Abstract Geometric dimension inspection and overall alignment evaluation of arch ribs remain challenging during the manufacturing and installation of prefabricated arch bridges. Enabled by three-dimensional laser scanning for rapid and accurate capture of complex geometries, this study proposes a rapid alignment evaluation method based on virtual trial assembly for prefabricated arch bridges. First, the high amount of noise in flange-plate plane point cloud data is mitigated by a SVD-based plane-fitting denoising algorithm. Second, geometric features of the bolt holes and the flange plane are obtained by applying an alpha shape-based feature extraction method to the flange-plate point cloud. Subsequently, under a “face-to-face, hole-to-hole” flange-plate alignment mode, the flange assembly error is constrained to the submillimeter level through coarse assembly that aligns flange-plate centers and fine assembly that aligns bolt hole centers. Finally, an axis extraction algorithm for the arch rib’s main chord tube is proposed. By utilizing the geometric properties of curved members with near-circular cross sections and adjusting the slicing plane orientation, the method overcomes axis extraction challenges and enables the efficient recovery of the main chord tube axis. The method has been successfully applied to Shuangbao Bridge, enabling reliable detection of the overall alignment of multiple segments during arch rib fabrication. Rapid inspection and high precision are achieved, yielding a threefold improvement in measurement efficiency over traditional methods while maintaining alignment accuracy within 2.6 mm.
Zhu et al. (2026) studied this question.