ABSTRACT
The high effort of operating and maintaining existing infrastructure facilities resulted in a large stock of structurally deficient bridges in most industrialized countries. Today, the condition assessment of bridges is conducted mostly manually. To relieve effort and costs, the digitization and automation of conventional manual, labour-intensive methods is necessary. We interpret the ambiguously used term ”digital twin” (DT) in this study as a semantic-geometric model of an existing asset (here, a bridge) that contains all information required for assessing its current condition. This paper proposes an approach to automatically generate the DT of existing bridges from point cloud data (PCD) and images captured from the structure. The PCD of the bridge is semantically segmented by means of ML techniques, and a digital model is created using parametric modeling. Subsequently, detected damages and data from existing bridge maintenance systems are linked to the model to create a full DT. This paper reports the main results of the TwinGen research project: The digital twinning process of bridges can be automated to a large extent, in order to efficiently support the maintenance process of bridges.
