Semi-automated discrete-element modelling of arch structures incorporating SfM photogrammetry

Abstract

This paper presents Image2DEM, a framework for the semi-automated structural analysis of arches. Numerical models of arch specimens were semi-automatically developed from structure-from-motion (SfM) photogrammetry, by employing image-processing techniques, to bypass laborious computer-aided design (Cad)-based processes. Then, the numerical models of the proposed framework were assessed by comparison with a conventional Cad-based framework in terms of (a) geometrical agreement (i.e. joint and block properties) and (b) structural capacity agreement (i.e. stiffness, load multipliers and normal forces between joints at each hinge formation). The results firstly demonstrated SfM photogrammetry as an effective avenue of structural surveying for numerical modelling. Afterwards, the numerical models developed from the proposed and conventional frameworks were demonstrated to have good agreement in terms of geometry. Furthermore, good agreement was also found in the predicted structural capacity, with differences in structural capacity of less than 7%. Although only the potential to capture both the geometry and structural capacity of small-scale arch structures was demonstrated here, this study lays the foundation for automated geometry acquisition for arches for their structural analysis.