Getting fishy

Elaine Maslin
Friday, April 1, 2016

The race is on to create an autonomous, resident underwater vehicle capable of long-term inspection operations and eventually a level of intervention. Elaine Maslin takes a look at BG Group’s version.

FlatFish during shallow water seatrials. Photos from BG Group.

Late last year, a select few were given a preview of a technology many would like to get their hands on – a resident subsea autonomous inspection vehicle.

The industry has an ever increasing amount of infrastructure on the seafloor, in ever deeper depths. Getting at it, for inspection and intervention, means hiring remotely operated vehicles (ROVs) and paying the day rates for the vessels that support them. The deeper ROVs have to go, the bigger they tend to get, and the costlier the support.

Operators are looking for a new tool for the job, and the Holy Grail is a form of autonomous underwater vehicle (AUV) that can get up close and personal, hovering on station without a tether to the surface, or indeed a support vessel. The aim is for a machine that can be left on the seafloor and sent on missions, autonomously collecting data and information, and actively replanning its route according to what it finds, such as a hydrocarbon leak or an unknown object.

Creating such a vehicle has been a goal for the industry for some time. Eni has been working on its own AUV technology, Clean Sea, which the firm says is conceptually similar to a drone, but in the marine environment. OE has also covered Subsea 7’s autonomous inspection vehicle (AIV) development and Total’s inspection AUV project in the past year (OE: October 2014). Late last year, Shell announced a US$7 million prize, the Shell Ocean Discovery XPrize, to “accelerate breakthroughs for rapid and unmanned ocean exploration.”

Shell could already have what it wants under its nose via BG Group, which it officially acquired in February. Over the past two years, BG Group has been developing the FlatFish – a lightweight, subsea resident autonomous vehicle, which it eventually hopes to deploy from floating production units, platforms or from a subsea docking station. The first prototype was built and tested in Germany, and last year BG Group tested a second, tethered prototype, in shallow waters offshore Bahia state, Brazil.

The first FlatFish is a 275kg unit rated to 300m water depth and measuring 220cm-long and 105cm-wide. It’s been designed to be able to run for 16-24 hours, covering up to 20km, including a return journey. It runs on Lithium-Ion batteries, and has six hubless ring thrusters. It has lighting, a laser line projector, sonar, four cameras, depth sensor, altitude and heading and inertial heading (INS) referencing systems, Doppler velocity log (DVL) and multiple communications systems. The navigation system and payload system will be separate, but connected, via Ethernet.

Adam Hillier

While two years might not seem like a long time to develop this unit, BG Group chose the Brazilian Institute of Robotics in Salvador as a partner, and has been working with the Robotics Innovation Centre, DFKI, Bremen, Germany, which has a 3.5 million liter seawater test tank, and the Robotics Group, at the University of Bremen, which has meant it can build on at least a decade worth of work into artificial intelligence and robotics, says Adam Hillier, BG Group’s chief technology officer.

He says getting a resident AUV is an industry goal. “It is one of the technologies the industry is crying out for and we see a need for it. There is a big relevance to our assets in Brazil and on other fields around the world where this could be of interest.”

What sets FlatFish apart, he says, is the level of intelligence built into the software on the machine. “It is the software that drives the navigation controls and the ability to interpret and optimize route plans during a mission, as well as processing technology for 3D, high-precision images. It’s difficult to judge, [in comparison to] others, but certainly FlatFish is exciting. It’s the ability to be down there almost permanently and move away from the need for a specialist crew and a vessel. ROVs exist and are quite mature, but they have limitations.”

Normal ROVs or AUVs either position themselves acoustically, using a mother vessel as a reference or using its INS or DVL with dead-reckoning, which can accumulate an error over time. FlatFish uses a combination of INS/DVL, ultrashort baseline (USBL) acoustic, identification of pipelines and umbilicals, and optical-acoustic feature detection to navigate. In fact, it’s this navigational technology that has held back previous efforts to create such a unit, according to a paper presented at Oceans ’15 in Washington in October.

The software architecture has been designed so that components are integrated and coordinated, but have a single purpose, so fault finding is easier. It also uses Syskit, a model-based approach which allows the different configurations of the systems component networks to be combined into more complex systems, with different subsystems able to be switched on and off, according to the Oceans ’15 paper.

“The software is essentially the brains of the AUV, which controls navigation, real-time plan optimization and the high precision 3D image reconstruction and interpretation,” Hillier says. The ability to hover is also key, enabling closer and clearer structure inspection as well as, high definition images to be captured, he adds.

FlatFish on display in Brazil. 

The various AUV offerings on the market differ in size and shape, with few looking like a traditional AUV. FlatFish is no different in that sense. “We have been going through tank testing and everything is looking good in that space. It’s a combination of being able to get the right hydrodynamic characteristics against being able to package the sensors in the right space.”

The second FlatFish prototype was demonstrated at the Senai CIMATEC University campus in Bahia, Brazil, in December. The current testing, offshore Bahia, is putting the FlatFish to work looking at a number of ship wrecks, to test controls, imaging and image processing, and creating 3D images of the wrecks.

“The next stage is building on the experience of Phase 1 and extending it in to slightly deeper water and starting to think about the docking station,” Hiller says. “The docking station isn’t everything. FlatFish could be useful without being permanently resident. There’s also a driver to get FlatFish in the field as soon as possible. That could happen before the docking station is finalized. All of this is due to be determined in the next year.”

Some of the other resident AUV concepts have looked at deployment in a cage or having it dock on infrastructure in place. Another option could be setting it off in shallow waters to its deeper water destination. “The docking station is the ultimate goal, but it is key to get the interfaces right, for power and data, with existing subsea infrastructure,” Hiller says. “Certainly, if we are going to retrofit it, there are going to be quite a few challenges. Green field would be a bit different.”

At the moment, the plan is for data collected from the AUV to be downloaded when it returns from a mission, either via docking or by being retrieved back to a vessel. But, Hiller says: “There’s potential for data to be transferred during the mission, which might be a future development.”

To be able to intervene on infrastructure is also a longer term goal for FlatFish, as it is for Subsea 7’s AIV. Subsea 7 has set out a four-step vision for its technology, to “sense, see, touch, and do.” “It’s not part of the initial version of FlatFish, but it’s part of future versions,” Hiller says. “We share that vision. I don’t think we see the next version as doing any significant intervention, but daughters of FlatFish will be going there.”

For long-term deployment, there will also be practical considerations to consider, such as bio-fouling, degradation of seals and the blockage of moving parts, according to the Oceans ’15 paper.

The wider operator community is interested in BG’s work, Hiller says. But the next step will be working out a commercialization strategy. “We know we might need another partner, maybe more, but we want to move quickly,” Hiller says. “We think we are on the cusp of something special.”


The Shell Ocean Discovery XPrize is a three-year global competition challenging teams to advance ocean technologies for rapid and unmanned ocean exploration. An additional US$1 million bonus prize is being put forward by the National Oceanic and Atmospheric Administration (NOAA) for teams that demonstrate their technology can “sniff out” a specified object in the ocean trough biological and chemical signals.

Announced in December 2015, the three years incorporates nine months for team registration, 12 months for initial solution development, and 18 months for complete two rounds of testing and judging. Tasks will include making a bathymetric map, producing high-resolution images, and identifying archaeological, biological or geological features. Reliability and durability will also be assessed down to 4000m water depth.

Categories: Technology Subsea Robotics Software Support Vessel Europe South America ROV & Dive Support Noaa Vessel

Current News

BW Energy Signs $300M Reserve Based Loan to Fund Work Offshore Gabon

Oil Tank Platform Collapses, Oil Spills in Terrebonne Bay, Louisiana

Saipem in Talks to End Its Arctic LNG 2 Projects in Russia

Keppel Terminates Semi-sub Rig, Jack-up Rig, and Liftboat Orders as Clients Fail to Pay Up

Subscribe for OE Digital E‑News