Flexible Pipe Riser Torsion Detection in Underwater Video Images

When flexible pipelines are subjected to high axial loads and exposed to a mixture of CO2 and water, their metallic armor can corrode and crack, leading to a twisting deformation that, if left untreated, may cause collapse. Monitoring the entire length of the pipeline for twisting is crucial for early detection, but manual assessment is challenging due to the subtle nature of these anomalies and the analysis of video footage might take hours. This work aims to automate the inspection of flexible pipeline twisting by proposing an algorithm that processes video data obtained from standard pipeline inspection procedures and accurately detects sections of twisted pipelines. The algorithm was validated using simulated data and successfully detected twisting of a magnitude of 3mm on a 300mm pipeline. Results on real data were shown to be consistent with experts’ opinion, however further development is required to ensure reliability in the presence of other equipment and marine life.