Physics‐Informed Digital Twins: Enhancing Concrete Structural Assessment Based on Point Cloud Data
Year: 2025
Type: article
Abstract: Finite element modeling is widely regarded as an effective method for simulating structural responses, but maintaining geometrical consistency with damaged physical structures remains insufficiently explored. This paper proposes a new physics‐informed digital twin framework for concrete structure modeling and implements the twinning/synchronization process between the physical model and its counterpart finite element analysis (FEA) model. This framework starts with point cloud scanning for damage and point cloud processing. Subsequently, a direct mapping method called Voxel–Node–Element (VNE) is proposed, which can improve mapping efficiency and reduce mapping errors. Furthermore, a multiscale modeling method is adopted to enhance digital twin modeling updates, dramatically reducing the number of elements and improving computational efficiency. An experimental case study was conducted to evaluate this method, showing good alignment between point cloud and physics models with a geometric error of less than 5%. Additionally, computational efficiency was improved by 95% compared to traditional methods. This method can also be used for full‐scale structure modeling, which was validated in the case of damage updates for large bridges. This study enables a highly accurate and efficient method for updating digital twin models. This capability was validated through damage updates applied to large‐scale bridge structures.
DOI: 10.1155/stc/5605927
Authors: Honghong Song, Xiaofeng Zhu, Haijiang Li, Gang Yang, Tian Zhang
Language: English
Subfield: Civil and Structural Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $2,490
Awards: G2350806196, G6335587300
OpenAlex ID: W4413217460