Abstract
Additive manufacturing has enabled the fabrication of advanced reactor geometries, permitting larger, more complex design spaces. Identifying promising configurations within such spaces presents a significant challenge for current approaches. Furthermore, existing parameterizations of reactor geometries are low dimensional with expensive optimization, limiting more complex solutions.
To address this challenge, we have established a machine learning-assisted approach for the design of new chemical reactors, combining the application of high-dimensional parameterizations, computational fluid dynamics, and multi-fidelity Bayesian optimization. We associate the development of mixing-enhancing vortical flow structures in coiled reactors with performance and used our approach to identify the key characteristics of optimal designs. Our results demonstrate that coupling advanced manufacturing techniques with ‘augmented intelligence’ approaches can give rise to reactor designs with enhanced performance.
Additionally, in order to include valuable physical insights from the domain in this talk, we will introduce collaborative Bayesian optimization that re-integrates human input into the data-driven decision-making process. By combining high-throughput Bayesian optimization with discrete decision theory, experts can influence the selection of experiments via a discrete choice. We propose a multi-objective approach to generate a set of high-utility and distinct solutions, from which the expert selects the desired solution for evaluation at each iteration.
Our methodology maintains the advantages of Bayesian optimization while incorporating expert knowledge and improving accountability. The approach is demonstrated across various case studies, including bioprocess optimization and reactor geometry design. This indicates that even with an uninformed practitioner, the algorithm recovers the regret of standard Bayesian optimization. The proposed method enables faster convergence and improved accountability for Bayesian optimization in engineering systems by including continuous expert opinion.
Biography
Antonio del Rio Chanona is head of the Optimisation and Machine Learning for Process Systems Engineering group at the Department of Chemical Engineering, Imperial College London. He is the Director of Education at the Sargent Centre for Process Systems Engineering and Co-director of the Centre for Doctoral Training in Next Generation Synthesis & Reaction Technology (rEaCt).
Antonio’s primary research interests include data-driven optimization, reinforcement learning, control, and hybrid modeling, which are applied to process systems engineering.
Antonio received his MEng from UNAM in Mexico and his PhD from the University of Cambridge, where he was awarded the Danckwerts-Pergamon Prize for the best doctoral thesis of his year. He received the EPSRC fellowship to adopt automation and intelligent technologies into bioprocess scale-up and industrialization and has received awards from the International Federation of Automatic Control (IFAC), the Institution of Chemical Engineers (IChemE) and the Association of European Operational Research Societies (EURO) in recognition for research in areas of process systems engineering, industrialization of bioprocesses, and adoption of intelligent and autonomous learning algorithms to chemical engineering.
Head of the Optimisation and Machine Learning for Process Systems Engineering group, Department of Chemical Engineering, Imperial College