Oil & gas inspections with Hadron Collider robots
Ross Robotics’ Philip Norman and Nick McCormick talk to Khai Trung Le about their collaborative modular robotics and imaging inspection system.
Amodular robotic inspection system that had courted interest from Network Rail, Sellafield and the CERN Large Hadron Collider, among others, hopes to provide safe and flexible inspection for the oil and gas and nuclear industries. Philip Norman, Director of R&D at Ross Robotics, and Dr Nick McCormick, Principal Research Scientist in Structural Health Monitoring at the National Physical Laboratory (NPL), UK, profiled their technologies that pairs Ross Robotics’ Scale 2.0 modular robot technology and the NPL DIFCAM (Digital Imaging for Condition Asset Monitoring) system.
Cost reduction and automation are not the only drivers behind the technology. In the past ten years, 22 people have died in accidents related to oil and gas facility inspections. Norman sees the robot as an opportunity to reduce incidents. He said, ‘There is the consideration of liability in performing these inspections – not only efficiency but in not endangering people in hazardous environments. It’s also very disruptive for the oil and gas industry to have to prepare for inspections. We remove the cost and inconvenience, as well as the safety issues.’
DIFCAM has been primarily developed with Network Rail and Omnicom, a Balfour Beatty company, and has successfully completed a series of feasibility tests in several situations, including Old Dalby Test Track, the Mier Tunnel at Stoke-on-Trent and pilot studies at the Haymarket tunnel near Edinburgh. The system uses a combination of high-resolution photography and low-resolution LIDAR to create a high-density dataset. McCormick told Materials World, ‘The work we did on the tunnels typically used ten cameras in each arc, capturing a ten megapixels photo every metre of the tunnel per camera. That equates to a resolution of around 1mm per pixel, and the LIDAR captures a million datapoints per second. Inspecting a 700m tunnel, creates a dataset around 700,000 pixels long and 200,000 pixels wide, with 700 million position datapoints.’
Although the DIFCAM system was developed separately from the Ross Robotics Scale 2.0 modular robot system – which has previously been used in CERN’s Large Hadron Collider because of its strong tolerance to magnetic fields and background radiation – the pair began collaborating in 2015. Norman stated that much of the Ross Robotics system’s appeal came from its modularity. ‘It reduces downtime, and the significant expense associated with it, by simply substituting modules,’ he said.
The robotic system is not expected to replace human inspection, but rather augment it. McCormick said, ‘The main element is to capture multiple panoramic data sets, use the computer to do the donkey work of comparing sets for differences and show the areas of change to a human inspector who, through experience and training, can identify whether those differences are significant or not.’
Reducing the cost of inspection is less likely for rail tunnel inspections, with McCormick stating, ‘It’s quite difficult to make inspection any cheaper. It’s essentially the cost of a single human going through the tunnel.’ But is of greater appeal for oil and gas and nuclear inspection. Norman added, ‘The other valuable element to modularity is being able to reconfigure the system. Nick and I were initially looking at large tanks in the oil and gas industry, but the reality is those assets vary enormously. We have to flexibility to take the same modules but put them together on a smaller robot.’
The DIFCAM-equipped Ross Robotics system is currently undergoing a feasibility project, but McCormick hopes to see wider adoption soon. ‘We’re already exploring follow-on projects to get closer to applications, but those are still intended to run over a couple of years. We’re aiming for rollout and implementation in a few years rather than decades.’