We Were Asked to Weld a Crack That Should Not Have Existed

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

Underwater Welding & Cutting | Cathodic Protection & Anodes The call came in March 12th from a town that doesn't show up on most maps. The water superintendent did the talking: they had a crack in the raw water intake line, 28 feet down, and the town board was tired of paying for divers. They wanted someone to go down there and weld it so they could get on with their summer. I asked what the line was made of. He said steel. I asked how old. He said the plant was built in 2008. So a ten-year-old line, cracking already. That should not have happened. I told him I'd take a look. My dive partner and I drove up with the compressor and a welding rig. The reservoir was a muddy thing in spring runoff. Visibility was about four inches.

The Crack

We set up on the end of the plant's dock. The intake pipe was a 36-inch carbon steel line that came from a crib intake about 60 feet offshore. The crack was on the top of the pipe, near where it went through the concrete wall. That told me something right away: if it were just corrosion, you'd expect pits, not a clean through-wall crack. This crack ran parallel to the pipe axis, about 14 inches long, with a slight opening at the center. I got into the water with a wire brush and a needle gun. The mud made everything look the same. I had to feel for the crack with my fingers before I could see it. We slung a small light, and even then it was a battle. The first thing that struck me was the crack edges. There was no rust scale inside, no salt. The faces were shiny and crystalline. That's a fracture, not corrosion.

Why It Should Not Exist

A carbon steel pipe that's properly designed and installed does not crack after ten years unless something is forcing it. The water isn't that corrosive. The pressure is low. The line is half buried. So I started looking for the forcing function. I surfaced and asked the superintendent for the design drawings. He had them in a cabinet. We saw the pipe was supported every 12 feet per the spec. That's typical for this size. But when we dove along the outside of the pipe, we found that three of the supports had been taken out. They had removed them to run a new electrical cable, and never put them back. The pipe was free to vibrate. That's the kind of thing that happens when a town doesn't have a dedicated engineer. A contractor does a side job, saves a buck, and leaves a problem. The crack was a classic vibration fatigue failure. It starts as a small hairline, grows at the corners, then becomes a through-wall crack that weeps.

The Fix

Welding a crack like that without addressing the missing supports is a waste of time. We told the superintendent that. He wasn't happy. He said they'd budgeted for the weld, not for a pipe job. We gave him two options. Option one was to weld the crack and install temporary supports in the dry, by lowering the water level. That would take a week of planning, a cofferdam, and a crew that can do a certified weld. We don't do that kind of work. Option two was to do it in the wet. We'll grind the crack out, preheat with a torch, and use a butter pass and a fill pass. In that visibility, that's a risk. You can get a decent weld, but you can't inspect it properly. The client has to accept that. He asked how long it would last. I said 'forever, if you fix the supports.' He asked what if they didn't fix the supports. I said 'two or three years, maybe.' We ended up doing the wet weld because they couldn't shut the plant down. We used a steel patch plate, not just a bead. We ground out the crack to a proper groove, welded a 1/2-inch plate over it, with seal welds around the edges. That's a standard repair for a non-pressure line.

Underwater Welding, Explained

Underwater welding is a trade of its own. We used the same rods as you'd use for a dry weld, but they're covered in a wax or plastic coating to keep them dry. The heat is the enemy. Water sucks the heat out of the weld instantaneously, so you have to work fast and keep the arc short. You also have to be careful about hydrogen. The water molecules break down and can get trapped in the steel, causing cracks. That's why we preheat when we can, and why we always do a grinder pass after to remove any cracked weld. In this case, we were lucky the crack was on top of the pipe. That meant the gas from welding could vent upward, and we could see a little bit. If it had been on the bottom, we'd have had a harder time. The patch plate was the right choice. A single weld bead on a through-crack is a band-aid. A plate spreads the load and gives the area strength. We welded it in two passes, overlapping each other, then painted a seal coat on the edges. We also put a sacrificial anode on each end to help with corrosion. We left the site the next morning. The superintendent sent us a picture of the patch a week later. It was holding.

What I Learned

The weld itself took about six hours over two dives. The supports are still missing, as far as I know. Last I heard, they've scheduled the work for next summer. The thing that sticks with me is the town board's attitude. They'd called the local stainless steel shop and asked if they could weld it from the inside. The shop said no. Then they called a welder from a nearby city, who quoted them a price that started at five figures before he'd even looked at the water. We came in at a fraction of that, but we didn't get the support fix done. There's a lesson here for anyone who owns a piece of infrastructure: a crack is a symptom, not the disease. If you only treat the symptom, you'll be back in the water sooner than you think. The best dive job is the one that doesn't have to be repeated.

Frequently Asked Questions

Can you really weld underwater? Doesn't the water just put out the arc? Yes, it's real. The arc creates a gas bubble around the electrode that keeps the water away for a fraction of a second. It's not pretty, and it takes a certified diver-welder. The weld is never as strong as a dry weld, so we use patches and account for that. How do you know the welding rod is going to work underwater? We use special electrodes rated for underwater welding. They have a waterproof covering. The rod burns the same way, but the water cools it faster, so we move in short, quick passes. Why did you say it should have lasted forever if the supports were fixed? Because the pipe itself was sound. Vibration fatigue was the only reason the crack appeared. With supports back in place, the pipe would have no reason to flex, so the patch would hold. If the pipe had been corroded or the metal had degraded, I'd have said something different.