panagiotis D Christofides, Distinguished Professor, University of California

Panagiotis was born in Athens, Greece, in 1970. He received his Diploma in Chemical Engineering from the University of Patras in 1992, and his MS degrees in Electrical Engineering and Mathematics, and a PhD in Chemical Engineering from the University of Minnesota between 1995 and 1996.

Since 1996, he has been with UCLA, where he is currently a Distinguished Professor and Department Chair in the Department of Chemical and Biomolecular Engineering, and a Distinguished Professor in the Department of Electrical and Computer Engineering. His research spans nonlinear and predictive control, distributed parameter systems, multiscale systems, hybrid systems, and machine learning, with applications to chemical processes, advanced materials processing, particulate processes, energy, and water systems.

He has published extensively, holding a Google Scholar h-index of 87, and has guided about 100 graduate students, including 50 PhD graduates. His work has received numerous accolades, including awards from the National Science Foundation, Office of Naval Research, and the AIChE. Professor Christofides is a Fellow of AAAS, AIChE, IAAM, IEEE, and IFAC, and has served on editorial boards of leading journals and conferences.

The material presented has not been peer-reviewed. Any opinions are the presenter’s own and do not necessarily represent those of IChemE or the ‘insert group name’ in full. The information is given in good faith but without any liability on the part of IChemE.

Time

16:00–17:00 BST.

Software

The presentation will be delivered via Microsoft Teams. We recommend downloading the app from the Microsoft website, rather than using the web portal.