Overview 

Engineering diagrams are crucial for chemical process development, specifically through Process Flow Diagrams (PFDs) and Process and Instrumentation Diagrams (P&IDs). Despite advanced CAD software, these diagrams are often stored as PDFs or images, making them non-machine-readable and difficult to retrieve. To address this, we propose deep learning-based computer vision algorithms to digitise PFDs and P&IDs. Our algorithms detect unit operations, classify symbols using an ontology, and identify connections. It also extracts and associates text with relevant components. A web application enables users to upload engineering documents and obtain digital versions in formats like DEXPI, GraphML, and CSV.

This webinar celebrates the award of the 2024 Senior Moulton medal for the publication in Digital Chemical Engineering by the speaker and co-authors Maximilian Theisen, Kenji Nishizaki Flores and Lukas Schulze Balhorn.

Speaker

Artur M Schweidtmann, Assistant Professor, Delft University of Technology

Artur is a tenured assistant professor for chemical engineering at Delft University of Technology and director of the Process Intelligence Research lab. His research focuses on the combination of artificial intelligence (AI) and chemical engineering. He received his Master of Science from RWTH Aachen University in 2017 and defended his PhD from RWTH in 2021, both in chemical engineering. During his studies, he spent the academic year 2013/2014 at Carnegie Mellon University as a visiting student via the DAAD ISAP program. He performed his Master thesis at the University of Cambridge.

Any opinions are the presenters' own and do not necessarily represent those of IChemE or the Computer Aided Process Engineering Special Interest Group. The information is given in good faith but without any liability on the part of IChemE.