A novel hybrid approach for alignment modelling of virtual trial assembly of steel box girder bridge

Inspecting completed steel box girder bridges is a standard requirement. Before construction, the steel box girder is pre-assembled to ensure its alignment meets construction requirements. However, current inspection processes still largely rely on manual labour and are subject to the expertise of engineers. This paper presents a hybrid modelling approach that combines geometric and statistical methods to predict the bridge shape in the Virtual Trial Assembly (VTA) process using extracted corner points from great amount point cloud data. Results from a case study show that this hybrid modelling approach can accurately and efficiently extract the corner points from steel box girders and model the bridge alignment in VTA with maximum deviation of 1.6mm for more than 90% of point data and 1.4 mm for corner points, while requiring less data collection and modelling time. This method provides a more accurate and effective solution for arch ribs inspection and line shape prediction of steel box girder bridges, accelerating construction progress and improving construction quality.