Reinforcement Learning-driven Process Intensification Synthesis – Design and Optimization of Reaction/Separation Systems
Year: 2026
Type: conference-paper
Abstract: This work aims to systematically generate intensified process designs by integrating reinforcement learning (RL)-driven process synthesis and phenomena-based modeling via Generalized Modular Framework (GMF). Rather than considering flowsheet synthesis with conventional unit-operations, GMF utilizes fundamental building blocks, also known as mass and heat exchange modules, to describe the physiochemical phenomena and to enhance novel process discovery. At its core are driving forces which characterize the mass transfer feasibility based on the total change in Gibbs free energy of the system. RL is integrated with this phenomena-based modeling strategy to drive flowsheet generation by exploring much of the total action space and minimizing pre-postulation of stream connections. All possible inlets, outlets, and interconnections between modules are contained in a stream matrix. Deep Q-Network is used as the RL agent, which contains a multi-layer convolution neural network followed by a multi-layer feedforward neural network. This network computes q values for each possible stream connection contained in the stream matrix and determines an optimal action from the maximum q-value (e.g., varying the stream interconnections). New designs can thus be generated and iteratively improved. This approach is demonstrated using two case studies: (1) binary separation of benzene and toluene, and (2) membrane-assisted reaction for methanol production.
Source: Systems and Control Transactions
DOI: 10.69997/sct.168270
Language: English
Subfield: Control and Systems Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W7165398197