7 Jobs an ROV Still Cannot Do

By Ray Whitfield, ADCI-certified commercial dive supervisor. Reviewed by Dana Whitcombe, technical reviewer.

Underwater Inspection | Cathodic Protection & Anodes October 12, 2011. Bonneville Dam, Oregon. The forebay was running at 40,000 cfs and the water was the color of weak coffee. I was standing on the upstream gate with a VideoRay Pro 4 at my boots. The contractor had promised the client an ROV could 'inspect the trash racks' without a diver hitting the water. It could. It brought back beautiful video of algae and one whitetail bass. But when I asked it to find the missing ¾-inch bolt on the downstream face of gate 2, it couldn't feel a thing. The camera glided past it at four inches. I could see it in the video. In person, with a glove, I'd have caught it in two seconds. That's when I started keeping a list of the jobs I still send a human down for. ROVs have improved a lot since then. They've got thrusters that hold station in 6-knot currents, HD cameras, sonar, even little manipulator arms. But there are seven jobs on my list that still eat an ROV for lunch. If you're planning an underwater job, here's what to know.

The feel of a missing bolt

Job 1 on that list is what I call the zero-vis feel. In a river like the Columbia, you often get pockets of silt where the visibility drops to less than an inch. An ROV's camera becomes a grey screen. Its sonar doesn't have the resolution to see a missing fastener. A diver's glove can find it by touch. They'll run a hand down a bolt pattern and know exactly which hole is empty, and more importantly, they'll feel the corrosion around it that tells you the thread condition. An ROV can't do that. It can only show you what a light reaches.

Torque is a feeling

Job 2 is torque. On a penstock flange or a chemical line, the specs call for a specific torque, say 140 ft-lb with the right lubricant. An ROV manipulator isn't stiff enough to get a calibrated torque wrench in there, and it has no sense of feel. If a thread isn't biting, an ROV will just spin it. A diver can feel that, back it out, clean the thread, and try again. There are subsea torque tools, but they're for oilfield work, not for the cramped corners you find inside a power plant intake.

Cleaning a trash rack

Job 3 is cleaning trash racks. These are the steel bars that protect a turbine from logs and fish. They're usually spaced 4 to 6 inches apart. An ROV with a water jet gets pushed backward by the jet's own reaction, and it can't get its thrusters between the bars without stirring up sediment that makes vision useless. A diver with a pressure wand and a scraper can work along each bar in a methodical pattern, and they can actually see what they're cleaning because they're right there. In a fish bypass, the standard matters; a clogged rack can cost you a turbine.

Reading a crack with a hammer

Job 4 is crack assessment. I've been called to a dam where the concrete on the upstream face had a hairline crack. The ROV video made it look benign. But a diver with a chipping hammer is the only way to tell if it's living. They tap along the crack and listen. A clean, sharp ping means the concrete is sound. A dull thud means you've got delamination underneath. That's feel and hearing and experience combined. An ROV can't do that. It can't even hold the hammer steady against the face in a current.

The wet weld

Job 5 is welding. Despite millions spent on submarine welding robots, the standard for repair below the waterline is still a certified welder-diver in a hot water rig. An ROV can inspect the joint and carry a stinger, but it can't lay a bead that's going to fly a video. There are robotic welding systems in dry habitats for pipelines, but in the messy world of dams and intakes, I'll take a human with a stick electrode every time. The weld has to hold against water hammer and vibration; a machine can't adapt to the changes you see in weld puddle on the fly.

The tight squeeze and the hard call

Jobs 6 and 7 are less mechanical and more about being human. Job 6 is confined space entry inside a pipe. An 18-inch penstock to a Pelton turbine is too small for an ROV with any useful sensor payload. A diver can squeeze in with a brush and a light, and they're small enough to turn around. Job 7 is the abort call. I've been on jobs where the current spikes, or the silt drops, or a line gets fouled. An ROV cannot decide to stop. It'll keep flying and get itself tangled. A diver can make the call to surface, call me on the comms, and say, 'Ray, this is too dicey.' That judgment is the one thing nobody's put on a circuit board yet.

Frequently Asked Questions

If ROVs are so limited, why are they used at all? Because they're cheap and safe for quick inspections in clear water. I've used them to do pre-dive surveys and to document as-built conditions. They save money and keep divers out of the water when the risk isn't worth a paycheck. Should I rent an ROV or hire a dive team for my intake inspection? It depends on the questions you're asking. If you just need to see what's down there in clear water and you don't need to touch anything, an ROV might be enough. If you need to know whether a bolt is loose or a crack is active, you need a diver. Talk to your dive supervisor and show them the drawings. Do ROVs get better every year? They do, but there's a physics floor. Manipulators don't give force feedback through a tether, and the visual resolution in murky water goes down fast. Until they build a robot that can feel the difference between a stuck bolt and a stripped one, there will be jobs for divers.