Eight papers at ESCAPE35 in Ghent

Our group presented eight papers at the 35th European Symposium on Computer Aided Process Engineering (ESCAPE35), held from 6 to 9 July 2025 at KU Leuven, Campus Ghent, Belgium.

The papers cover generative AI for engineering diagrams, physics-constrained machine learning, optimization with embedded machine learning models, graph machine learning and large language models for process modeling. All papers are published open access in Systems and Control Transactions, volume 4:

Engineering diagrams (P&IDs)

• D. P. Goldstein, L. Schulze Balhorn, A. A. Alimin, A. M. Schweidtmann. pyDEXPI: A Python framework for piping and instrumentation diagrams (P&IDs) using the DEXPI information model. doi:10.69997/sct.139043
• A. A. Alimin, D. P. Goldstein, L. Schulze Balhorn, A. M. Schweidtmann. Talking like Piping and Instrumentation Diagrams (P&IDs). doi:10.69997/sct.159477
• L. Schulze Balhorn, N. Seijsener, K. Dao, M. Kim, D. P. Goldstein, G. H. M. Driessen, A. M. Schweidtmann. Rule-Based Autocorrection of Piping and Instrumentation Diagrams (P&IDs) on Graphs. doi:10.69997/sct.150968

Hybrid modeling and optimization

• G. Lastrucci, T. Karia, Z. Gromotka, A. M. Schweidtmann. Picard-KKT-hPINN: Enforcing Nonlinear Enthalpy Balances for Physically Consistent Neural Networks. doi:10.69997/sct.108423
• T. Karia, G. Lastrucci, A. M. Schweidtmann. Kolmogorov Arnold Networks (KANs) as surrogate models for global process optimization. doi:10.69997/sct.195815
• S. Rupprecht, Y. Hounat, M. Kumar, G. Lastrucci, A. M. Schweidtmann. Text2Model: Generating dynamic chemical reactor models using large language models (LLMs). doi:10.69997/sct.165009

Graph machine learning

• Q. Gao, D. C. Miedema, Y. Zhao, J. M. Weber, Q. Tao, A. M. Schweidtmann. Bayesian uncertainty quantification of graph neural networks using stochastic gradient Hamiltonian Monte Carlo. doi:10.69997/sct.111298
• M. F. Theisen, G. M. H. Meesters, A. M. Schweidtmann. Transferring Graph Neural Networks for Soft Sensor Modeling using Process Topologies. doi:10.69997/sct.185977

We thank the organizers from KU Leuven and the University of Liège and all co-authors for their contributions.