A new collaboration between N-Sea and the University of Strathclyde will research the potential of deep learning in automating subsea pipeline inspection
Subsea inspection services have been transformed by the advent of cheaper, more accurate sensors and higher-resolution imaging equipment. The ease with which operators can view their assets and the measurements that can be taken mean that installations are being examined and tracked in greater detail than ever before.
Yet some elements of this process remain manual. The annotation and interpretation of the images and data recorded are done by ROV pilots and engineers, limiting the speed at which the results can be compiled and passed on to the client.
A new tie-up between a Netherlands-headquartered subsea services firm and a Scottish university hopes that the application of deep learning could provide a breakthrough. Services provider N-Sea and the University of Strathclyde’s Institute of Sensors, Signals and Communications have been awarded new funding from the Data Lab Innovation Centre as part of a project to explore methods of automating subsea pipeline inspection. N-Sea is known for its work as an independent offshore subsea contractor, specialising in IMR services for the oil and gas, renewables and telecoms industries. Combining its experience in IMR (inspection, maintenance and repair) and analysis with the university’s data analytics capabilities, N-Sea’s hope is that computer systems will be able to process images and footage, spot patterns and report on pipeline integrity without human operators poring over the material.
Sea here Although deep learning has been applied to subsea operations before – BP’s use of GE’s Predix software, for example, to enable communication and data-gathering between hundreds of subsea wells – the project backers have affirmed that the approaches and techniques they are exploring have “never before been attempted in the subsea environment.” N-Sea survey and inspection data centre manager, David Murray, explains: “We hope [the project] will transform pipeline inspection operations… By working closely together, we aim to automate the inspection process and operate ROVs at previously impossible speeds.”
Commenting on the university’s partnership with N-Sea, Chancellor's Fellow at the University of Strathclyde, Dr Christos Tachtatzis, said: “Our research group is excited to partner with N-Sea in this innovative project for the hostile and extreme subsea environment. Deep learning approaches have great potential to speed up the inspection processes which remain manual and labour-intensive.”
“We have a long tradition of conducting industrially relevant research with high economic, environmental and societal impacts and this is a prime example of the challenges we seek to provide solutions for. Owing to the support of the Data Lab, we are looking forward to a fruitful collaboration with N-Sea.”