Deep Dive into Engineering the World’s Most Advanced ROV System: Q&A with Carl Barrett
In August 2017 a research group led by explorer and philanthropist Paul G. Allen used ultra-high-tech underwater equipment to locate the wreckage of the USS Indianapolis, a ship that sunk in the final days of WWII after it was struck by Japanese torpedoes. The discovery was made by Mr. Allen’s company, Vulcan Inc., using a new expedition ship it acquired for the purpose of seabed discovery—the RV Petrel.
Petrel was outfitted with cutting-edge technologies, including an autonomous underwater vehicle (AUV), which uses side-scan sonar to locate objects on the seabed, and a remotely operated vehicle (ROV) for further investigation and video documentation.
While AUVs and ROVs are becoming more common, the USS Indianapolis was discovered at a depth of nearly 6,000 m, and technologies suitable for robust research at great depth can be hard to find.
Carl Barrett is a project manager at the underwater engineering firm 3U Technologies LLC (Underground, Underwater and Under Ice) who worked with Vulcan to outfit an ROV that could handle extremely deep depths. Sea Technology sat down with Barrett to talk about engineering an ROV system that helped make RV Petrel one of the most advanced deep-sea expedition vessels on the planet.
Barrett told us about the process of specifying and sourcing equipment for the ROV that recorded the broadcast-quality video we saw on the PBS live-streamed tour of the USS Indianapolis ruins. He provided plenty of insight into the complexities of deep-sea expedition, and some neat ROV graphics as well.
Tour the RV Petrel ROV in HD! (See the video after interview.)
ST: So Carl, tell me how you first got involved with the RV Petrel and Mr. Allen’s team?
CB: 3U Technologies started working for Vulcan in 2013. We assisted with developing upgrades for their existing ROV, the Octo ROV on Mr. Allen’s yacht Octopus. The system was outfitted with a larger umbilical for higher power capacity. It’s a hydroelectric machine, so upgrades included a larger HPU and larger thrusters. We did pretty much a full rebuild of the power system. Due to the success of this and other projects, Vulcan tasked 3U with performing a study to develop options for design and build of a 6,000-meter ROV system.
ST: So Vulcan was originally looking at putting this new ROV on Mr. Allen’s yacht, the Octopus?
CB: Yes, the initial thinking was to perform exploration from the Octopus, the vehicle system was built for that purpose. However, a decision was made to buy the Petrel when it became available, and plans shifted over to installing the vehicle on the Petrel, which was a very good move. The Petrel was an ROV support vessel purchased from Subsea 7.
CB: The installation consisted of taking the [existing] vehicle system off [Petrel], which was a much bigger vehicle, although it was only a 3,000-meter system. We repurposed the winch and handling system and made a few modifications, specifically a different spooling shell and some inserts for the over-boarding sheave to fit the smaller umbilical. Otherwise, it was a really good fit, and there was certainly a lot of room on the Petrel. There’s a nice control room on the Petrel, [the team is] able to spread out with a lot of consoles and a lot of operating space.
ST: Did the Petrel and its ROV undergo field trials before it was put to the task of deep-sea discoveries?
CB: Absolutely. The Petrel and diving systems completed final outfitting in February of this year. We ran sea trial operations, called shallow-water trials, in the fjords of Norway in February and March. In the fjords, we dove to perhaps 700 meters, just testing all the systems on board, diving the AUV, finding targets. As you can imagine, the fjords of Norway are littered with a lot of boats on the bottom, so there were a lot of targets to test the system on. Once everything was operational the team went to the Mediterranean for their first set of deep-diving trials.
ST: What kind of testing is required for a new umbilical, the cable that connects the ROV to the ship?
CB: The thing to do there is to dive just the umbilical, without the vehicle. You blank-off the end of the cable with just an oil-filled junction box, a container, and find a place in the ocean deep enough to spool all the cable off the winch, let it spin out and find its happy home, and then spool it back onto the winch. You do that several times to season the cable and get the cable rotationally neutral. As you can imagine, you want to try for balanced torque on the cable so that as you tension it up it won’t want to twist. You want to get all that worked out so that it’s a rigid armored umbilical, stored on the winch under tension and without twists built in. Without that it would tend to hockle, or generate loops in the cable.
After cable seasoning was complete, the ROV was connected, and they did a series of deep diving trials which effectively resulted in finding a WWII Italian vessel, Artigliere. Then, in April the Petrel started steaming to the Pacific, going through the Suez Canal and out to the Philippine sea, with operations commencing around May. The months of May, June, July and August were spent running AUV operations, checking targets with the ROV, and searching for the USS Indianapolis. And then of course the Vulcan team found it in late August.
ST: In the beginning, what was Vulcan’s main request to you? I assume they said: “We need the ROV to do____.”
CB: Everything! Their biggest need is a good camera and video platform. They plan to use the system primarily for investigating wreck sites. It needed a lot of lighting, it needed a lot of data bandwidth and data communications capabilities to run multiple sensors, multibeam sonar systems, 4K HDTV video cameras, high megapixel stills, and to do all of this at once.
Six-thousand meters is a pretty unique depth. There are a whole range of sonar and camera systems for 3,000 meters. We had to find products that were packaged for 6,000 meters. If anything, the depth drove a lot of the decisions. Once we had the basics there, we had a whole stack of equipment that needed to transmit data to the surface, so the next step was coming up with a MUX [multiplexer] solution to transmit all of it.
ST: I keep reading this is the most advanced ROV system in the world. Is the 6,000 meter depth the reason for that?
CB: Right. This would be the most powerful and advanced 6,000 meter system. There are certainly plenty of systems in the world with this level of capability at 3-4,000 meters, but there are very few systems rated for 6,000 meter operations, most of which are older designs with more limited capabilities. We investigated all the major ROV manufacturers initially, but found that most were limited by the operating voltage of their existing designs, so the power level at the ROV would be limited. Jumping from 3,000 to 4,500 volts would be a custom change for a lot of these manufacturers.
The requirements for the vehicle and power system itself were number one on the priority list: we wanted to push as much power down to the machine as we could to maximize propulsion on the vehicle. Vulcan wanted a responsive system. It needed to be able to drag around 6,000 meters of cable (the system does not use a TMS) so the umbilical needed to be small to minimize drag and support its own weight.
ST: And you said the umbilical for this project was highly specific?
CB: That was certainly a big portion of the project, sourcing the umbilical and testing it. It’s a new 17mm design. I won’t say it’s never been done before, but the combination of everything in the umbilical [was unique] because we were looking for a 4,500 volt capability in order to get power down to the [6,000 m] depth. A lot of systems still in-use operate at 3,000 volts, and there are 3,000-volt cables, but that was a significant limitation on how much power we could transmit. Cortland cable offered a solution designed for our planned 4,500-volt system and was able to offer the largest conductor size for maximum power transmission.
Another key was the ROV manufacturer, Argus Remote Systems, one of their big advantages is that they were offering an electric vehicle system, as opposed to an electro-hydraulic machine. The big advantage there is that we only need three power conductors in the umbilical to feed power distribution on the ROV as opposed to a typical electrohydraulic machine which requires five conductors: three to run the HPU and then two to run instrumentation separately. By simplifying the cable, we could maximize the size of the conductors powering everything. Additionally, details of the Argus power system allowed for a larger voltage drop (voltage regulation) range in the system, which further increased the power throughput capability. As a result, we are able to supply up to 90 kW of electric power to the ROV, 6,000 m below Petrel.
All that power can be distributed as needed on the vehicle, and it gave us the ability to run a lot of lighting while maximizing propulsion capability. There are five kilowatts of lighting capacity on the vehicle distributed through eight switchable and dimmable lighting circuits, which typically run one or two lights, so it can be outfitted with up to 16 different LED lights, that makes for a lot of lighting down there on the machine, resulting in spectacular video.
I don’t think you’ve heard the last of Vulcan and new discoveries from them. This whole system was built for exploration, and this was just the first set of operations…
ST: How does data move from the ROV 6,000 m underwater up to the ship?
CB: As we started into design of the system, we settled on a Focal MUX [multiplexer] for all the data transmission via fiber optic elements in the umbilical. It is a rather extensive MUX with a lot of high-speed serial data ports, 4x gigabit Ethernet channels and multiple SD and HD video transmission channels. All of the various sensors were integrated directly into the MUX as opposed to, say, a control system subsea on the vehicle.
ST: How does all this equipment work together in one vehicle?
CB: One of the key features on the system is the Greensea software and hardware. The ROV is outfitted with a full inertial navigation system (INS) including DVL (doppler velocity log) and USBL, which is an ultra-short baseline tracking system from Kongsberg. There is also a heading reference system included, a ring laser gyro with accelerometers. All these sensor systems feed into an INS processor supplied by Greensea, which provides all the system automation.
Greensea also provided what we call sensor fusion. Initially when we started looking at integrating all of this into a system we realized we had an issue. Every sensor—whether a sonar, a pan-and-tilt connected through serial interfaces, DVL—all these pieces of equipment, each one of them has its own vendor-supplied software package to operate it. That means we would have needed something like 20 different computers operating top-side to gather up all this data and present it. It was looking unwieldy initially, until we started talking with Greensea. What they provided was a centralized, network-based solution, where they effectively emulated the vendor-supplied software on a common platform. Greensea supplied drivers for each of these pieces of hardware, although they already had a lot of these drivers in their existing Opensea inventory. That way, all of this could run on one processing system and be integrated through a surface network, combining all of this data into one platform.
As I described it to them: it’s a system for our system! What that allowed is that now we’ve got all of this information feeding into Greensea’s package, where all the data, including SD Video, is stored on a running black box hard drive.
ST: How many cameras are on the ROV and how do they work?
CB: There are a total of up to nine video cameras and capability for multiple stills on this system. Six of the cameras are just standard definition cameras used for maneuvering—a camera looking out back, one looking up at the umbilical, various other views so the pilots can look all around the ROV for maneuvering, but the SD video isn’t necessarily presented in the [PBS] show or other presentations. All of the SD video, as well as the HDTV converted to SD, is saved on a black box. The ROV features three HDTV cameras (both SDI and 3G) for presentation and broadcast purposes. The broadcast-quality video flows to other Vulcan server systems on Petrel which are designed to handle and store high-bandwidth HDTV. Beyond 3G capability, the ROV data circuits can support 4K video in the future, but Vulcan has yet to find a camera system that meets their needs on this system.
All the video and data information can be displayed where needed. Inside the control room, we’ve got a multitude of HD displays, and the pilots have the ability of switching any of the sensor outputs to any of the displays. They can route scanning sonar, multibeam sonar, SD cameras, or HD video, etc. to any of those various screens. And they have a lot of video feeds coming in beyond just the vehicle. There are multiple views of the winch and handling system topside, all of which the pilots can monitor. During diving operations, the pilots can monitor the angle of the umbilical hanging from the handling system, and they can drive the winch from the pilot’s chair directly. In fact we even tied in an automated winch control such that the winch will pay in/out automatically as the ROV depth changes, which helps to maintain a consistent level of slack at the ROV to prevent entanglement and maximize operational efficiency.
ST: How do the pilots on-board RV Petrel drive the vehicle at the end of a 6,000 meter cable below?
CB: The ROV can be flown in a sort of standard (old school) operation mode using a joystick with the pilots directly controlling propulsion on the machine. But it can also operate in a semi-autonomous mode, where the ROV will hold station, and the pilots operate it by movement of the joystick to move the vehicle relative to an existing position. For instance, they can command the ROV to move five meters one way or another, or move to various waypoints. The beauty of running in this automated mode is that the computer uses the INS system for positioning reference, which operates on a very fast update rate, so ROV propulsion and positioning control is much more stable than any pilot could ever achieve.
One big feature that Greensea offers is sonar target-based tracking. Using the multibeam sonar, you can select a sonar target of interest, and the ROV will hold station and navigate relative to that target. This is a big advantage for what Vulcan does: a lot of their operations are in the vicinity of large sonar targets, large steel wreck sites, so as they mapped out the wreck site, they would have selected a large series of sonar waypoints, and they could navigate the vehicle around the site in that manner. All this automation makes the vehicle very stable as a camera platform and it helps decrease a lot of the task loading on the pilots. So rather than having to manually hold station, they can focus more on the overall mission, improving operational efficiency.
We also customized the cyber chairs [in the console] with our controls, there’s a joystick on the right-hand side and then there’s a paddle on the left hand side, which is used for vertical control. The chairs also feature all of the required switches and dimming controls for lights, camera controls, pan-and-tilt control etc., all from the comfort of the chair itself. Each pilot then has their own small touch screen which allows them to directly access all aspects of the ROV control system and data that is coming up from the vehicle.
ST: Do you think this particular system will be involved in more deep finds like the Indianapolis?
CB: Absolutely. I don’t think you’ve heard the last of Vulcan and new discoveries from them. This whole system was built for exploration, and this was just the first set of operations, this is the first year they’ve had this capability in operation. I can’t predict what else they have planned, but I expect more exciting news in the future!
Interview by Amelia Jaycen
Carl Barrett has worked in the ROV and deep-diving industry since 1986. He started with Oceaneering as a divers’ tender and then worked his way up through ROV operations, ultimately joining the engineering group, Ocean Systems Engineering (OSE), in Houston, Texas. At OSE he developed tooling such as torque tools, docking buckets, tooling skids and other ROV equipment, as well as supporting new and unique offshore projects. He is an author on the first version of the API 17D standards for ROV tooling, the basis for the current ISO 13628 ROV tooling standards. He has worked at Perry Tritech, where he led multiple cable burial and oil field ROV system developments and builds. He has been a project manager at 3U since 2003.
Interview has been edited for style and clarity.