End-to-End Reinforcement Learning of Koopman Models for eNMPC of an Air Separation Unit
This work addresses demand response optimization for industrial processes like air separation units, but it is incremental as it extends a previously proposed method to a more complex case study.
The paper tackles the challenge of scaling a reinforcement learning-based method for training Koopman surrogate models to a large-scale air separation unit, showing that it achieves similar economic performance to a system identification-based approach while avoiding constraint violations.
With our recently proposed method based on reinforcement learning (Mayfrank et al. (2024), Comput. Chem. Eng. 190), Koopman surrogate models can be trained for optimal performance in specific (economic) nonlinear model predictive control ((e)NMPC) applications. So far, our method has exclusively been demonstrated on a small-scale case study. Herein, we show that our method scales well to a more challenging demand response case study built on a large-scale model of a single-product (nitrogen) air separation unit. Across all numerical experiments, we assume observability of only a few realistically measurable plant variables. Compared to a purely system identification-based Koopman eNMPC, which generates small economic savings but frequently violates constraints, our method delivers similar economic performance while avoiding constraint violations.