Automated point-cloud to FEM generation for historic masonry bridges
Moving Toward Smart, Resilient and Sustainable Bridges, CRC Press, ss.2383-2388, 2026
- Yayın Türü: Kitapta Bölüm / Araştırma Kitabı
- Basım Tarihi: 2026
- Doi Numarası: 10.1201/9781003778677-288
- Yayınevi: CRC Press
- Sayfa Sayıları: ss.2383-2388
- Orta Doğu Teknik Üniversitesi Adresli: Evet
Özet
FE modeling of historic masonry bridges is difficult since idealized geometry often deviates from actual dimensions due to differential settlement, partial reconstructions, and large de-formations at highly stressed regions. A 3D laser-scan based automated generation greatly en-hances modeling accuracy. 3D scanning is increasingly practical using drones, handheld scanners, photogrammetry, and even low-cost LiDAR. The scanned point cloud can be used as a vessel; slices of it are filled with rectangular, cube-like solid elements to generate the bridge volume in a realistic manner. The objective is to eliminate labor-intensive CAD steps and keep a highly accurate overall geometry modeled as close as possible to the existing shape. 2D sur-vey approximations or shell elements for massive volumes often lead to important modeling and simulation errors. Some pre-processing is necessary before turning the scan into a 3D-FEM. A practical sequence is adopted: registration of interior/exterior scans, segmentation to remove ground and non-structural objects, hole repair on incomplete surfaces, and re-sampling to obtain a clean, uniformly distributed point set. An auto-generator is programmed to define solids by mapping points to a virtual grid to represent the true geometry. Eight-node hexahedral elements are used, and mesh smoothing is also automated for exterior surfaces us-ing the point cloud. Support conditions are defined after mesh generation. The method is vali-dated on a stone masonry bridge. The algorithmic steps, data and runtime considerations, and model export are discussed. The current version prepares an S2K file for SAP2000 input but can be easily adapted to other FEM software. The resulting model may be used for self-weight and vehicle loads, calibration with static/dynamic tests, modal analysis, stress/deflection limit checks, and future nonlinear studies. Restoration–rehabilitation decisions can thus be based on realistic analytical models.