From 7 to 12 September, four doctoral candidates from the DT-HATS project, Arnob Kundu, Yihan Yang, Pedro John and Tommaso Baffetti, participated in two consecutive events hosted at the Université libre de Bruxelles in Belgium: the CYPHER COST Action Training School on Scientific Machine Learning for Digital Twins, followed by the Workshop & Hackathon on Reduced-Order Modeling for Advanced Combustion Systems.
The Training School provided lectures and practical insights into scientific machine learning, with a focus on reduced-order modelling, modal analysis and system identification for reactive dynamical systems. The following Hackathon focused on the development of reduced-order models for a dynamical combustion system.
Working in teams, participants developed and evaluated models based on training data from an externally forced laminar flame, with the submitted approaches assessed in terms of accuracy, computational speed and physical consistency. Pedro John’s team achieved the best performance on the final evaluation data, while Tommaso Baffetti developed the data challenge on Codabench.
We would like to thank Prof. Alberto Procacci and Prof. Alessandro Parente for hosting and organising these events at ULB, and the CYPHER COST Action for creating these opportunities. We also acknowledge the support of the DT-HATS project and its supervisors, whose support enables doctoral candidates to participate in international research activities.