A SMALL dinghy bobs on the ocean a few miles off the Irish coast. Two men in dark clothes lower a pipe into the water, switch on a pump and chemicals start to flow. To the untrained eye, it looks distinctly dodgy. But these men are from the Irish navy and the chemicals they are dumping are creating a simulated oil spill, to be sniffed out by a team of robots lurking in the depths below.
I’m aboard LÉ RĂłisĂn, an Irish Naval Service patrol ship, with a team of scientists who think they have a uniquely effective way of dealing with oil spills.
The project leader is Javier Gilabert of the Technical University of Cartagena, Spain. Inspiration came to him following the huge Deepwater Horizon spill in 2010. The accident drove home how strangely oil can behave in the water. Much of it didn’t rise straight from the ruptured well to the surface but remained trapped in the water column, where it spread out and then “popped up in unexpected places”, says Gilabert.
Advertisement
Emergency responders can use satellites to spot the rough location of oil on the surface, then track it closely using radar on planes or by eye from boats. This helps them decide on their tactics: whether to use floating booms to hem in the oil and vacuum it up, or add chemicals called dispersants to help dissolve the oil. But we have no good way of tracking oil underwater, and it is tough to make those calls without knowing if there is a lot of oil sloshing around out of sight.
Gilabert’s idea was to deploy a team of autonomous underwater vehicles (AUVs) with instruments that can detect the way oil fluoresces when you shine a light on it. They can then provide responders with an evolving 3D map of the plume.
To simulate the spill, the naval officers in the dinghy are releasing a dye called rhodamine, which has chemical bonds that fluoresce in a similar way to oil. The five AUVs then begin to criss-cross the patch of dye.
The difficult part is getting the communications to work, both so the AUVs know their position and so they can relay information back. Radio signals don’t travel well underwater but sound does. So the team send signals via Wi-Fi to floating stations, which then relay them to the AUVs using an acoustic signal. Soon things are working and researchers with laptops on plastic tables have a 3D model of the spill.
This isn’t the first exercise of its kind. Gilabert and his team also tested things out in the Mediterranean in 2015 and 2017. But there they worked from a special ship that could use its thrusters to remain still in the water. This time things were more challenging and realistic.
The navy ship is built primarily for speed, not stability, and the rocking of the antennas meant that the team twice lost contact with the AUVs, some of which are worth more than ÂŁ100,000 because of the sophisticated sensors on board. On the second day, tests had to be called off; the black waters of the Atlantic were so rough that the AUVs could easily have smashed against the ship while being retrieved using mechanical winches.
There are ways of fixing those problems. You might use a static balloon as a communications base station, for example, says Gilabert. And to get around the problems of retrieving the AUVs, you could program them to head to shore on their own. “We have learned many things and we’re close to making this a reality,” Gilabert told the team below decks at a wrap-up meeting.
Plenty of people are interested in the project, which is known as Enhanced Underwater Robotics Ready for Oil Spill (eURready4OS), including professional response firms funded by the oil industry to help clean up big spills.
Then again, you can only learn so much without working with real oil, which is why – just occasionally training exercises like this release the real stuff into the ocean. Industry-funded cooperative Oil Spill Response Limited is planning one such exercise in the next few years.
It may sound like a terrible idea, but when the next big spill happens we need to be ready for it.
This article appeared in print under the headline “Robots hunt for oil spills underwater”
