Newest Research Published in IECR
Excited to share that our paper, “Data-Driven Multi-Objective Optimization of an Industrial Hydrotreater Using a Multi-Output Gaussian Process Surrogate,” has been published in Industrial & Engineering Chemistry Research and featured with ACS cover art.
This work, first-authored by Souvik Ta, develops a multi-output Gaussian Process surrogate trained on multi industrial catalyst cycles from a diesel hydrotreating unit. The model captures coupled responses in sulfur removal, liquid yields, hydrogen consumption, and fuel quality, while providing uncertainty-aware predictions. By integrating the surrogate with NSGA-II, we map Pareto-optimal operating strategies and reveal practical trade-offs for data-driven refinery decision support.
This study reflects our continued effort at the SustainAI Group to advance AI-enabled digital twins, surrogate modeling, and multi-objective optimization for sustainable and efficient chemical processes. Happy to connect with colleagues and partners interested in this direction.
Read the article here:
https://pubs.acs.org/doi/pdf/10.1021/acs.iecr.6c00111
