AUTOMATED REVERSE Modeling and examination of as-built steel truss based on point cloud data: Machine learning and bim
Master thesis
Published version
Permanent lenke
https://hdl.handle.net/11250/2774033Utgivelsesdato
2021Metadata
Vis full innførselSamlinger
Sammendrag
Building Information Modeling (BIM) has improved the way structures are designed and constructed over the last decades. Although BIM implementation is an expected part of newer construction projects, the technology is heavily under-represented for use cases other than design. The research presented in this paper has a basis in creating a reliable way of reverse modeling an as-built steel truss to evaluate whether or not the component is suited for use. To do so, a reverse model of the as-built steel truss is created using scanned point cloud data. Results find the proposed method to deliver very accurate results. Digitally examining the reverse modeled steel truss ensures that the component is very similar to the designed model, and further FEA concludes that the component is suited for use. The proposed method lays the foundational ground for using modern technologies such as 3D laser scanning, parametric BIM modeling, machine learning, computer vision, and other BIM software to develop algorithms, scripts, and workflows for reverse modeling and examining as-built steel components digitally. However, making improvements and building on top of the proposed method will open up the possibilities for the technology to be used for other purposes than described in this paper.