Topological Relationship Modelling for Industrial Facility Digitisation Using Graph Neural Networks Haritha Jayasinghe University of Cambridge, United Kingdom Ioannis Brilakis University of Cambridge, United Kingdom This is a section of CONVR 2023 - Proceedings of the 23rd International Conference on Construction Applications of Virtual Reality (DOI: 10.36253/979-12-215-0289-3) by Pietro Capone, Vito Getuli, Farzad Pour Rahimian, Nashwan Dawood, Alessandro Bruttini, Tommaso Sorbi Firenze University Press Florence 2023 https://doi.org/10.36253/10.36253/979-12-215-0289-3.88

Available for academic research purposes

Open Access

Copyright Author(s)

Content licence CC BY-NC 4.0

Metadata licence CC0 1.0

This is original content, published for academic research purposes

 Digital edition XML powered by Booksflow

There is rising demand for automated digital twin construction based on point cloud scans, especially in the domain of industrial facilities. Yet, current automation approaches focus almost exclusively on geometric modelling. The output of these methods is a disjoint cluster of individual elements, while element relationships are ignored. This research demonstrates the feasibility of adopting Graph Neural Networks (GNN) for automated detection of connectivity relationships between elements in industrial facility scans. We propose a novel method which represents elements and relationships as graph nodes and edges respectively. Element geometry is encoded into graph node features. This allows relationship inference to be modelled as a graph link prediction task. We thereby demonstrate that connectivity relationships can be learned from existing design files, without requiring domain specific, hand-coded rules, or manual annotations. Preliminary results show that our method performs successfully on a synthetic point cloud testset generated from design files with a 0.64 F1 score. We further demonstrate that the method adapts to occluded real-world scans. The method can be further extended with the introduction of more descriptive node features. Additionally, we present tools for relationship annotation and visualisation to aid relationship detection

 BIM Digital twin GNN machine learning

It is available online at https://doi.org/10.36253/10.36253/979-12-215-0289-3.88

References Agapaki, E., S. M., Glyn-Davies, A., Mandoki, S., Brilakis, I., M., Candidate, P. D., O’, L., & Reader, R. (2019). CLOI: A Shape Classification Benchmark Dataset for Industrial Facilities. 10.1061/9780784482445.009 Agapaki, E., & Brilakis, I. (2022). Geometric Digital Twinning of Industrial Facilities: Retrieval of Industrial Shapes. 10.48550/arxiv.2202.04834 Agapaki, E., Miatt, G., & Brilakis, I. (2018). Prioritizing object types for modelling existing industrial facilities. Automation in Construction, 96, 211–223. 10.1016/j.autcon.2018.09.011 Bloch, T., & Sacks, R. (2020). Clustering Information Types for Semantic Enrichment of Building Information Models to Support Automated Code Compliance Checking. Journal of Computing in Civil Engineering, 34(6), 04020040. 10.1061/(asce)cp.1943-5487.0000922 Buruzs, A., Šipetić, M., Blank-Landeshammer, B., & Zucker, G. (2022). IFC BIM Model Enrichment with Space Function Information Using Graph Neural Networks. Energies, 15(8). 10.3390/en15082937 Hamilton, W. L., Ying, R., & Leskovec, J. (2017). Inductive Representation Learning on Large Graphs. http://arxiv.org/abs/1706.02216 Ismail, A., Strug, B., & Ślusarczyk, G. (2018). Building Knowledge Extraction from BIM/IFC Data for Analysis in Graph Databases (pp. 652–664). 10.1007/978-3-319-91262-2_57 Kipf, T. N., & Welling, M. (2016). Semi-Supervised Classification with Graph Convolutional Networks. 5th International Conference on Learning Representations, ICLR 2017 - Conference Track Proceedings. 10.48550/arxiv.1609.02907 Nguyen, T. H., Oloufa, A. A., & Nassar, K. (2005). Algorithms for automated deduction of topological information. Automation in Construction, 14(1), 59–70. 10.1016/J.AUTCON.2004.07.015 Oh, I., & Kwang, H. (2021). Automated recognition of 3D pipelines from point clouds. The Visual Computer, 37, 1385–1400. 10.1007/s00371-020-01872-y Qi, C. R., Su, H., Mo, K., & Guibas, L. J. (2016). PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation. http://arxiv.org/abs/1612.00593 Sager, C., Zschech, P., & Erlangen-Nürnberg Niklas Kühl, F.-A.-U. (2021). labelCloud: A Lightweight Domain-Independent Labeling Tool for 3D Object Detection in Point Clouds. 10.48550/arXiv.2103.04970 Shin, Y.-M., Tran, C., Shin, W.-Y., & Cao, X. (2021). Edgeless-GNN: Unsupervised Representation Learning for Edgeless Nodes. http://arxiv.org/abs/2104.05225 Son, H., Kim, C., & Turkan, Y. (2015, June 18). Scan-to-BIM - An Overview of the Current State of the Art and a Look Ahead. 10.22260/ISARC2015/0050 Tang, P., Huber, D., Akinci, B., Lipman, R., & Lytle, A. (2010). Automatic reconstruction of as-built building information models from laser-scanned point clouds: A review of related techniques. In Automation in Construction (Vol. 19, Issue 7, pp. 829–843). Elsevier B.V. 10.1016/j.autcon.2010.06.007 Veličković, P., Casanova, A., Liò, P., Cucurull, G., Romero, A., & Bengio, Y. (2017). Graph Attention Networks. 6th International Conference on Learning Representations, ICLR 2018 - Conference Track Proceedings. 10.48550/arxiv.1710.10903 Wang, Z., Sacks, R., & Yeung, T. (2022). Exploring graph neural networks for semantic enrichment: Room type classification. Automation in Construction, 134. 10.1016/j.autcon.2021.104039 Xie, Y., Li, S., Liu, T., & Cai, Y. (2023). As-built BIM reconstruction of piping systems using PipeNet. Automation in Construction, 147. 10.1016/J.AUTCON.2022.104735 Yin, C., Wang, B., Gan, V. J. L., Wang, M., & Cheng, J. C. P. (2021). Automated semantic segmentation of industrial point clouds using ResPointNet++. Automation in Construction, 130. 10.1016/j.autcon.2021.103874