Magnetic Particle Inspection Underwater
By Ray Whitfield, ADCI-certified commercial dive supervisor. Reviewed by Amanda McCallister, editor.
Underwater Inspection | Commercial Diving Contractors Alabama I remember my first underwater MPI job like it was yesterday. A sluice gate on the upstream face of a concrete dam, 70 feet down, where the intake had been leaking past the leaf for a decade. My glove was screaming at the cold, and the yoke felt like a small anchor. The gate had cracks at the toe of a fillet weld, and the client wanted them mapped and measured before the next outage. Magnetic particle inspection is the bread-and-butter method for finding surface cracks in ferromagnetic steel. You put a magnetic field across the area, spray iron particles in a liquid suspension, and watch where they collect. The particles are attracted to the flux leakage at a crack, and they line up like your favorite bar magnet does in a pile of filings. Underwater, it's the same physics, but the conditions are meaner: clear water is rare, currents are real, and your own breath bubbles can disturb the particle pattern. This is the way we actually do it on inland and industrial dive sites, not the brochure version.
What Underwater MPI Can Actually See
MPI only works on ferromagnetic materials — iron, nickel, cobalt, and their alloys. If the steel doesn't stick to a magnet, you're wasting your time. Call it the magnet test. It finds surface-breaking cracks, and only those. A crack has to open to the surface, or at least be close enough to distort the field. If you're trying to see a flaw six millimeters down, forget it. We mostly chase fatigue cracks in weld toes, like the one on that sluice gate. Or stress corrosion cracks along a heat-affected zone. Those are the ones that will bite you if you ignore them.
The Gear and Rigging
A basic underwater MPI kit is close to what a topside tech carries, only everything is waterproof and typically battery-run. We use a DC electromagnetic yoke — Magnaflux and Parker Research are the common ones — with a 12-volt or 24-volt drop, not AC. AC can set up eddy currents, and in the wrong steel it can leave residual magnetism that makes the next test impossible. The yoke is a horseshoe magnet on a stick. You hold it flush against the steel and pulse it with a switch on the handle. Then you need a UV-A light source. For fluorescent testing, the particles are coated with a dye that glows yellow-green under black light. The light sits in a waterproof housing with a battery pack, and you use it with a filter that cuts white light. Non-fluorescent black particle is still used in low-visibility water, but the fluorescent is far easier to read on a bright day. Also a spray bottle or a pressure pot for the particle suspension. The particles come as a powder you mix with water, usually one to two percent by weight. And a short stainless steel brush to prep the spot. Your dive helmet has a communication rig, a standby diver is dressed, and the topside check-in keeps a float line with a lift bag to haul the gear if you have to bail. The whole kit, give or take a scrubbing brush, fits in a tackle box. That's the part the brochure doesn't show.
How We Run a Shot
First, we drop to the work site and set up a safety line from a stanchion or a ladder. We can't see much in dirty water, so a lot of the work is by sense of touch. The diver gives topside a thumbs-up, then we start at the bottom of the scan if there's a known crack, because the last thing you want is to lose your place on a fresh field. Next is cleaning. You use a wire brush or a rotary grinder on a hydraulic tool in bad cases. The surface has to be clean and dry. Dry is a funny word underwater, but you can wipe the steel with a rag and a little solvent, and the water will bead off if you let it sit a second. If it's steel that's been underwater for years, it's slime-covered. Scrape that off first. Then you hold the yoke firmly against the steel, poles touching square. Hit the switch for a few seconds to let the field build. Spray the particle solution across the area while the field is on. The particles start moving toward the crack, and you watch through your mask or helmet. After about ten seconds, the pattern sets. A sharp line of particles means crack. A fuzzy line might be a lap or a grind mark. After each test, turn the yoke off and wipe the area. Take a photo with an underwater camera, both UV and white light, so the report has teeth. A GoPro can do it, but you need a filter or all you get is blue. We usually demag afterwards if the part is going to be welded or machined. An AC coil can fade a residual field, or you drag a permanent magnet off the part.
Reading Indications and Reporting
The hard part is not falling for false positives. Barnacles and zebra mussels make their own little bumps, and particle mixture can pool in a groove and look like a crack. A real crack is sharp, continuous, with particle buildup at the tips. The contrast is worse in water, your eyes are dealing with refracted light, and you've been breathing hard from the swim. You have to respect the field orientation. A crack perpendicular to the field will show up; one parallel to it won't, because there's no leakage. So you run two tests at 90 degrees, but a practical guy already knows the weld toe is usually the high-stress line. Point the yoke across the weld and you'll catch the ones that want to run along it. One time we were checking an outlet pipe in a paper mill in Maine. Visibility was near zero, and the plant couldn't shut down. The only way to do MPI was by feel. We'd spray, then run gloved fingers across the steel. If a cluster of particles was dense enough, you'd feel a faint hump. I swear some of the best NDT divers are blind men with a wet finger. That's not in any manual, but it gets you the answer. The report says what you saw: crack number, location from a reference point, length and depth if you can gauge it, and a photo. You don't put 'probably a crack' in a report. You either saw it or you didn't. If you're not sure, clean the spot again and do another shot. That's why they pay us the big money, or what's left of it. Depend on the plant, but that's the way the work goes.
Frequently Asked Questions
Can underwater MPI find cracks in steel piles and jackets? Yes. It's used a lot on offshore and freshwater structures, mostly for welds and parent steel that shows fatigue cracking. If it's ferromagnetic and has a crack open to the water, you can usually find it with a yoke and particles. Depth doesn't change the physics, but the diver's ability to hold a yoke on a vertical surface does. What's the difference between UV fluorescent and black particle underwater? Fluorescent is easier to see in clear water because the particles glow against the dark background. Black particle works in low-visibility water when you run your fingers over the surface, but it's less reliable in current. Most dive teams carry both and pick based on turbidity. How long does an underwater magnetic particle inspection take? A typical inspection on a straight weld can be done in 20 to 30 minutes per linear foot, give or take, once the diver is on the part. But if there's marine growth, high current, or a large area to cover, call it double that. Depends on the plant and the access. Is underwater MPI safe to do around a live plant? It can be done, but it takes careful planning. We use low-voltage DC yokes, and the dive team stays physically clear of any crossing currents. The intake flow is the biggest risk, not the electricity. If the plant has to keep running, the diver works in a safe zone and keeps a tender on the line.