2026-10-06T22:36:29Z https://api.figshare.com/v2/oai

oai:figshare.com:article/32641263 2026-10-01T16:24:36Z portal_1172 item_type_8 month_year_10_2026

Scour depth prediction, using computer vision at a FlexiArch bridge Ben Millar (12857222) PUREID: 641402470 Scour masonry arch bridge computer vision finite element modelling FlexiArch neural network prediction Scour induced failure of masonry arch bridges is the most common cause of bridge collapse, especially during extreme flood events. Over recent years there has been a significant number of documented scour-induced failures of this bridge stock across the world. The extent of the scour can erode the supporting soil or sediment around the bridge piers or abutments, leading to a loss of support for the structure. As the foundation becomes unstable, it can result in settlement of the bridge, potentially causing structural damage, while the ongoing nature of scour leads to the progressive deterioration of the bridge’s ability to safely transfer load. The development of the scour hole is not immediately obvious without the use of monitoring techniques since it is below the water line. Traditional monitoring methods are expensive to install, monitor and maintain as well as being sensitive to external interference and have the potential to be damaged during flood events. This research links the changes in arch displacement to the easement of support at the bridge foundation due to scour. Scour predictions are generally made using empirical equations which do not provide accurate estimations of scour and scour modelling has not reflected the true scour profile created by the fluid-structure interaction at individual bridges. Scale experiments and finite element modelling on a typical FlexiArch under loading with realistic simulated scour conditions and accurately modelled 3D scour, informed the development of a framework for computer vision-based monitoring (ArchIMEDES) and an artificial neural network (Keystone). Scour depth and load carrying capacity predictions are made by Keystone for a monitored bridge based on the observed deformation in the arch ring extracted using ArchIMEDES. This contributes to the development of a low cost, non-intrusive sensor to identify scour development so it can be investigated and remediated before a bridge collapses. <br><br><i>Thesis is embargoed until 31 July 2030</i> 2026-10-01T16:24:36Z Text Thesis 10.17034/32641263.v1 https://figshare.com/articles/thesis/Scour_depth_prediction_using_computer_vision_at_a_FlexiArch_bridge/32641263 All Rights Reserved Open Access after 2030-07-31