How Does Underwater Welding Work? Wet vs. Dry Explained

A tanker is taking on water in the North Sea, the tow line snaps, and the repair barge cannot get alongside in a rising swell. So the diver-welder goes down anyway, with a waterproof electrode in one rubber-gloved hand and a surface tender holding the only switch that can kill the arc. Thirty metres down, in the dark, with a weld pool the size of a dinner plate and no way to see what he is making. It looks impossible. It is also one of the oldest tricks in heavy fabrication, and the physics behind it are genuinely fascinating.

How Underwater Welding Actually Works: The Short Answer

Here is the plain answer, before any nuance. An electric arc is struck between a flux-coated electrode and the workpiece. The heat of that arc vaporizes the surrounding water into a cavity of steam and gas, and the cavity wraps itself around the arc and the molten pool. That bubble does the job that shielding gas does in air welding. The electrode melts, drops a puddle of liquid metal, the puddle fuses to the parent metal, and as the diver drags the electrode along the joint, the puddle solidifies behind it into a bead of weld.

That is the whole trick. It is simple physics and unforgiving metallurgy, in that order.

The sequence, step by step

1. Preparation. The joint is ground or flame-cut clean, fitted, and clamped. Marine growth, thick paint and heavy rust have to come off first, because the arc has no tolerance for contamination.
2. Arc initiation. The diver makes contact with the workpiece and the power source energizes. The electrode’s flux coating decomposes, and the arc strikes.
3. Vapour bubble formation. Arc heat boils the surrounding water away, creating a cavity of steam and gas that encloses the arc, the electrode tip and the weld pool. As the bubble grows, the arc settles into it and stabilises.
4. Electrode melting. The electrode melts at the tip, and the arc transfers heat to the workpiece, melting a small volume of the parent metal.
5. Puddle formation and fusion. Molten electrode metal mixes with the melted parent metal to form the puddle, which fuses to the surrounding base metal.
6. Solidification. As the diver moves along the joint, the puddle is drawn out behind the arc and solidifies into the weld bead, cooling almost instantly in the surrounding water.

Nothing about the circuit is new

Here is the part that surprises people. The arc circuit itself is essentially identical to topside shielded metal arc welding. Same electrode, same flux chemistry, same constant-voltage behaviour, same need for a stable arc length. If you already know why SMAW still wins on topside jobs, you already understand90% of the mechanics of underwater welding.

What changes is the environment, and the environment is hostile in four specific ways:

• Cooling. Water pulls heat out of the weld far faster than air, so the metal quenches almost on contact.
• Hydrogen. Decomposing moisture produces hydrogen, and hydrogen is the enemy of weld metal toughness.
• Visibility. A murky North Sea gives the diver very little to see, and rising bubbles actively obscure the bead.
• Electrical safety. The diver is surrounded by a conductive medium, wearing conductive equipment, touching a live electrode.

Two families of method

Every underwater repair falls into one of two camps. Wet welding means the arc burns in open water with no barrier between the water and the weld. It is fast, cheap, portable and needs no fixtures at all, which is why it remains the default for emergency work. Dry welding means the welder works inside a sealed enclosure: a chamber, a habitat, a one-atmosphere pressure vessel, or a small transparent dry-spot hood held over the joint. Dry methods are slower and considerably more expensive, but they reproduce above-water weld quality.

The rulebook: AWS D3.6M and OSHA Subpart T

Underwater welding is not an improvised trade. It is a codified discipline governed by two documents. The AWS D3.6M standard for underwater welding defines five basic methods of underwater welding and three weld classes: Class A, for welds comparable to above-water welding; Class B, for less critical applications; and Class O, for welds that must meet some other designated code or specification. AWS first published its Underwater Welding Code in 1983, with later editions in 1989, 1993, 1999 and 2017, each one tightening qualification, procedure and inspection requirements as the practice matured.

Safety sits under OSHA 29 CFR 1910 Subpart T, Commercial Diving Operations, which treats this as hot work performed by commercial divers under a diving plan, with its own rules on current switching, grounding, insulation and gas-freeing.

What has changed over those decades is the standing of the work itself. Underwater welding began as a stopgap for damaged steel-hulled vessels, and it is now accepted practice for engineered repair of offshore structures, submerged pipelines, port facilities and nuclear power plant components.

Over the next sections we will walk through the five environments, the physics of the vapour bubble, why hydrogen causes cracking, the safety framework in detail, the craft itself, and where this work actually gets done. If you remember one thing from this introduction, make it this: the arc is not the hard part. The water around it is.

The Five Underwater Welding Environments: Wet, Dry and Everything Between

“Underwater welding” sounds like one process with one set of rules. It is not. The AWS D3.6M:2017 standard for underwater welding recognises five basic methods in use today, arranged on a spectrum from fully dry to fully wet. Where you land on that spectrum carries real cost, so each environment is worth understanding before you commit.

The Five Methods, From Driest to Wettest

1. Dry Welding at One Atmosphere

The closest thing to a topside shop job. A sealed chamber is lowered to the site, water pumped out, and the welder works inside at roughly one atmosphere. No diving suit, no helmet, no umbilical. Because internal pressure matches the surface, there is no physiological load on the welder at all. It is the most comfortable and highest-quality option, and the most elaborate and expensive, since you need a pressurised vessel placed and secured over the joint. Standard choice for critical work on nuclear components, pressure-retaining structures and high-criticality joints.

2. Dry Welding in a Habitat

A habitat is a working chamber, usually a bell or cylinder, flooded to ambient pressure, then sealed and dewatered inside. The welder enters dry and works at surrounding water pressure, breathing from an air supply. Free of a bulky suit, productivity and weld quality are considerably better than wet, and the work is far less taxing. Habitats are serious diving infrastructure: gas supply, life support, communications, handling procedures. They suit structural repairs where quality outweighs mobilisation cost, and can host several workers for extended shifts.

3. Dry Chamber Welding

The simplest dry option. An open-bottomed enclosure, often a rigid box or bell, is lowered over the joint so the diver’s head and shoulders sit in a dry gas-filled space while the body stays in the water. The diver enters from below in full gear and works with breathable gas inside. Still a diver and still suited, but head and arms are dry, which is where it matters. Visibility is vastly better than wet welding, the electrode holder need not be waterproof, and the arc is shielded by gas rather than a vapour bubble. Cost and complexity land between a habitat and a wet weld.

4. Dry-Spot Welding

Dry-spot welding inverts the chamber idea: it encloses the weld rather than the welder. A small transparent enclosure, often acrylic, is fixed over the joint and filled with inert or breathable gas. The welder stays outside in a suit and reaches in through gloves or a sealing sleeve to work in a dry, gas-shielded atmosphere. Cost and complexity stay down while the weld metal gets a proper gas shield and the welder a clear view of the puddle. The limitation is obvious: accessible joint area is small, and the reach is awkward on complex geometry.

5. Wet Welding

The baseline method. The diver works at ambient pressure with no mechanical barrier between the water and the arc. The electrode sits in a waterproof holder, the joint is submerged, and the arc is struck through water. The vapour bubble around the arc does the shielding. By far the most versatile: no chamber, no habitat, no fixtures, and a team can be welding within hours of mobilising. Often the only realistic option, because there is rarely a way to place a chamber over an awkward hull penetration or a pipeline tie-in in open water.

Why the Environment Changes the Weld

The environment is not a cosmetic detail. It changes how the arc behaves, how metal transfers, how the puddle solidifies and how the finished weld performs. AWS D3.6M notes that metal transfer characteristics, solidification behaviour, weld appearance and mechanical properties can all vary with pressure, so any of these methods may behave differently from conventional surface welding. Compare that to why SMAW still wins in so many shops: a topside stick weld has a stable, repeatable arc because the environment is dry, static and predictable. Submerged, the same process fights water flow that quenches the puddle, fragments the shielding bubble and introduces hydrogen into the arc.

As the TWI explainer on arc welding underwater puts it, the process itself is largely familiar; the difficulty is controlling the weld. A dry chamber removes the variable entirely, and the welds it produces are generally equal in quality to welds made above water. That is the argument for the dry side of the spectrum.

The Practical Trade-Off

Factor	Wet welding	Dry methods (chamber, habitat, one-atmosphere)
Mobilisation time	Hours	Days to weeks
Cost	Low	High
Equipment needed	Waterproof holder, surface power and tender	Chamber, gas supply, life support, handling gear
Joint access	Any geometry the diver can reach	Limited to what the enclosure fits over
Visibility	Poor, murk and rising bubbles	Good
Weld metal quality	Lower purity, higher hydrogen	Comparable to above-water welds
Typical use	Emergency and general structural repair	Critical, code-governed and pressure-retaining work

Wet welding wins on speed, cost, versatility and logistics: no fixtures, no special fabrication, applicable to almost any shape a diver can reach. The penalty is paid in the weld itself, with more hydrogen pickup, more contamination risk, an unstable arc and a less predictable microstructure. Researchers studying the mechanics of underwater welding and hydrogen cracking behaviour consistently find the underwater environment degrades mechanical properties, particularly toughness, in ways dry methods do not.

Dry welding wins on quality and control. It is slower, more expensive and far less flexible about deployment. When a project specifies a critical weld class, or the structure is a pressure boundary, a nuclear component or a load-carrying member with no tolerance for a weak fusion boundary, the dry side is where the job belongs. Often the strongest argument for dry welding is accountability rather than paper quality: a gas-shielded dry environment is easier to defend than a vapour bubble at 20 metres in a North Sea swell.

Why the Arc Survives: Vapor Bubbles, Steam and Electrolysis

An electric arc is hot enough to melt steel in a heartbeat. The plasma column in a stick arc reaches several thousand degrees Celsius, and it does not care that it is underwater. The problem is the cold water around it, which pulls heat away and flashes to steam so violently that an unprotected arc would be snuffed out in a fraction of a second.

So the arc builds its own shelter: a cavity of vapour. Understanding that cavity is the key to everything else here.

How the vapour cavity forms

The arc does not exist in the water directly. Heat from the arc, the electrode tip and the molten pool flashes the surrounding water into steam, and that steam pushes water back out of the hot zone. The result is a small gas bubble wrapped around the electrode tip, the arc and the weld pool, often only a few centimetres across. It is not a stable structure. It forms, wobbles, collapses and reforms as the arc heats and the water pushes back.

The composition of that bubble is remarkable, and it is not just steam. According to The Welding Institute, the vapour around the arc typically comprises about 70% hydrogen, about 25% carbon dioxide and about 5% carbon monoxide. Hydrogen comes from the electrolysis of water itself; the carbon oxides come from the electrode coating burning back inside the cavity.

That bubble does two jobs at once, and both matter.

• Thermal shield. The gas pocket insulates the arc and molten pool from the surrounding water. Without it, the pool would freeze almost instantly. With it, the steel stays above its melting point long enough to fuse properly.
• Chemical shield. The bubble keeps liquid water and dissolved nitrogen away from the molten metal, so the weld is not contaminated the way you would expect. What does get in is hydrogen, and it comes through the vapour rather than the water.

Hydrogen reaches the pool through that vapour, so diffusible hydrogen can reach up to 100 ml per 100 g of deposited metal when rutile electrodes are used, dropping to around 25 ml per 100 g with oxidising electrodes. That single number, hydrogen pickup, is where the cracking risk in the next section begins.

Bubbles rising to the surface

Gas escaping the cavity does not vanish. It rises, right through the diver’s line of sight. A curtain of bubbles streaming past the work face is the classic underwater welding image, and it is not merely an inconvenience. Rising bubbles disturb the boundary of the vapour cavity, which makes the arc wander, spit and spatter. The diver is welding through a shifting gas pocket that is constantly torn apart and rebuilt.

Electrolysis and polarity

Salt water is a far better conductor than fresh water, and that changes the physics in a useful direction. A more electrolytic environment raises arc heat and improves arc efficiency, which is why industry guidance notes that wet welding works best in higher-salinity water. The North Sea, in that sense, is a friendly welding environment compared with a freshwater harbour. The technical literature agrees, as you can see in this overview of underwater welding metallurgy and arc characteristics.

Wet underwater welding runs on direct current. DC with the electrode negative (DCSP) is the usual choice because it concentrates more heat at the workpiece and behaves predictably underwater. DC reverse polarity (DCRP) can be used, but it demands higher current for the same result. Alternating current is not used, for two separate reasons: it is an electrical safety problem in conductive water, and it is hard to hold a stable arc underwater when the current reverses twice a cycle and the cavity keeps collapsing. Those safety concerns are why OSHA’s commercial diving rules require the welding machine frame to be grounded and the current supply switch to be tended by a dive team member in voice contact with the diver.

One more polarity consequence catches newcomers out. With DC positive polarity, electrolysis is active and causes rapid deterioration of metallic components inside the electrode holder. Contacts, jaws and housing parts corrode quickly, which is why holders are built with insulating shrouds and treated as consumables rather than permanent equipment.

What the diver actually feels

At the work face the arc hisses, wanders and spits, and the pool is a churning, shimmering mess rather than the calm mirror a welder expects topside. The diver cannot hold a fixed arc length like they would in the shop, so the technique changes. Instead of maintaining a gap, the diver keeps the electrode in contact and drags it along the joint, letting the coating burn back slower than the core wire so the coating extends past the tip and sets the arc length automatically. That process is covered in our guide to why SMAW stick welding still wins, and rod selection for awkward positions in our all position welding rod guide.

None of this makes underwater welding mysterious. It makes it a different trade with different rules, and AWS D3.6M treats it that way, requiring qualified procedures and qualified welders for work that would be routine on a dry deck. Everything else follows from one simple fact: the weld is made inside a bubble of hot hydrogen gas, with cold water pressing in on all sides.

What Water Does to the Weld Metal: Quenching, Hydrogen and Cracking

The vapour bubble keeps the arc alive, but the moment the puddle freezes, water has done two things a topside welder never has to fight. It cooled the joint faster than any shop-floor quenchant, and it filled the deposit with hydrogen. Together those effects explain nearly everything that makes underwater weld metal different from the same weld made in air.

The Quench You Cannot Control

In air, a weld bead cools over minutes, giving the heat-affected zone (HAZ) beside the fusion line time to transform into a balanced mix of ferrite, pearlite and, in modern steels, tough acicular ferrite. Underwater there is no such luxury. Water conducts heat away roughly a thousand times faster than still air, so the weld and its HAZ cross the transformation temperatures in seconds.

The result is a harder, finer microstructure: martensite and bainite where you would normally expect softer ferrite-pearlite. On paper that is an upgrade, because harder microstructures translate directly into higher tensile strength. The problem is that martensite and bainite have very little ductility and poor resistance to crack initiation. A weld that is stiff and strong but cannot absorb strain or impact energy is a weld waiting for a bad day. As the TWI notes on wet welding point out, the surrounding water cools the weldment far faster than the filler can properly fuse, and that microstructural penalty cannot be designed out of the procedure.

Hydrogen and the Clock That Never Stops

Hydrogen is the second problem, and the more dangerous one. Moisture feeding into the arc decomposes into hydrogen and oxygen, and the vapour bubble that shields the arc is roughly 70 percent hydrogen. That hydrogen sits in contact with the molten pool, and molten steel absorbs it readily.

Here is the trap. Hydrogen normally escapes a weld by diffusing out of the hot deposit over minutes or hours. Underwater, the quench freezes the deposit long before that diffusion is complete, so the weld retains high levels of diffusible hydrogen and the risk of hydrogen-assisted (cold or delayed) cracking climbs with depth and cooling rate. This is the dominant metallurgical challenge in wet welding of high-strength steels, particularly where the carbon equivalent exceeds roughly 0.4 wt%. In those grades hydrogen collects at grain boundaries, weld inclusions and the fusion boundary, and combined with high residual tensile stress it opens microcracks that may appear hours after the arc has died.

What the Research Actually Shows

Several studies have quantified the effect of depth and water chemistry. A depth study on 5G position wet welds in low carbon steel recorded tensile strengths of 277 MPa at 2.5 m, 303 MPa at 5 m and 313 MPa at 10 m. The trend is counterintuitive but consistent: deeper water means a faster quench, a finer transformation product and more acicular ferrite, so the weld gets stronger as it goes down.

An A36 marine steel investigation showed the same pattern with more detail:

Condition	Tensile strength	Elongation
Air-welded reference	416.16 MPa	4.0 percent
Underwater, cold water, no flow	481.08 MPa	6.137 percent
Underwater, non-uniform flow, no baffle bulkhead	585.09 MPa	6.93 percent

The same study measured absorbed energy of up to 41.9 J and impact strength of 2.05 J/mm², which is where the underwater weld starts to fall behind. Higher strength, yes. Better ductility than the air-welded baseline, marginally. Lower toughness, consistently.

Salinity adds a twist. Research at 0, 7.5 and 35 parts per thousand found HAZ hardness falling from 211 HV10 to 193 HV10 as salinity rose, impact strength climbing from 82.5 to 101.3 J/cm², and fewer microcracks at higher salinity, with diffusible hydrogen reaching up to 65 ml/100 g. Seawater is a slightly better welding environment than fresh water, not because salt removes hydrogen but because it changes how the pool quenches and how the crack-prone microstructure forms.

The most sobering result came from an open-sea trial in which EH40 ship steel was wet welded by SMAW at 7 m depth using austenitic electrodes. Hydrogen-induced microcracks in a lath martensite layer at the fusion boundary became the fracture initiation region in both bending and transverse tensile tests. The crack did not start in the softer, more ductile weld metal. It started at the hardest, most hydrogen-rich place in the joint, which is exactly where nobody is looking with a magnifying glass on a dark, cold deck plate.

Strength Up, Toughness Down

Industry guidance citing naval facility test results shows underwater mild steel fillet welds consistently developing over 80 percent of the tensile strength and 50 percent of the ductility of similar welds made in air. Those are usable numbers for many structural repairs, but they are not interchangeable with an above-water weld, and any code that specifies a critical weld class under AWS D3.6M needs to know the difference.

Post-weld heat treatment, the traditional remedy for hydrogen cracking in pressure-retaining and high-strength components, is essentially impossible underwater. You cannot heat a submerged joint uniformly, hold it at temperature and control the cooling rate, which is why dry chamber and habitat methods exist at all. When a wet weld does need to come out, the practical fallback is mechanical: air arc welding and carbon arc gouging to cut out the suspect metal underwater, clean back to sound parent material, and reweld under better control.

The bottom line for engineers and welders alike: underwater strength numbers can look impressive because they are partly a byproduct of an uncontrolled quench. What you are really trading is toughness, ductility and confidence that the joint will not crack weeks later. That is why quality underwater welds get inspected like welds, and why the industry guidance on underwater weld performance treats these results as a starting point for qualification, never as a blanket guarantee.

Safety First: OSHA Rules for Underwater Welding and Diving

Underwater welding is not something a fabrication shop picks up on a Friday afternoon. It is hot work performed by commercial divers, and in the United States it sits squarely inside OSHA 29 CFR 1910 Subpart T, Commercial Diving Operations. That is the difference between an arc underwater and a controlled job: the diver is also a worker in a hostile environment with an electrical hazard running through the water around them, and the paperwork matters as much as the weld procedure.

1910.422(g): the switch that can kill the arc

The most important line in the rule for anyone following along is 29 CFR 1910.422(g). In plain terms:

• A current supply switch must be provided to interrupt current flow to the welding or burning electrode.
• That switch must be tended by a dive team member in voice communication with the diver.
• It must be kept in the open position except when the diver is actually welding or burning.

Treat it as a two-way safety interlock. On the surface a tender stands at the switch with a hand on it and a radio to the diver. If the diver calls for a cut, if the hose snags, if the arc chases out of the puddle, that tender kills the power in a second. The diver cannot reach the plug, and nobody topside reaches into the water. The switch is the only emergency stop, and the person holding it has to hear the diver clearly.

Grounding, insulation and the gloves on your hands

The same rule requires the welding machine frame to be grounded by solid metal-to-metal contact on the vessel’s structure and/or shore ground connections. That is a grounding point, not a wire clipped to a convenient bracket. On a wet steel hull in salt water, a marginal ground is the same as no ground at all: the whole structure sits in a conductive bath.

Cables, electrode holders and their connections must be rated for the maximum current and properly insulated, and the diver must wear insulated gloves. Personnel operating electric welding equipment used in diving operations must have experience or training in its safe use. That covers the surface crew as much as the diver, and it is why the electrode holder, ground clamp and switch are treated as life-critical equipment rather than consumables. Electrode selection and all-position practice are worth understanding before you ever get wet, so our all-position welding rod guide and notes on why SMAW still wins make useful pre-job reading.

A flooded compartment is not automatically a safe compartment

OSHA’s enforcement directive CPL 02-00-151 carries a warning that catches experienced crews out. Filling a compartment with water does not make it safe for hot work, because two gases can still collect up top: oxygen escaping from the diver’s welding and cutting oxygen supply hose, and hydrogen generated by electrolysis as the arc works in the water. Both accumulate overhead, where the diver cannot see them and where gas escapes poorly.

The directive also gives practical direction on sequencing: vertical cuts and welds are best started high, at shallow depth, and finished low, at greater depth. Starting high puts the hydrogen-rich gas at the far end of the joint, where it rises into space the diver has already left instead of into the diver’s face.

Depth limits and the chamber that has to be standing by

Hot work does not cancel diving limits. They apply in full:

Diving mode	Maximum depth	Conditions
SCUBA	130 fsw	Standby diver available; diver line-tended or in continuous visual contact with another diver
Surface-supplied air	190 fsw	Standard surface-supplied air limits apply
Surface-supplied air	220 fsw	Only where bottom time is 30 minutes or less

Surface-supplied air is the normal mode for underwater welding, since the diver needs a helmet, communications, power for lighting and tools, and a steady air supply. For surface-supplied air dives deeper than 100 fsw and shallower than 220 fsw, a decompression chamber capable of recompressing a diver to a minimum of 165 fsw (6 ATA) must be at the dive location. It must be dual-lock, multiplace and within five minutes of the dive site. That single line sets the logistics of the whole job: the chamber is not a nice-to-have, it dictates where the barge anchors and whether the tide window works.

Gas-freeing the vessel first

When the job is aboard a vessel rather than on an open structure, an inspection for flammable or combustible gases must be conducted by a certified marine chemist before hot work such as welding or burning begins. That requirement comes through 46 CFR 115.710, which points to NFPA 306 for the standard approach to evaluating and testing enclosed spaces on vessels.

The pattern holds: gas-free first, ground properly, then light the arc with a tender on the switch. Below the surface the same discipline applies as in any certified shop, only the consequences are larger. Above-water marine work has its own rules, covered in our marine welding safety and standards guide, and training programs like our aluminium welding course overview help anyone build a qualification record.

The Craft: Electrodes, Power Sources and the Drag Technique

Understanding why the arc survives underwater is one thing. Doing the job is another. Every element of the equipment stack solves one awkward fact: the diver must never touch the steel bare-handed, and the person holding the power switch cannot see the arc, the puddle or the bead.

The Equipment Stack

A wet weld is assembled from a short, deliberately simple chain of components:

• A DC power source on the surface, usually a generator or rectifier watched by the tender. Constant current is preferred because the diver cannot adjust the machine once the arc is struck.
• Double-insulated welding cables lowered to the diver, so neither the diver nor the wet rope carries a live surface to ground.
• A diver in rubber gloves, ideally over a dry inner glove, working in a helmet with independent air supply.
• A waterproof electrode holder sealed against the water with rubber and tape.
• A knife switch or interrupt switch on the surface, held open except while welding.
• The electrode itself, which behaves in a way no shop-floor rod does.

The surface switch is the safety heart of the operation. Under OSHA 29 CFR 1910.422(g), a member of the dive team in voice communication with the diver must tend the current supply switch and keep it open except while welding is taking place. The rule exists because a diver working blind in dark water may not be able to reach a kill switch, or even feel a problem developing. The tender in the boat is the eyes and the off switch.

Why the Coating Is the Point

Wet welding uses a coated SMAW electrode, and the coating does something with no topside equivalent. On a normal stick the flux burns back at roughly the same rate as the core wire, so arc length is whatever the welder’s hand makes it. Underwater, the flux is formulated so the covering burns away more slowly than the core and projects past the end of the steel.

That projection is an automatic arc length regulator. The diver does not have to hold a precise standoff in swirling water. Instead the rod is buried in the puddle and dragged along the joint: the tip of the coating touches the work, the core is held back a short distance inside it, and the arc runs in the gap between the core and the burning coating, protected inside the vapour bubble. It is why the process works at all.

That also explains why SMAW remains the dominant wet process: no shielding gas, no wire feed, and no moving gas line beyond the diver’s air hose.

Setting the Current

Industry practice sets the underwater current higher than the topside current for the same electrode, because the surrounding water absorbs heat rapidly and robs the arc of energy.

Electrode diameter	Typical working range (horizontal)
1/8 in (3.2 mm)	140–170 A
5/32 in (3.97 mm)	170–210 A
3/16 in (4.76 mm)	220–260 A

Overhead, the ranges narrow. Once the arc is above the puddle, the pool cannot help hold the bead, and heat that would have gone into the weld instead heats the diver’s suit, hood and air hose.

Angle, Gap and Travel

• Electrode angle: roughly 30 degrees to the weld line, with a working range of 15 to 40 degrees depending on electrode and diver preference. Drag makes the angle forgiving in a way a fixed-length arc never is.
• Joint gap: 1/16 in (1.59 mm) is the maximum for wet fillet welding. Wider gaps call for the feeding in technique, where the electrode is held in the opening so weld metal runs into the gap rather than bridging it in one pass.
• Travel direction: weld toward yourself, not away. The bubble rises, and if it crosses your line of sight it blacks out your view of the bead exactly when you need to see it.
• Vertical work: progress downward using the drag technique, keeping the puddle small and letting heat sink and buoyancy work with you rather than against you.

Rods are selected the same way as for any multi-position job, and a good all-position rod guide is worth reading before a job, not after a reject. Material matching matters too: a wrong electrode choice on alloy steel will not be forgiving underwater.

Automated Alternatives

Hand dragging is not the only option, and semi-automatic wet processes suit work that repeats itself or runs long:

• Water-curtain GMAW, developed in Japan and used on the Megafloat semi-submersible, where solid wire is fed through a nozzle with the shielding gas arranged as a curtain of water that keeps the arc region clear.
• A wire-brush shield variant, in which a rotating brush breaks up the bubble and steadies the arc in the same way.
• A self-shielded flux-cored wire process developed in the former Soviet Union, which needs no shielding gas, since the flux core generates its own protection in the vapour bubble.

All three exist because the physics that makes wet welding simple also makes it slow, tiring and inconsistent. Automating the bead is one of the few ways to raise the consistency of an underwater arc without moving the welder into a dry chamber.

Where Underwater Welding Is Used: Real Repairs in Real Water

Underwater welding earns its keep wherever the alternative is to lift, tow and dry-dock a steel structure. That covers a lot of industrial work: a trawler holed below the waterline, a jacket node on an offshore platform, a leaking subsea pipeline, a lock gate at a busy port, a bridge pier eaten by scour.

The shift in attitude is worth noting. AWS D3.6M records that underwater welding has evolved from a means of temporary repair for damaged steel-hulled vessels into an accepted method of construction and repair of engineered structures. It names offshore structures, submerged marine pipelines, underwater port facilities and nuclear power plant components as legitimate applications, and notes that wet welding is routinely used to repair offshore structures, underwater pipelines, docks and port equipment.

The Scenario List

• Emergency hull and pipeline repair. A vessel takes a hit and floods a compartment, or a cargo line splits on the seabed. The fix is a patch plate, a welded sleeve or a spool piece, done in place, in whatever water the boat is floating in.
• Offshore platforms and subsea structures. Jackets, deck legs, risers, clamps and pipeline supports all corrode and all sit in water. Wet welding is the default repair method because a saturation diving spread can reach a platform faster than any heavy-lift campaign.
• Harbour and port facilities. Lock gates, sheet piling, quay walls, fenders, dolphins and navigation structures are permanently wet and permanently abraded. Port authorities schedule diving repairs during weather windows instead of waiting for a dry dock slot.
• Dams, intakes and trash racks. Water treatment and hydro infrastructure accumulates debris, erosion and corrosion on submerged gates and racks. Cleanout and repair happen with divers in the flow path.
• Nuclear power plant components. Intake and discharge structures, screens and conduits sit in water for the plant’s entire life. AWS specifically lists nuclear power plant components among engineered structures where underwater welding is accepted.
• Tie-ins and hot tapping. Repairing a line without shutting it down means welding to a fitting while it is pressurised or live, which raises the stakes on hydrogen cracking and heat input.
• Bridges and piers. Marine bridge foundations, pier caps and piling repairs are done in place to avoid closing a public route or building a temporary cofferdam.

The Economics: Dry-Dock Versus Dive

The business case is blunt. Docking a vessel, towing a platform to a sheltered yard, pumping a cofferdam dry and hiring a full yard crew can cost several times more than the weld itself. AWS puts it more simply: underwater welding reduces repair cost by avoiding transport of offshore and marine structures into the air, and lets the work happen directly in the environment where the damage is.

What that buys is speed and flexibility, and what it costs is metallurgical quality. A wet weld quenched by surrounding water sits somewhere between a temporary fix and a permanent one, and is generally not a substitute for a dry weld on a pressure boundary or a fatigue-critical joint. For the broader fabrication picture, see our guide to marine welding standards and practices.

When Dry Welding Gets Specified

Codes do not treat every underwater weld the same. AWS D3.6M sets three weld classes:

Class	What it means	Typical use
Class A	Requirements comparable to above-water welding	Critical joints, nuclear and pressure-retaining work
Class B	Less critical applications, relaxed criteria	Structural repairs and non-critical members
Class O	Must meet the requirements of another designated code or specification	Work governed by a separate standard, such as a vessel or pipeline code

When the specification says Class A, wet welding is off the table. The job moves into a dry chamber, a dry-spot enclosure or a one-atmosphere habitat, where the welder works without diving gear at ambient pressure and the weld cools like a topside weld.

That is the real picture of where underwater welding lives. Not a single trick with a single use, but a spectrum from fast emergency wet patches on a hull to habitat welding on a nuclear intake, chosen by what the joint has to do for the next twenty years.

Common Underwater Welding Mistakes and How to Fix Them

Wet welding is not forgiving, and the failures a diver makes at depth are usually the same ones a shop welder makes topside: spatter, undercut, porosity, slag in the bead. The difference is the rate and the consequence. A rough bead in a fabrication shop can be ground out and re-run before anyone notices. An underwater weld sits on the bottom of the sea on a hull that still has to sail, and the diver making the repair may be doing it in the dark with a vapour bubble in his face.

Research on wet welding of A36 steel has recorded spattering, irregular bead surfaces, porosity and undercutting as the characteristic defect set, essentially the same catalogue of defects seen in air but appearing at much higher rates (Metals, 2021).

Symptom	Common cause	Fix
Arc wander, heavy spatter	Bubble collapse, current too low for depth	Raise current into the depth-rated range, steady the travel
Porosity, brittle fracture	Hydrogen pickup, excessive heat input	Lower heat input, clean every pass, use dry chamber where quality is critical
Undercut, lumpy bead	Poor travel speed, excess current, fouled steel	Slow the travel, trim current, grind the joint
Slag and inclusions	Slag left between passes	Clean every bead before restarting
Arc will not strike	Coated tip not scraped, damaged coating	Scrape to bare core, inspect holder and cables
Cracks in bead or HAZ	Hydrogen-assisted cold cracking	Reduce heat input, consider post-weld heat treatment onshore

Arc instability and spattering

The vapour bubble around the arc is constantly forming and collapsing. When a bubble wall breaks, the arc length jumps, the pool is disturbed and molten metal is thrown out as spatter. Currents that would be perfectly normal topside often leave the arc hunting at depth, because the water is absorbing heat faster than the electrode can replace it. The first fix is simply to raise the current to the range published for the electrode diameter and the depth. The second is to stop chasing the arc and use the drag technique instead, keeping contact and letting the coating set the arc length automatically. Underwater SMAW is described in full in our guide to why stick welding still wins.

Porosity, hydrogen and brittle fracture

This is the failure that costs the most. Hydrogen from decomposed moisture is absorbed by the molten pool and has almost no time to diffuse out, so pores form and the weld comes up brittle. Both get worse with depth and with excessive heat input, because a hotter, longer-lived pool absorbs more hydrogen. The answer is counterintuitive for welders used to driving heat in: use less, not more. A single pass with a clean, short bead breaks far less hydrogen into the joint than a heavy multi-pass build-up.

Undercut, irregular bead profile and slag

Undercutting is the classic signature of travel speed that is too fast for the current setting, or current that is too high for the thickness, aggravated by a dirty surface. Fix the speed first, then trim the amperage, then look at the steel. Slag behaves differently underwater: the flux floats and does not lift away with the bead, so it is routinely trapped between passes. Industry manuals are unambiguous that each bead must be chipped and cleaned before the next pass is laid down, a step that is tedious, uncomfortable and routinely skipped. It is also the step that decides whether you have a sound weld or a laminated one.

Striking failures, electrode condition and fit-up

An arc that will not strike underwater is usually a coated tip that has not been scraped back to bare core. The coating is the point of the wet electrode, but it will not strike until it is exposed. Other causes are a damaged or water-saturated coating, an electrode holder with compromised insulation, and electrolysis currents bleeding through the salt water. If the holder feels warm, stop.

Fit-up is equally unforgiving. AWS D3.6M limits the joint gap to 1/16 in (1.59 mm). Wider than that and the arc cannot bridge the gap, so the diver feeds the electrode into the joint ahead of the puddle to fill it as he goes. Electrode choice also matters: rutile coatings are widely reported to give the best wet results because they strike easily and hold a stable arc, though they are less resistant to hot cracking during solidification, and flux-cored designs with Fe₂O₃ additions have been reported to cut diffusible hydrogen to 14.6 ml/100 g at 26 kJ/cm. Picking the right rod for the job is covered in our all-position welding rod guide.

Overheating, overhead work and the diver’s own safety

Too much heat input produces a hard, low-ductility weld with a wide heat-affected zone, and it is the most common route to hydrogen cracking. Some of the heat goes somewhere worse: into the diver. Industry guidance is to minimise welding in the overhead position, because falling molten material can seriously damage the diving suit and the air hoses, and a suit breach or a severed hose underwater is a life-threatening event. Where overhead work is unavoidable, the crew on the surface needs to know exactly when the arc is live.

Surface preparation

Before any arc is struck, thick paint, rust, scale and marine growth must be ground away from the joint. Foul and paint burn into the bead, generate gas, poison the flux and block the arc. Welders should treat preparation as a prerequisite rather than a nicety, and remember that grinding underwater throws a cloud of abrasive and debris into the water the whole team is breathing.

Inspecting and Testing Underwater Welds

Completing an underwater weld is only half the battle. Verifying its integrity is often the grueling part. Between limited visibility, strong currents, high pressure, and cramped access, inspection underwater is drastically harder than it is topside. Because of this altered environment, AWS D3.6M requires that special quality requirements and inspection procedures be established for every underwater weld before construction begins.

These procedures are not a formality. For any repair tied to critical marine operations, such as those described in the AJ Marine Welding guide, a failed inspection means rework at depth, extended downtime, or, in the worst case, a catastrophic structural failure.

The AWS Weld Classes

Not every underwater repair has to match the quality of a brand new, shop-fabricated weld. To account for that, AWS splits underwater welds into three classes, each with its own expectations.

• Class A – The highest quality level. Intended to be comparable to above-water welding and typically specified for critical, pressure-retaining, or load-bearing structures.
• Class B – A step down from Class A, used where the consequences of failure are significantly lower.
• Class O – Meets another designated code or specification rather than the default AWS mechanical or soundness requirements for Class A or B.

For each class, AWS defines two sets of criteria: one for weldment properties that must be established during qualification, and one for weld soundness that must be verified during construction. A weld therefore cannot simply look right. It has to prove its mechanical properties through qualification and be confirmed sound through inspection on the job. That testing follows the standard’s referenced documents, most notably AWS B4.0, along with ASTM E165 for penetrant testing, ASTM E709 for magnetic particle testing, ASTM E1416 for acoustic emission, ASTM E164 for ultrasonic testing, AWS D1.1, and ANSI Z49.1 for welding safety.

The Standard NDT Toolbox Underwater

Depth and physical restrictions mean inspectors rely on a selective toolbox of non-destructive testing (NDT) methods. As echoed in a 2021 NAVSEA ship drydock specification, the standard methods used for underwater welds are these.

Method	What It Reveals	Underwater Reality
Visual Testing (VT)	Cracks, undercut, poor bead shape, surface damage	Diver inspection with a light, or topside cameras and submersibles for larger or deeper repairs
Dye Penetrant Testing (PT)	Surface-breaking flaws in non-porous materials	Surface is cleaned and degreased, penetrant applied and removed, then developer draws out indications; governed by ASTM E165
Magnetic Particle Testing (MT)	Cracks and laps near the surface in ferromagnetic material	A magnetic field is induced and ferrous particles gather at discontinuities; governed by ASTM E709
Radiographic Testing (RT)	Internal defects such as porosity and lack of fusion	Gamma or X-ray source with shielding and strict radiation control; logistically intensive and needs a specialist crew
Ultrasonic Testing (UT)	Internal discontinuities using sound waves	A popular alternative to RT, but it needs purpose-built submersible probes and strict adherence to ASTM E164

Qualification, Personnel, and the Hardest Obstacles

All of these methods share one massive hurdle: they cannot be performed by unqualified personnel or with topside procedures copied over verbatim. Underwater NDT demands its own qualification trials, tied to conditions that are practically unavoidable at depth, such as poor visibility, awkward orientation including vertical and overhead positions, limited dexterity in bulky gloves, the difficulty of repositioning large equipment, and the absolute necessity of thorough coating removal before testing.

Every part of the inspection must also be planned for the environment in which the weld was produced. A procedure that passes on a dry, one-atmosphere weld may be wholly insufficient for a wet weld, which is exactly why the AWS requirement for site-specific quality and inspection procedures is non-negotiable. Only personnel qualified for both their NDT discipline and the underwater application can interpret results reliably, and a false call or a missed flaw is exponentially more expensive to correct at depth than it is on land.

Inspection is what upholds the line between an acceptable, engineered repair and a short-term patch. With weld classes setting the level of scrutiny, a verified toolbox of NDT, and a hard demand for qualified divers and inspectors, the industry holds underwater welds to a very high bar. It is not about how fast the arc was struck. It is about the proof, dive by dive, that the joint will hold for years to come.

Conclusion: Choosing the Right Underwater Welding Method

There is not one technique but five, and picking the wrong one is expensive in money and risk. The choice is less technical than it sounds, because AWS D3.6M:2017 already lays out the framework. It recognises the full spectrum of environments, from fully wet to fully dry, and leaves weld class selection to the customer.

The Default: Wet Welding

For most emergency repairs, wet welding wins by default, for practical rather than technical reasons. A diver can be in the water and depositing weld metal within hours of mobilising: no chamber, no habitat, no dewatering, no crane lifting a sealed enclosure into place. The diver descends, cleans the joint, starts the arc and drags the electrode along the seam.

Wet welding also handles awkward geometry better than any dry method. A dry chamber needs space, access and a surface structure to hold a diver dry at working depth, and subsea pipelines, congested hull plating and slender members often offer none of that. Where access is poor, wet is the only realistic option.

You accept lower ductility and impact toughness, a rougher bead, more hydrogen pickup and higher diffusible hydrogen than an equivalent above-water weld. If the code calls for an AWS Class A weld, wet welding will not satisfy it. But for secondary structural repairs and keeping a vessel in service, the trade is worth it. The all-position welding rod guide shows how underwater consumables compare with conventional filler metals.

When Dry Methods Win

Move up the ladder when quality, not speed, sets the agenda.

• Dry chamber welding places the diver in an open-bottomed chamber flooded to the surface, giving a dry working volume at depth. Suits long, repetitive seams with a built fixture.
• Dry-spot welding clamps a small transparent gas-filled enclosure over the joint while the diver stays outside. Fast to set up, good for local repairs on critical members.
• Habitat welding seals a working chamber at ambient pressure so a welder in ordinary coveralls works at the surface without diving gear, with better results and ergonomics.
• One-atmosphere hyperbaric welding depressurises the space to roughly one atmosphere, using topside procedures. Weld metal approaches above-water quality. The premium option for nuclear and pressure-retaining work.

The decision rule is simple: if the weld must approach above-water quality, if the specification demands AWS Class A, or if failure would be catastrophic, dry methods are not optional. Their extra cost buys metallurgy you can specify, document and defend.

The Physics, in One Paragraph

All of these share one piece of physics. The arc vaporises a cavity of water into a gas bubble around the electrode and molten pool, and that improvised shield keeps the arc alive. Roughly 70 percent of that bubble is hydrogen, produced as water decomposes and absorbed by the molten metal. Because water quenches the weld and its heat-affected zone far faster than air, there is almost no time for hydrogen to diffuse out. The result is hard martensitic and bainitic microstructures, high diffusible hydrogen and a hydrogen-assisted cracking risk that rises with depth, and that 70 percent hydrogen share is the root of the risk that separates wet welding from every other arc process. TWI explains the arc-in-water problem in similar terms, and researchers continue to refine the data on underwater weld microstructures and hydrogen levels.

The Safety Frame Does Not Bend

Whichever environment you choose, this is commercial-diver hot work governed by OSHA 29 CFR 1910 Subpart T and the procedures in 1910.422. The most important safety control in wet welding is the surface-tended current interrupt the diver can call at any moment: if something goes wrong, the tender kills the arc without the diver reaching for anything. That switch, a grounded machine frame, insulated cables and holders, insulated gloves and gas-freeing of closed compartments are the reason the job is survivable.

The Honest Takeaway

Underwater welding is not a shortcut around proper welding. It is a different trade with different rules, metallurgy and failure modes. The vapour bubble that keeps the arc alive is also the source of the hydrogen that threatens the weld, and the water that reaches the joint in minutes is the same water that quenches it brutally. Knowing how stick welding works helps, but the environment changes the fundamentals.

So choose deliberately: wet for speed and access, dry when the specification or the consequence demands it, hyperbaric when the stakes are nuclear. Then treat it as hot work performed by commercial divers under a specific code, with a tender holding the only switch that matters.

Frequently Asked Questions

Can you actually weld underwater with an electric arc?

Yes. The arc between a flux-coated electrode and the workpiece vaporizes the surrounding water into a cavity of steam and gas that surrounds the arc and molten pool, and that cavity acts as an improvised shielding atmosphere. Everything else about the circuit is ordinary shielded metal arc welding – only the environment is brutal.

How is underwater welding different from dry welding?

Dry welding isolates the welder and the arc from the water, usually in a chamber or hyperbaric habitat, and produces weld metal comparable to above-water welds, but it is slow and expensive. Wet welding leaves the arc in the water, so it is faster and cheaper but produces more hydrogen, faster cooling, more porosity and lower ductility.

Are underwater welds weaker than welds made above water?

Not necessarily in strength. Research on A36 steel recorded underwater tensile strengths above the air-welded comparison, and naval testing has produced underwater fillet welds with over 80% of the air-weld tensile strength. The real penalties are ductility, toughness and crack risk – roughly 50% of air-weld ductility in those naval tests, plus hydrogen-assisted cracking.

Why is hydrogen such a problem for underwater welds?

The vapour bubble around the arc is roughly 70% hydrogen, and the molten pool absorbs it, but the surrounding water quenches the weld so fast that the hydrogen has no time to diffuse out. Diffusible hydrogen can reach around 100 ml/100g with rutile electrodes, and the risk of hydrogen-assisted cracking grows with depth.

Do I need a diving chamber to weld underwater?

Only for the dry methods. Wet welding needs no chamber at all – just a diver, a waterproof electrode and a current source on the surface. A dry chamber, dry-spot enclosure or habitat is used when the weld quality has to match above-water work.

What electrodes are used for underwater welding?

Wet welding is mostly SMAW with a waterproof-coated electrode whose flux burns back more slowly than the core, giving an automatic arc length, and rutile or basic types chosen for arc stability and hydrogen control. Automated wet processes use solid wire with a water curtain or wire brush shield, and self-shielded flux-cored wire that needs no shielding gas.

Is underwater welding safe?

As a commercial trade operation, yes – when it is run as regulated hot work. OSHA 29 CFR 1910 Subpart T requires a surface-tended current interrupt, a grounded machine frame, insulated cables and gloves, and gas-freeing of closed compartments before hot work. It is absolutely not a recreational or self-taught activity.

Does salt water make underwater welding easier?

Partly. Dissolved salts make the water more electrolytic, which tends to produce a hotter, more stable arc, and one salinity study found HAZ hardness dropping from 211 to 193 HV10 and impact strength rising from 82.5 to 101.3 J/cm2 as salinity increased from fresh water to 35 parts per thousand, with fewer microcracks.

How deep can underwater welding be performed?

Depth affects the weld metal, not the legal limit. Studies show tensile strength rising and microstructural changes as depth increases from 2.5 m to 10 m, and hydrogen content rising with depth. The practical limits come from diving rules – SCUBA to 130 fsw and surface-supplied air to 190 fsw under OSHA.

How are underwater welds inspected?

Using the standard toolbox – visual testing, dye penetrant, magnetic particle, radiographic and ultrasonic testing – but every procedure needs its own underwater qualification because of access, orientation and coating problems. AWS D3.6M ties inspection to the weld class, with Class A expected to be comparable to above-water welding.