LAUSR

Abstract The aim of this paper is to quantify the effect of geometric uncertainty on the structural behaviour of arches developed using traditional geospatial and remote sensing techniques. Geometric models of twenty-five dry-jointed, rigid block arch specimens were developed using the proposed “Image2DEM” framework, encompassing the Structure-from-Motion (SfM) photogrammetry pipeline. Assessment of the framework was carried out with respective geometric models developed from tape measurements, in terms of geometry and structural behaviour. The geometric variables investigated were the: a) joint inclination angle; b) joint mid-point location; c) joint length; d) block volume; and e) block centroid location. Concerning structural behaviour, experimental testing of the arch specimens was numerically simulated with the Discrete Element Method (DEM), and the stiffness, load multiplier and normal forces between joints were obtained. Results showed that even small variations (approx. 8%) between geometric models developed from SfM photogrammetry and tape measurements can influence the collapse load and stiffness of the arch by more than 10% and 46% respectively. Also, of the geometric variables investigated, the joint inclination angle was found to influence the collapse load the most. These findings highlight the importance of developing accurate geometric models, and subsequent employment of accurate geometric data acquisition techniques, to reliably capture the structural behaviour of arches. Although an investigation on arch specimens constructed in the laboratory is demonstrated here, the developed outcomes have important implications for the broader topic of data-driven masonry diagnostics using SfM photogrammetry and high-level numerical modelling using micro-modelling strategies such as the DEM.