Development of mass, energy, and thermodynamics constrained steady-state and dynamic neural networks for interconnected chemical systems

• Published in: Chemical engineering science Vol. 309; no. C; p. 121506
• Main Authors: Mukherjee, Angan, Bhattacharyya, Debangsu
• Format: Journal Article
• Language: English
• Published: United Kingdom Elsevier Ltd 01.05.2025; Elsevier
• Subjects: Energy conservation; Interconnected systems; Mass conservation; Neural networks; Thermodynamics constraints; Thermodynamics constraints; Neural networks; Energy conservation; Mass conservation; Interconnected systems
• ISSN: 0009-2509; 1873-4405
• DOI: 10.1016/j.ces.2025.121506

•Neural network models developed for exactly satisfying mass, energy, and thermodynamics.•Algorithms developed for training and forward problems for steady and dynamic data.•Best thermodynamics model selected from a family of candidates given the data.•Studies conducted by training against complex dynamic noisy measurements.•Constraints exactly satisfied even when training data violate those constraints. This paper discusses the development of steady-state and dynamic modeling algorithms for mass, energy, and thermodynamics constrained neural networks (METCNNs) for interconnected chemical process systems. The METCNN models can ‘exactly’ conserve the overall system mass and energy balances, as well as certain thermodynamics constraints during both training and forward problems. The proposed approaches can accommodate an outer layer integer programming problem for selection of the best thermodynamics model from a family of candidates given a particular transient dataset. The developed algorithms for both steady-state and dynamic METCNNs are tested for an interconnected chemical system in presence of noise and bias in training data. For all case studies considered in this work, it has been observed that the optimal METCNN models ensure exact conservation of system physics and consistently converge close to the system truth, even when trained against complex dynamic noisy measurements that do not necessarily satisfy the system physics.