Chapter 5: The Arc

The arc was a point of light in the darkness.

At 150 metres, in the North Sea, in winter, with the visibility reduced by suspended sediment to something between one and three metres, the welding arc was the brightest thing in the world. It was brighter than the helmet light. It was brighter than the bell's external lamps. It was brighter than anything Kwame had seen since the last time he had struck an arc at depth, because at depth there was nothing else — no sun, no reflected light, no ambient illumination of any kind — and the arc existed in contrast to the total dark the way a single voice exists in contrast to total silence: absolutely, without competition, the only thing of its kind.

The arc was approximately 5,000 degrees Celsius at its core. The seawater three centimetres from the arc was 3.2 degrees Celsius. The gradient between these two temperatures occurred over a distance shorter than a man's thumb, and it was Kwame's job to manage that gradient, to keep the arc stable and the weld pool molten and the electrode consuming at the correct rate while the ocean pressed against him at fifteen atmospheres and the current tried to push him sideways and the cold crept into the margins of his hot water suit and his breathing gas hissed through the demand regulator with each inhalation, a sound that was both mechanical and intimate, the sound of his own survival.

He was welding the first joint of the spool piece — the upstream connection, where the new section of pipe met the existing pipeline. The joint was a single-V butt weld, the standard configuration for pipeline girth welds, the edges of the two pipes bevelled at thirty-seven and a half degrees to create a groove that would be filled with weld metal, layer by layer, pass by pass, until the joint was complete and the two sections of pipe were one.

The first pass was the root. The root pass was the most critical — it was the foundation of the joint, the first layer of weld metal deposited into the narrow gap at the base of the V, and it had to be perfect. Not good. Not acceptable. Perfect. Because the root pass formed the inner surface of the pipeline, the surface that would be in contact with the product — in this case, natural gas at high pressure — and any defect in the root, any lack of fusion, any porosity, any crack, would be a point of weakness that the gas pressure would find and exploit, the way the ocean found and exploited the stress corrosion crack that had brought Kwame to this pipeline in the first place.

He positioned himself. Kneeling on the seabed beside the pipe, his left hand bracing against the alignment clamp that held the spool piece in position, his right hand holding the electrode holder with the 7018 rod inserted, the rod angled at fifteen degrees to the pipe surface, the tip positioned at the root gap, the gap itself approximately three millimetres wide, wide enough for the molten metal to penetrate through to the inner surface, narrow enough for the weld pool to bridge.

He had done this ten thousand times. The number was not a figure of speech. Eleven thousand hours of hyperbaric welding, at an average of perhaps one root pass per hour when accounting for preparation, positioning, rod changes, and inspection, gave a conservative estimate of ten thousand root passes in his career. Ten thousand times he had positioned the rod at the root gap and struck the arc and watched the metal melt and flow and solidify, the transformation from solid to liquid to solid occurring in the small bright space beneath the arc, the metallurgy happening in real time, the iron and carbon and manganese and silicon combining according to the phase diagram that governed their behaviour, the crystal structure forming as the metal cooled, the grain boundaries establishing themselves, the weld becoming steel.

He struck the arc.

The light erupted. Behind his welding filter — the darkened glass in his helmet viewport that reduced the arc's brightness to a level the eye could tolerate — the arc was a sharp white point surrounded by a halo of blue-white light, and in the light the weld pool was visible: a small oval of molten metal, bright orange at the centre, darker at the edges where the metal was solidifying, the pool moving along the root gap as Kwame moved the electrode, the rate of travel controlled by his hand, by the years of practice that had calibrated his hand to move at the speed the weld required — not too fast, which would produce a thin, underfilled bead, and not too slow, which would produce an overfilled bead with excessive heat input and a wider heat-affected zone.

The heat-affected zone was the enemy. It was always the enemy. The zone of metal adjacent to the weld that was heated by the welding process but not melted — heated to temperatures that altered the steel's microstructure, that could make the metal harder, more brittle, more susceptible to cracking. The stress corrosion crack that had damaged this pipeline had initiated in the heat-affected zone of the original construction weld, and Kwame was aware, with the awareness that came from understanding the metallurgy at a level deeper than procedure, that his repair weld would create its own heat-affected zone, its own band of altered metal, its own potential vulnerability.

The difference was in the execution. The difference was in the heat input, the interpass temperature, the cooling rate, the electrode selection, the technique — all the variables that a welder controlled, or attempted to control, in the small bright space beneath the arc. The original construction weld had been made from a lay barge on the surface, the pipe rotating as the barge moved forward, the welders working in the open air, in daylight, with gravity doing half the work. Kwame's repair weld was made at 150 metres, in a helium-oxygen atmosphere at fifteen atmospheres, in darkness, in cold, with the pipe fixed and the welder moving, the welder's body positioned in whatever orientation the joint required — flat, vertical, overhead — the body adapting to the geometry of the pipe the way water adapts to the shape of its container.

The root pass took forty minutes. Kwame worked in segments — each electrode lasted approximately two minutes before it was consumed, and between electrodes he stopped, chipped the slag from the previous bead, inspected the weld with his helmet light, and inserted a new rod. The rhythm was: weld, stop, chip, inspect, rod, weld. A percussive rhythm, the tapping of the chipping hammer on the pipe audible through the water, transmitted through his helmet as a dull metallic ring, the sound of quality control.

Mac was beside him, tending the umbilical, keeping the fifty metres of bundled hose and cable from tangling on the seabed structures, passing tools when needed, watching. Mac was a good diver and a competent welder, but Kwame was the lead welder on this rotation, the man who would make the critical joints, and Mac's role during the welding was support — umbilical management, tool passing, observation. Mac watched Kwame weld the way a musician watches another musician play: with professional attention, noting the technique, the arc length, the travel speed, the angle, the things that could be taught and the things that could not.

Kwame completed the root pass and stopped. He keyed his communications. Kwame to dive control. Root pass complete on the upstream joint. Requesting visual inspection before proceeding to hot pass.

Sarah Webb's voice: Confirmed. Hold position. We're reviewing the video.

They watched through his helmet camera. The dive superintendent, the welding engineer, the client representative — they all watched the video feed from Kwame's camera, the image showing what Kwame saw: the completed root pass, a continuous bead of weld metal running around the circumference of the pipe at the base of the V-groove, the bead uniform in width and profile, the surface smooth, the colour — visible in the light of his helmet lamp — the silvery-grey of properly deposited low-hydrogen weld metal.

The welding engineer came on the line. His name was Andrew, and he had the particular authority of a man who understood welding at the molecular level and who could read a weld bead the way a cardiologist reads an ECG — not just the shape but the story, the thermal history, the metallurgical narrative encoded in the surface texture and the colour and the width and the profile.

Andrew said: Root looks good. Consistent profile. No visible porosity. No undercut. Proceed to hot pass. Maintain interpass temperature below 250 degrees.

Kwame checked the interpass temperature with a contact thermometer pressed against the pipe surface near the root bead. Two hundred and twelve degrees Celsius. Within limits. The pipe was cooling, the heat from the welding being conducted away by the steel and the concrete coating and, ultimately, the seawater, the ocean acting as an infinite heat sink, drawing the thermal energy away from the weld and into the three-degree water and dissipating it into the vastness of the North Sea, where it would raise the water temperature by an amount so small it could not be measured — the heat of Kwame's arc, the energy of his work, absorbed into the ocean's indifference.

He began the hot pass — the second layer, deposited directly onto the root, filling the V-groove to approximately one-third of its depth. The hot pass was less critical than the root but more demanding physically, because the welder had to work faster to maintain the interpass temperature, and working faster at 150 metres, in a hot water suit, with an umbilical trailing behind and a current pushing from the side, required a sustained physical effort that was closer to athletics than to craftsmanship, the body working at a rate that the helium atmosphere taxed — the breathing harder, the muscles less efficient, the coordination fractionally impaired by the pressure.

Kwame welded. He did not think about technique. He had passed the point, years ago, where technique required thought. His hands knew what to do. His hands knew the arc length — two to three millimetres, the distance between the tip of the electrode and the surface of the weld pool, maintained by the continuous small adjustment of the wrist and the forearm and the shoulder, the body acting as a machine, a biological positioning system that held the arc at the correct distance through the feedback loop of sight and sound and the feel of the rod against the metal.

What he thought about, while his hands welded, was the light.

The arc was a small sun. It existed only because he had struck it, only because he held the electrode at the correct distance and maintained the current through the circuit — electrode to weld pool to pipe to ground cable to welding machine to electrode — and if he withdrew the rod by even five millimetres the arc would extinguish and the light would die and the darkness of the North Sea at 150 metres would return, instant and complete, the way darkness returns when you blow out a candle.

He was making light. At the bottom of the sea, in the dark, in the cold, he was making a small, intense, impermanent light that existed only as long as the work continued, only as long as the rod was in contact with the metal, only as long as the current flowed. When the rod was consumed, the light died. When he inserted a new rod and struck a new arc, the light was born again. This cycle — light, darkness, light — was the rhythm of his work, and it was, he had come to understand, the rhythm of his life.

He was illuminated only when he was working. When the work stopped — when the rotation ended, when the decompression was complete, when he surfaced and returned to Aberdeen and the flat on Crown Street — the light went out, and he moved through his surface life in a kind of darkness, the darkness of a man who was not doing the thing he was made for, the thing that gave him his small sun.

This was not a metaphor. This was a statement about competence. Kwame was competent at welding. He was competent at diving. He was not competent at the rest of it — the surface, the flat, the relationships, the letters, the phone calls, the being-present that his mother and his brother and the world required of him. At 150 metres, with an arc burning and a weld pool flowing and the darkness held at bay by his own hands, he was exactly the person he needed to be. On the surface, at one atmosphere, in the Aberdeen daylight, he was a man in a flat with a pile of unread letters, and the letters were evidence of his incompetence at the thing that mattered most.

He finished the hot pass. Then the fill passes — three of them, each one building the weld higher in the V-groove, the joint thickening, the metal accumulating, the cross-section of the weld growing from a thin line at the root to a broad mass at the surface. The fill passes were the body of the weld, the structural bulk, and they required consistency more than precision — each bead laid parallel to the last, each one fusing to the previous pass and to the groove walls, the weld metal building up the way sediment builds up on the sea floor, layer by layer, each layer a record of the conditions under which it was deposited.

The cap pass was last. The final layer, the surface of the weld, the face that would be inspected visually and by radiography and ultrasonics. The cap had to be smooth, uniform, with a slight convexity — the weld raised fractionally above the pipe surface, the reinforcement — and with no undercut at the toes, no porosity on the surface, no cracks.

Kwame laid the cap with the care of a man signing his name. Each bead was a signature — the width, the profile, the surface texture, all of it particular to him, recognisable to anyone who had seen his work before. Andrew the welding engineer could identify Kwame's welds by sight, the way a typographer could identify a typeface, because every welder's technique produced a distinctive bead, and Kwame's beads were characterised by a consistency that bordered on mechanical, each one the same width, the same height, the same smooth surface, as though they had been deposited by a machine rather than by a man kneeling on the sea floor in the dark.

But they had not been deposited by a machine. They had been deposited by Kwame Asante, thirty-six, from Osu, Accra, whose hands had been trained first by his uncle Kofi in a welding shop behind Makola Market, where the rods were cheap and the steel was scrap and the training method was repetition — weld, grind, weld, grind, weld, until the hands knew what the mind could not yet articulate — and then by the instructors at the welding school in Aberdeen, where the training was more systematic but the principle was the same: the hands had to learn, and they learned by doing, and doing meant hours, and hours meant years, and years meant the accumulation of a skill that was, in the end, not knowledge but sensation, the feel of the rod against the metal, the sound of the arc, the smell of the flux — all the sensory data that the body collected and processed and stored, bypassing the conscious mind, becoming reflex, becoming the thing that allowed Kwame to weld a perfect root pass at 150 metres while thinking about light and darkness and letters and the distance between Aberdeen and Accra.

He completed the upstream joint in four hours. Four hours of welding, with breaks for rod changes and inspection and communication with the surface, the work proceeding at the pace that depth and pressure demanded — not slow, not fast, but deliberate, each action considered, each decision weighted, each movement made with the awareness that at 150 metres there was no margin for error that could not be easily corrected.

A bad weld at depth could not be ground out and re-done the way a bad weld at surface could. At depth, every repair was a major operation — removing the helmet, returning to the bell, re-rigging the equipment, descending again, repositioning, re-starting. A bad weld could cost hours, could push the work scope into the next shift, could extend the rotation. So Kwame did not make bad welds. He had not made a bad weld — defined as a weld that failed inspection — in three years, a record that he did not mention to anyone because in saturation diving the reward for consistent quality was not recognition but continuation, the opportunity to descend again tomorrow and do the same work in the same conditions with the same precision and return to the chamber and eat the cardboard food and sleep the fragmented sleep and wake and do it again.

Mac's voice came through the communications. Looking good, Kwame. Cap's clean.

Kwame ran his gloved hand over the completed joint. Through the neoprene, he could feel the contour of the weld — the slight convexity of the cap, the smooth transition at the toes, the uniform width. It was, by touch, a good weld. The radiography would confirm it. The ultrasonics would confirm it. But his hand already knew.

He keyed the communications. Kwame to dive control. Upstream joint complete. Ready for NDE.

Andrew's voice: Confirmed. We'll run the crawler tomorrow. Good work, Kwame.

Good work. Two words that, in the dive industry, carried more weight than any commendation or certificate, because they came from a welding engineer who had reviewed the video and the parameters and the technique and had concluded, on the basis of evidence, that the weld was acceptable. Not beautiful. Not inspired. Acceptable. Which was, at 150 metres, the highest praise.

They returned to the bell. The ascent. The trunk. The chamber. The routine of return — equipment removed, shower taken, food consumed, the body transitioning from the intensity of the sea floor to the stasis of the chamber, the adrenaline receding, the fatigue arriving, the particular exhaustion of six hours of physical and mental work at fifteen atmospheres settling into the muscles and the bones.

Kwame sat at the table and ate his dinner — roast chicken, potatoes, vegetables, all tasting of the helium atmosphere's version of these things, which was to say they tasted of approximately nothing, the flavour stripped by the pressure and the gas the way colour is stripped from fabric by bleach, leaving the shape intact but the quality altered.

Mac sat across from him. They ate in the comfortable silence of men who had just completed a dive together, the silence of shared effort, the silence that was different from the silence of the chamber's dead hours because it contained something — the residue of the work, the satisfaction of the completed joint, the professional regard of two men who had done well in difficult conditions and who did not need to say so.

Kwame opened his notebook after dinner. He wrote: Upstream joint complete. Root, hot, fill x3, cap. Total time 4 hours 10 minutes. 7018 rods, low hydrogen, 2.5mm diameter for the root, 3.25mm for the fill and cap. Arc voltage 24-26V. Travel speed approximately 100mm per minute on the root, 120mm on the fill. Interpass temperature maintained below 250 degrees. Visual inspection satisfactory. NDE tomorrow.

This was the technical record. The facts. The data.

Then he wrote: Uncle Kofi would not recognise what I do. He taught me to weld in his shop in Accra — mild steel, 6013 rods, no shielding gas, the sparks falling on concrete and the smoke rising into the Accra air and the ceiling fan turning slowly above the bench because the electricity was not strong enough to make it spin fast. He taught me the thing that cannot be taught: the feel. The sound of a good arc. The way the pool moves when the technique is right. He did not know the metallurgy. He did not know the phase diagrams or the CCT curves or the hydrogen diffusion coefficients. He knew the feel, and he taught me the feel, and the feel is what I use at 150 metres when the metallurgy and the phase diagrams and the procedures are all in place but the weld is still made by hands, by my hands, by the hands that Uncle Kofi trained on scrap steel behind Makola Market.

He closed the notebook. The chamber hummed. Sean was talking — something about a football match, the words distorted by the helium into a high-pitched narration that Kwame could follow if he concentrated and that he chose not to concentrate on because the sounds of Sean's voice had become, over the days, a kind of ambient music, a human frequency in the mechanical symphony of the chamber, and Kwame let it wash over him the way he let the current wash over him at depth — present but not resisted, a force he had learned to accommodate.

He thought about his uncle Kofi. He had not seen Kofi in six years either. Kofi was sixty-eight now, if he was alive. He might not be alive. Kwame did not know. This was the mathematics of distance — the longer you stayed away, the more the information degraded, the fewer data points you had, the less accurate your model of home became. Six years ago, Kofi was sixty-two, healthy, still working in the shop, still welding. In six years, anything could have changed. Everything could have changed. Kwame would not know until he surfaced, until he opened the letters, until the accumulated dispatches from Accra filled in the gaps that his absence had created.

Or he could call. He could pick up the chamber phone and call Yaw and ask. The unscrambler would process his helium voice into something approximating human, and Yaw would hear something approximating his brother, and they could exchange something approximating information about the family, about Kofi, about their mother, about the house and the neighbourhood and the market and all the small particulars of a life that Kwame had left and not returned to.

He did not call. He lay in his bunk and breathed the heliox and listened to the chamber and felt the pressure — all of it, the physical pressure of fifteen atmospheres and the psychological pressure of the things undone and the temporal pressure of the days passing, each one a day closer to the surface and a day further from the last time he had been home — and he closed his eyes and waited for sleep, which came as it always came in the chamber, in fragments, in intervals, in the small surrenders of consciousness that the helium atmosphere permitted.

The pressure held. The weld held. The distance held.

Everything held, because holding was what pressure did — it held things in place, it maintained the state, it prevented the change that would occur if the pressure were released, the explosive expansion that the industry called decompression and that life called returning, and both were processes that required time and care and the slow, controlled release of the forces that had built up during the compression, during the dive, during the years of being at depth, at distance, in the between.

Kwame slept. The chamber kept him. The North Sea kept the chamber. And somewhere in Accra, in the house in Osu, his mother sat at the kitchen table and wrote a letter to her son, her pen moving across the blue aerogramme with the care of a woman who believed that written words carried weight, that they travelled not just through space but through the particular medium of a mother's attention, and that if she wrote carefully enough, precisely enough, with enough love folded into the letters of his name, the words would reach him wherever he was — even at the bottom of the sea, even inside steel, even in the place between the surface and the floor where her son lived his strange, pressured, illuminated life.