Rust: The Longest War

by Jonathan Waldman

  While he showed me the tank, a pump alarm chirped twice, because the resistance on the impeller was increasing. Neogi needed to take it out, soak it in vinegar, and reinstall it. But he already knew that, because his fish tank is remote controlled. Sensors in the tank monitor pH, oxygen level, specific gravity, ozone, dissolved oxygen, total dissolved solids, temperature. Other sensors measure power draw, humidity, lighting, and leaks. From his phone, Neogi can monitor and control his fish tank. He can adjust the lights. If the power goes out, a backup battery will take over, and he’ll get an email. If the power remains out for more than a day, a backup generator will keep his tank running for a week. If a fish detector detects no moving fish, he’ll be sent an email. If the temperature changes more than two degrees in two hours, he’ll get an email. Thanks to a motion detector, he can tell if a caretaker is feeding his fish as much as he or she ought to be. A video camera monitors his fish tank in real time. In other words, Neogi has built a miniature simulacrum of a triple-redundant fail-safe pump station in his basement.

  MILE 160

  By the time the pig was launched on March 26, it had been cleaned and recalibrated, its ceramic sensors swapped with others less brittle, its batteries replaced. Neogi made sure that the technicians knew how much oil the pig demanded to get moving, and launched the pig on schedule at four in the morning. He timed the launch so that the pig was still twelve hours behind another scraper pig, and so that it would crest Atigun Pass at noon, and then—in daylight—face the major challenge of the run.

  That challenge was the far side of Atigun Pass, down which a pig, unconstrained by oil—in what’s known as slackline—could hit ninety miles per hour. At that speed, the sensors cannot gather data and can easily melt or crack, rendering the pig blind. If blindness ensued, Neogi would have to rerun the whole section from Pump 4 to Valdez. That would be a multimillion-dollar mistake. It was little consolation to Neogi that it had happened before. He didn’t want a failed pig run on his watch.

  Alyeska had been dealing with slackline at the bottom of Thompson Pass, just north of Valdez, for twenty years before residents in a neighborhood called Heiden View began complaining of earthquakes. The shaking woke them up at night. The pulses were caused by a torrent of oil crashing into a flat pool of oil at the bottom: a waterfall within the pipe. Engineers determined that the vibrations threatened the integrity of the pipe and that the induced fatigue had a small chance of cracking it. Alyeska responded by installing back-pressure control valves in Valdez that pushed the “slackline interface” 1,800 feet higher, but not quite all the way to the top of the pass. Over the Chugach Mountains, the oil still cascades almost a half mile down, like a truck with no brakes.

  Atigun Pass, over the Brooks Range, is almost twice as high as Thompson Pass. Pulsing at the bottom compelled Alyeska to install a sleeve—essentially a metal cast—there in 2003. Because there are no back-pressure control valves to constrain the oil through the descent at Atigun, engineers at Alyeska rely on a complicated, concerted technique known as tightlining to guide the pig through.

  The plan for Neogi’s run, formed months earlier, was to stockpile crude in tanks north of the pass and push as much oil as possible through the line ahead of the pig. Then, so as not to overpressure the pipe on the south side of the pass, Alyeska would inject a drag-reducing chemical into the oil, divert the extra oil into other tanks, and send the rest down the line. By splitting the flow, Alyeska would put a bend in the hydraulic gradient at that pinch point. In other words, it would keep the pipeline from bursting. It was a tricky procedure, evaluated and refined by hydraulic engineers, tested virtually in Alyeska’s pipeline simulator, scrubbed by operational engineers, assessed for hazards, and practiced ahead of time by Alyeska. Because the procedure entailed bringing the actual pressure so close to the maximum allowable pressure, the company sought permission to go ahead from the federal Pipeline and Hazardous Materials Safety Administration. PHMSA, a branch of the Department of Transportation, wants to ascertain the condition of the pipeline as much as Alyeska but not at the expense of prudence. The agency granted permission. It also asked to watch the procedure.

  Early in the morning, Neogi sent the smart pig southbound. In an ultrasecure, unmarked, nearly windowless concrete building hundreds of miles to the south, one of Alyeska’s operators watched carefully, on high alert. Operators there respond to emergencies much like pilots. They do not panic. In the building, regulators from PHMSA watched the same screens. The smart pig was represented by a little pink capsule labeled “030,” for pig number 30. Shortly after noon, it crested Atigun Pass. One screen flashed brightly, indicating that the leak-detection system was offline. The flow, normally 560,000 barrels per day, more than doubled. On a large flat-screen displaying the hydraulic profile of the pipeline, a yellow line (actual pressure) bumped right up against a blue line (maximum allowable pressure).

  Back in the Brooks Range, just below the pass, three crews were stationed at gate valves in case anything went wrong; in case the pipeline burst. Neogi remained at Pump 4, where he listened to updates over the radio.

  Nothing went wrong. Shortly after one o’clock, Neogi called his boss and said, “We’re good. We only went up to nine hundred psi.” He was in good spirits. The pressure was right on the nose. The major hurdle between Prudhoe Bay and Valdez had been cleared in style. To celebrate, the operators and hydraulic engineers feasted on a tray of pulled pork.

  Smart pigs weren’t smart enough to find corrosion during the pipeline’s first decade of operation. The first ones could detect only wall losses of 50 percent and, according to the engineer who ran them, “didn’t work very well; they couldn’t find anything.” As another Alyeska engineer put it, “A high level of verification of corrosion signals to actual natural corrosion was not achieved.” At least twice in the 1980s, pigs found buckles on the verge of becoming leaks and earned some respect. In two spots, the forty-eight-inch pipe was only forty-two inches across. Responding to a buckle under the Dietrich River, where the pipe had settled fifteen feet, Alyeska sent out crews immediately. It was 67 below. But by investigating settlement of the pipeline, which led to a couple of leaks, crews found corrosion. They knew their pigs were missing something.

  In the spring of 1987, Alyeska ran its first high-resolution MFL pig, made by a Canadian company called International Pipeline Engineering. Along the line, it found a dozen potential anomalies. Alyeska dug up each site and found corrosion at every spot. The next year, Alyeska ran IPEL’s pig again. Nineteen eighty-eight had been a banner year for TAPS. On one day in January, more oil made it through the line than on any other day in the line’s history. In June Alyeska celebrated its tenth anniversary, having shipped five billion barrels of oil. No pigs got stuck that year. No corrosion repairs were made. No emergency shutdowns were necessary. The pig run was in the fall, and everything changed.

  Reading thirteen miles of paper, analysts determined that the pig had found 241 potential anomalies. Field investigations verified two-thirds of them, revealing pits as big as quarters. Alarmed, Alyeska had IPEL reanalyze the data. The second analysis was worse: over 400 anomalies. “We thought we wouldn’t have a problem with corrosion,” Alyeska’s pigging engineer at the time said, “but our world changed when the pigs told us otherwise.”

  Nineteen eighty-nine was rough from the outset. First, in January Alyeska dug up the pipe and installed thirty sleeves. Then, in March there was the Exxon Valdez oil tanker spill. Cleanup notwithstanding, it had little to do with Alyeska. If anything, it revealed the merits of shipment by pipelines over barges—but the result was increased scrutiny and ill will. In June Alyeska ran its first high-resolution ultrasonic pig. It was made by NKK, and it could measure 10 percent wall-thickness loss. The three-ton, red-cupped, titanium tool was five times more precise than IPEL’s pig, accurate to the millimeter. It stored information in a “black box” of the type employed in aircraft. NKK had five engineers on its crew, and they brought gifts for Alyeska employees.
They lived at Pump 1 for a month, and did calisthenics in the manifold building. They wore matching green helmets and uniforms, bloused where they entered their boots, and communicated by hand signals. Before loading their pig, they rolled up a Shinto prayer, put it in the tool, then got on their hands and knees and prayed to it. The Alyeska guys should have prayed too.

  NKK’s pig also found hundreds of anomalies. The total, now, was nearly a thousand. Investigating them, Alyeska found that nearly three-quarters of the pig’s calls were, indeed, corrosion anomalies. Almost a third would need to be repaired. Corrosion damage was centered in two spots on opposite sides of Atigun Pass: to the north, under the Atigun River, and to the south, under the Chandalar Shelf. By August, Alyeska was weighing replacing nine miles of pipe under the Atigun River. That January, before it could do the big fix, the company dug up the pipe and installed eighty-six more sleeves in that section. That Alyeska did this in winter is proof that the damage was severe; that it ran NKK’s pig again in the summer of 1990 is, too. An Alyeska spokesman told the Reuters news agency, “It is normal that pipes buried in the permafrost, in water and earth, are subject to some form of corrosion.” So much for rustproof. Alyeska’s engineering manager, Bob Howitt, reassured a writer from Popular Mechanics. “These are not the kinds of rust patterns that cause a total rupture,” he said. “They’re the kind that might cause a pinhole leak.” That fall, Alyeska began replacing all nine miles of pipe in the upper Atigun River.

  By then, Alyeska had dealt with the two final stings of 1989. First, the Department of Transportation’s Office of Pipeline Safety told Alyeska that it planned to conduct inspections annually. “What’s surprising to us is the degree of corrosion,” the DOT’s pipeline safety director, George Tenley, said after seeing Alyeska’s corrosion reports. “I don’t think anybody expected to see this much rust this soon.” Then, in December, Alyeska billed Alaska for the cost of the corrosion repairs. That’s the financial arrangement for a pipeline down the middle of the state: Alaska gets oil tariffs minus the operational and maintenance costs of the pipeline that delivers the oil. (Alyeska wanted to raise its delivery rates from $3 per barrel to almost $4 per barrel.) In that sense, the pipeline isn’t just a cash cow, it’s a cash whale. Flowing into state coffers, oil money dwarfs the cumulative contributions of gold, fish, and timber. Alaska, seeing nine digits in red on Alyeska’s prefiled tariff sheet, did not like the size of the operational and maintenance costs. The state said, Wait: replacement? Why aren’t we getting thirty years out of this pipeline? Alaska objected, citing operational “imprudence.” It was the unlikeliest dispute imaginable, like the Israelites denouncing God for providing too little manna. The state sued: Alaska v. Alyeska.

  Alaska’s attorney general, Douglas Baily, was vocal about his displeasure. “They tell us they just got this state-of-the-art surveillance system, and that’s why they found this,” he told a reporter. “But we’re not buying that story. This corrosion didn’t just happen. They’ve had other technology to use.” While Baily was in DC looking into the matter, he heard just the thing. NPR aired an interview with Kevin Garrity, the genial and patient public affairs chairman of NACE. Garrity, who was also a partner at a small firm called CC Technologies, understood corrosion. The AG got in touch, and Garrity was on a plane to Alaska the next day. He brought his colleagues Kurt Lawson and Neil Thompson with him.

  For a few weeks, Garrity, Lawson, and Thompson evaluated whether or not the corrosion damage in Atigun was normal wear and tear. Their initial assessment was negative. It’s worth noting that Garrity’s company usually worked for oil companies. In fact, CC Technologies at the time was doing a lot of work for the American Gas Association and the Pipeline Research Council, which was funded by the same oil companies that owned TAPS. Agreeing to work for the state, he and his partners drew a line in the snow—and their other work evaporated. They became threats to the industry.

  After the men presented their initial findings in Seattle, the Federal Energy Regulatory Commission allowed a full forensic investigation to proceed. For the next three years, Garrity, Lawson, and Thompson, with the help of nine more people, looked at corrosion in TAPS. They checked out the coating, tested the cathodic protection system, examined the passive anodic system, sampled the soils, the water, and the rust they found. Alyeskans, many of them bitter, didn’t make it easy. Supervisors denied them access to dig sites. Others refused to cooperate. Alaska resorted to depositions. Garrity recalled that the situation was “quite contentious,” and Lawson recalled that it was “pretty uncomfortable.” Out along the pipe, they relied on Alyeska for food, shelter, communication—everything. They were, to say the least, vulnerable.

  The corrosion guys formed a hypothesis about the cause of rust in the Atigun River area: sharp crushed limestone (rather than gravel) beneath the pipe was tearing through the high points of spirally wrapped tape, and because the tape happened to be made of polyethylene, current from the cathodic protection (CP) system couldn’t get through it and protect the pipe. Water was then getting to the steel, and doing what water and steel and oxygen do.

  Alyeska had wrapped the pipe in tape because the coating was imperfect and because the federal government told the company it could improvise. The coating, an epoxy called Scotchkote 202, had been introduced in 1965. (It’s made from the same Epon resin as can coatings.) By 1972, it was apparent that it cracked in the cold, and by 1974, it was apparent that it had “disbondment problems.” It didn’t stick. Alyeska, committed to three million pounds of the stuff, sued its manufacturer, 3M, and settled out of court for $24 million. Because Alyeska wasn’t about to scrape all of the stuff off, it asked the Department of Transportation and the Department of the Interior if it could wrap over it with two products, Royston Greenline and Raychem Arcticlad II. The departments okayed the move. “Rather than do the smart thing,” said Lawson, “they Band-Aided it. I think they did what they thought would work, but I don’t think they put a lot of thought into it. The whole concept of taking a failing coating and wrapping it with something else is kind of mind boggling today.”

  Lawson recalled presenting their findings in a weeklong minitrial at the Watergate Hotel, but it might have been any DC conference room. Alyeska blamed saltwater that had started corroding the pipe the minute it was shipped by barge from Japan, years before it was coated, welded, and buried. Alyeska pointed to its world-renowned design, its high-for-the-industry standards, its state-of-the-art pigging program. Yes, Alyeska admitted, construction was rushed at the end. But recall the oil crisis of 1973 and 1974: the rationing, the shortages, the eight-hour lines at gas stations. The country wanted the pipeline in a hurry.

  It took ten minutes for the mediator to make a decision: the state could sue. The mediator declared that to better oversee the pipeline, the various state and federal regulators would combine forces in one Joint Pipeline Office. And the mediator instructed Alyeska to hire CC Technologies for the next two years as technical experts. The corrosion guys went out and got drunk at the Dancing Crab.

  Returning to Alaska, Garrity and crew changed their image. Alyeskans realized that they too wanted to see the pipeline operating for thirty more years. On TAPS, they focused on making the CP system testable; Alyeska put tens of millions of dollars into it. They did some probabilistic modeling, a lot of finite element analysis, and figured out where corrosion would manifest. They compared their predictions with data from pigs, and were spot on. Lawson claimed that Alyeska didn’t need to run smart pigs anymore, but that’s an exaggeration. Said Lawson, “There were repercussions, but the science was good and benefitted the industry.” Their work had integrity.

  CC Technologies survived, grew, and was sold to a giant Norwegian company called DNV. Garrity, who called his work on TAPS some of the best in his career, became the president of NACE. “I always tell people,” he said, “if you had asked us to help you and figure out what’s going on, we would have, but you didn’t ask us. Now at the end of the day, all that researc
h came back, and the pipeline turned out better.” In the meantime, although Alyeska and Alaska signed a cooperative agreement declaring their intent to hunt down and battle corrosion, they both suffered. The Government Accounting Office accused regulators of complacency and Alyeska of making unjustified assurances about the prevalence of corrosion on the pipeline. Then the US Court of Appeals in Washington, DC, dismissed the case, and left Alaska with the bill. The state never got over it.

  MILE 450

  South of the Brooks Range, the pipeline weaves, above ground and below, through remote territory. Down the Dietrich River Valley—past the Chandalar Shelf—it passes under bands of broken yellow limestone cliffs. It passes Nutirwik Creek, Kuyuktuvuk Creek. It passes Sukakpak Mountain, looming like Nez Percé. It goes through gold territory, past Gold Creek, Nugget Creek, Prospect Creek, and Bonanza Creek. Mostly, it’s wild country, with Sheep Creek, Wolf Pup Creek, Cow Creek, Moose Creek, Porcupine Creek, Bear Creek, Grayling Creek. Apparently it’s Bigfoot territory, too. To the pipeline, though, ravens pose a greater threat. Ravens pick at the pipe’s insulation, and then water gets in. Alyeska spent millions installing bands around the seams of the insulation, and the ravens persisted, outsmarting engineers.

  The pipeline wanders past Coldfoot: a frozen truck stop more than town. It crosses an ice road. It crosses the Arctic Circle. It crosses swaths of burned patches, where stubby trees cover the land like obtuse cattails. To the west, the peaks resemble Colorado’s Sawatch, broad and expansive above the tree line. Southern faces harbor poplars. Northern faces harbor only snow. Narrow valleys harbor spruce, and the flat valley bottom harbors red willows.