Treffer: Integrated dual adaptive control of continuous chromatographic separation processes via reinforcement learning.

Continuous chromatographic processes involve nonlinear dynamics, cyclic operation, and strong interactions between operating variables, which make real-time control inherently complex. As process variability and integration increase, there is a growing need for adaptive control frameworks that can maintain performance under changing conditions. In this work, a reinforcement learning (RL) control framework is developed for adaptive operation of continuous chromatographic systems. The controller is trained based on a mechanistic simulation environment and learns to coordinate multiple process inputs based on observed system states. A phase-dependent reward formulation captures the trade-offs between product purity, yield, and operational efficiency, guiding the agent toward stable, high-performance operation from process start-up. The RL controller maintains the target purity under feed variations, flowrate disturbances, and measurement noise, achieves cyclic steady state (CSS) within 10 process cycles, improves process yield from 86 % to 91 %, and increases product recovery by 79 % compared to nominal operation. These results demonstrate that RL can deliver robust and adaptive control for complex multivariable chromatographic systems, providing a framework for intelligent and adaptive process operation.
(Copyright © 2026 The Authors. Published by Elsevier B.V. All rights reserved.)

Declaration of competing interest The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Maria M. Papathanasiou reports financial support was provided by Engineering and Physical Sciences Research Council. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.