When the smart pig passed by, it also stopped a clock on a toaster-oven-sized apparatus that picked up the signal from the pig’s transmitters. This second, more exact method worked this time, but it doesn’t always, even if the transmitter is the Alaska model. For some reason, it’s less reliable than a pair of ears, and because the men didn’t understand exactly how it worked, they didn’t trust it. In any case, when the pig passed by, the men radioed Alyeska’s Operations Control Center (OCC), relaying their identities, location, and the time the pig passed.
Listening carefully to those calls not far south was a young engineer named Ben Wasson. Neogi’s lieutenant, Wasson was coordinating the pig run out in the field. (Neogi had since flown home to Anchorage.) Camped out in the cafeteria at Pump Station 3, Wasson was drinking coffee from a plastic cup and carrying around two handheld radios, two cell phones, and a logbook. During every call, he jotted down notes. He noted when another crew, twenty miles south of the first, reported hearing the scraper pig go by, and he noted when another crew, farther south still, reported having raised valves along the line, so that both pigs could proceed unobstructed. From these notes, Wasson compiled daily tracking updates, and these he emailed to eighty Alyeska executives, including Neogi. Attached to each update was a spreadsheet called “Pig Passage Summary.” For all those antsy about the pig getting stuck in the pipeline, this document calmed their nerves. It demonstrated that the pig, surveilled every few miles, wasn’t hung up or lost but on schedule.
Wasson was wearing Carhartt pants, a gray flannel plaid shirt, a black fleece jacket, and a green cap. A Mainer, from Bar Harbor, he’d shaved his lumberjack beard seven years ago, leaving a pair of rosy and windswept round cheeks. At the Formica table where he sat, he was joined by a grizzly bearded man who resembled a weary sailor. This was Dave Brown, the superintendent of the three field crews. Wearing a tan baseball hat, and resting his glasses on the second button of his blue T-shirt, he looked more the Mainer than Wasson. He sat down and began eating a hamburger. A NASCAR race unfolded on a TV nearby.
Brown has been on every one of Alyeska’s pig runs since 1995, and on the first ten runs, after tracking pigs in something like 1,500 places, he only missed one. Hence his nickname: Super Dave. He later told me, “My crews don’t miss pigs.” It was his job to deal with any snags that came up. Earlier that morning, a snowcat en route to raise a valve at mile 81 broke through the ice of the Sagavanirktok River. Before it started sinking, the driver put it in reverse. He called Brown, who called Wasson, who called in a helicopter. Wasson had two helicopters at his disposal, as well as a front-end loader, two snowmobiles, and four satellite phones. When the radio system went down earlier, these proved handy. Neither Wasson nor Brown, though, had a spare snowcat. With one broken down, crews hadn’t been able to track the scraper pig between mile 18 and mile 40. Had it gotten stuck, the smart pig would have slammed into it. Things worked out, but Brown remained unhappy. He didn’t like missing pigs. After three days of near nonstop travel, he was also tired. After lunch, he took a nap.
With Neogi away, Wasson became the de facto guru of the pig run, its natural leader, choreographing the layers of logistics while drinking a lot of coffee. This pig run, in 2013, was his first since he started at Alyeska in 2010. He called the smart pig the ILI pig, for in-line inspection pig. More calls came in over the radio. Thanks to the tracking of Super Dave’s crews, Wasson learned that the twelve-hour gap between the scraper pig and the smart pig was down to ten and a half hours. Wasson relayed the information to Neogi, who decided to delay the smart pig at Pump Station 3 for three hours.
Wasson is an engineer, but not the mechanical or hydraulic type. He’s a civil engineer, and also a surveyor. He’s a map guy, and thinks like one. The first date in his logbook was July 2, 2012—nine months earlier. One hundred and five pages bore notes on the pig run, all written in pencil. Wasson tapped the book, and said, “This won’t break down. I can go without everything else.” That was a surveyor talking. Thirty-seven years old, Wasson could have talked until he turned forty about surveying methods, precision, corrections. In a long discourse on digital mapping and Geographic Information Systems (GIS) Wasson brought up the 2010 San Bruno, California, pipeline explosion. Pacific Gas and Electric, he said, knew that it had a deep pit, but thought the pit was on a thicker piece of pipe. “They had no integrity in their integrity management,” he said.
Contrary to received wisdom on the lines, pig is not an acronym. It does not stand for pipeline inspection gauge. The Texas roughnecks who came up with the term in the early twentieth century did so after shoving bundles of barbed wire and straw through their lines. Apparently these bristly messes, which emerged covered in muck, brought swine to mind. Plus they screeched like pigs. Before Texans called them pigs, others in Pennsylvania called them go-devils, moles, rabbits, and spears. Further proof that pipeline inspection gauge is a crafty modern backronym lies in the term itself. Objects shoved through pipelines didn’t do any inspecting until halfway through the century. Until then, they just scraped pipelines clean.
Almost any object can serve as a cleaning pig. Balled-up canvas rags and bundles of leather initially served the purpose. After a four-inch gas line in Montana was buried in a rock slide, an operator, in 1904, pumped a rubber ball through. One thrifty pipeline operator used chunks of foam mattresses. A jam manufacturer pigged his lines with loaves of bread. A paint manufacturer used plastic coffee cups. Tools engineered specifically for the purpose, more or less resembling encumbered plungers, have been available since 1892.
Some engineers have argued that the first smart pig was the first pig that got stuck in a pipe, but this seems like bestowing Phi Beta Kappa status on a preschooler. A failed cleaning does not an inspection make. The first pig to get stuck revealed that there was a problem but told nothing of the problem. If anything, it exacerbated it. On the other hand, unintelligent pigs did yield valuable information: some, made with soft aluminum discs or notched sheet metal, revealed evidence of encounters with dents, though not details about their locations. Tracking such pigs, even with chains or whistling noisemakers fastened to them, was difficult at best. Finding a stuck one was nearly impossible.
Then caliper pigs were devised. The first one had two fixed arms, sized differently, like a lobster, and a third articulated arm. Its motion, recorded on a chart, revealed where dents and dings were located. But it could go only so far, because it was powered by an electric cord trailing behind it. Not until 1955 was a battery-powered caliper pig designed. Not until 1959 did one really work.
In the meantime, much more capable pigs were conceived. Between 1953 and 1959, the field burst to life as nearly every type of modern pig was imagined. Four oil companies filed patents for pigs that could infer the thickness of a pipe based on magnetic flux leakage. If credit is due anyone, it is Howard EnDean, of Gulf Oil’s Pittsburgh research lab, who filed four pig-related patents on one day in the summer of 1956. In addition to scheming up a leak-detecting pig based on sound and pressure, EnDean designed an electrical-potential-gradient pig that would “ascertain the points where excessive corrosion is likely to occur.” The only pig not patented during this tiny window was an ultrasonic thickness pig, the technology for which was a dozen years out of sight. Nevertheless, EnDean—the grandson of an old Pennsylvania oil driller—was a visionary. He was obsessed with the insides of pipes. Before he died in 1996, he expressed his wish to have his ashes poured into an oil well so that he could finally see what happened at the bottom.
Though the principles of his smart pigs made sense in theory, in practice they didn’t work. Shell built the first MFL pig and soon declared it a failure. The tool was insensitive and could detect only corrosion on the bottom 90 degrees of a pipe. A subsequent pig developed by a company called Tuboscope could read all 360 degrees, but it could detect only major flaws, couldn’t tell if they were internal or external, and it didn’t have an odometer, so locating the pig’s discoveries was a challenge. An MF
L pig developed by Pipetronix in 1972 performed no better.
The major challenge in pig design lay in data storage. When British Gas, in the early 1970s, tested the MFL pigs on the market, it concluded that none was good enough and began developing its own. Storing the data, someone at British Gas said, was like “reading the Bible every six seconds.” In an age of magnetic tape and paper charts, taking billions of measurements wasn’t possible. As a result, early pigs were limited geographically, as if on parole, by their recorders. The Pipetronix pig could run only thirty miles, by which time its twelve-channel recorder had produced an inspection log a thousand feet long.
Pig engineers responded by designing pigs that collected different information. Instead of measuring metal thickness, they hunted for leaks. The best leak-detecting pig design relied on the detection of the inaudible, high-frequency squeal made by a liquid or gas escaping through a small hole in a pipe. Today this technique works marvelously. A generation ago, it did not. The sound of a pig traveling through a line was just too similar to the sound of a leak in that line.4 And still, storing the data was a challenge. Shell made such a pig that could run for no more than four days. After that, it ran out of room to print. Sophisticated for its day, the recorder printed a seven-digit code every couple of seconds. The first digit indicated low or high pressure. The second digit indicated a leak, a marker, or no sound. The next five digits indicated the time, in hours, minutes, and seconds. To avoid duplication, the recorder printed out only the first two digits, up to nine times, until something changed. The result was a long, skinny ticker tape that looked like this:
1L24392 1L 1L 1L 1T24515 1T
Into the 1990s, TAPS was considered a “staggering inspection assignment” on account of its scale. In data storage and analysis, only recently have computers caught up. Modern pig data—color coded and laid out horizontally, as if the pipe had been split open—can be deciphered more easily. On TAPS, it’s converted to a list of hundreds of thousands of calls and then winnowed to a short list, a few pages long, of anomalies that demand attention.
The second problem with early smart pigs was battery power. One had only enough juice to run for twenty hours. To save power, some tried to make pigs with generators that ran off wheels that spun as the pig traveled along, but the only way to keep the wheels from slipping was to give them teeth, which presented a catch-22. Designers tried to put a turbine in the center of a pig, so that fluid passing through would power it, but dirt, wax, and dust clogged it up every time. In an attempt to save power, more than one company designed a camera pig for use in gas lines. One company used a Hasselblad camera, the same as NASA had used on the moon landings. Its lens captured the bottom 60 degrees of the pipe. The film was Kodak. The camera had no shutter, since a pipeline interior is blacker than space. Every forty feet, a strobe fired, and the camera got a shot. When the film was developed and printed, analysts used trigonometry to determine the scope of corrosion problems.
Getting a photo on the moon was easier than getting one in a pipeline. Dust obscured everything. One such pig, trailing behind a caliper pig, took photos whenever the caliper pig detected a dent. The good news was that it took surprisingly crisp photos. The bad news was that when the pig hit a dent, it shook so forcefully that the trailing camera overreacted and, according to two pipeline historians, took “excellent close-ups of the opposite but undented pipe wall.”
An ultrasonic smart pig—which measured the thickness of pipe walls by listening to echoes—wasn’t patented until 1971. It took many companies many years to develop this ultimate of measurement technologies. In early lab tests, ultrasonic pigs were very accurate, finding voids, inclusions, pits, and laminations. The hitch was getting them to work on dirty, rough pipes at twenty miles per hour. The first challenge was holding the transmitters and transducers firmly against the pipe wall in a “known, consistent relationship.” The second was perfecting digital circuitry capable of measuring time in millionths of a second. This put ultrasonic pigs even with other varieties of smart pigs, whose history of runs in 1977 was, in the words of an engineer, “mixed.” It was in that era that the Alaska pipeline began operating. As a result, Alyeska endured a good bit of pigging heartache.
Neogi chose Baker Hughes’s tool, the Gemini, by first selecting the “threat categories”—the detection capabilities—of the pig he wanted to run. He wanted an MFL pig that could examine internal and external corrosion, particularly between the pipe and the seventy thousand clamps that hold the aboveground pipe to big stilts. The Trans-Alaska Pipeline traverses a great deal of permafrost, so about half of the pipe is elevated on H-shaped stilts. If the warm pipe were buried in these areas, the permafrost would melt, allowing the pipe to sink, bend, crack, and leak. The last time those regions had been examined was in 2001, back when Alyeska ran an ultrasonic pig. Neogi also wanted a pig that could detect dents and settlement. (Until 2009, Alyeska collected corrosion data and curvature/deformation data from two different pigs, in two different runs.)
From six bids, Neogi selected one from Baker Hughes, but until its pig was tested, or as Neogi put it, verified, the deal wasn’t done. To verify the pig, Alyeska had shipped five pieces of pipe, laden with manufactured anomalies, and two clamps, to Calgary, Alberta. There, after the pipes were welded together in the gravel lot behind its building, Baker Hughes pulled its pig through. They call this a pull test, and according to Devin Gibbs, a serious thirty-five-year-old Baker Hughes employee with much pigging experience, it was a bitch. The pig weighed so much and had so much drag that the motor pulling it was maxed out, and the pig didn’t want to budge. The thing was ornery then, too. But the pig found all the defects, Baker Hughes got the contract (a $2 million lease), and Alyeska was satisfied. Better: Neogi called this pig “a dream, a holy grail.”
Still, the pig would be further calibrated. Alyeska has 150 intentional defects in its pipe at various pump stations. Once the pig emerged in Valdez with all of its data, analysts would use these to fine-tune the pig’s defect-sensing algorithm.
In spite of Neogi’s obsessing, there was one trait he could not guarantee in the pig of his dreams: he could not certify that this pig wouldn’t get stuck.
In pipelines around the world, pigs get jammed in offtakes, wedged in valves that aren’t entirely open, stalled at branches or wyes, constricted by debris, stuck nose down, trapped in reducers, pinned in too-tight bends, or—as on TAPS—sucked into drain lines. If one pig pushes on another pig in front of it, the added pressure can translate into increased outward force on the front pig’s seals, stopping it like a cork. If the seals of a pig wear down or buckle, and product flows around it, the pig may stall in place, like a kayaker in an eddy.
A stuck pig may be jammed so tightly that the only way to remove it is by burning it away. Or a pig may get wedged so tightly in a sharp bend that, as happened in British Columbia, its batteries explode. A stuck pig has forced Canada’s Keystone pipeline to shut down. In a subsea pipeline, a stuck pig is a particularly big deal. Runner-up is a stuck pig in the Arctic in winter.
Often what transpires between pig and pipeline is mysterious. Pigs pop out inside out. Cylindrical pigs pop out spherical. Some pop out backward. A thirty-six-inch foam pig has arrived inside another. Four thirty-six-inch steel pigs—each four feet long—got stuck in a train wreck and emerged 75 percent shortened. Most often, pigs pop out in shards and get swept up with a broom. Sometimes only the front half pops out, or nothing comes out at all. A smart pig has gotten stuck for nine months, and then all of a sudden “woken up” and finished its run. Conoco launched a six-inch pig on a thirty-mile pipeline in 1972. It popped out in 1996.
A pig can do worse than get stuck, though. Propelled by gas, a pig can easily rupture a pipeline. Behind compressed air, a pig has reached 170 miles per hour. A pig traveling at that speed, or a fraction of that speed, will refuse to take sharp turns. It will shoot straight through a pipe. A pig has created its own exit in a twenty-inch pipeline. Another has created
its own exit at the bottom of a long downhill stretch. In late 1995, after a smart pig undergoing tests got stuck in Battelle’s Ohio test loop, representatives of the pig manufacturer, thinking themselves wise, inserted a pusher pig behind the stuck pig, to force it out. This result was not achieved. Instead, the pig heated up, ignited a fire (the line held only air), and the test loop exploded. Both pigs were destroyed. Battelle—a nonprofit research company—later sued the pig manufacturer.
More than a couple of pigs have created their own exits at the end of the ride. A six-inch pig went through the door of a receiving trap and into a chain-link fence fifty yards away. A forty-two-inch pig came in with such velocity that it knocked a three-thousand-pound door three hundred yards into a car. A forty-eight-inch pig broke through a steel grate, slammed through a shed behind it, and landed in a pile of lumber. The first witness to the scene said it looked like a tornado had hit. A foam pig projected out of its receiver has cracked a brick wall thirty yards away. Another has cleared eight hundred yards. One made it half that distance, crashed through a wall, and landed in a bedroom in Texas. The homeowner said the room “looked like a war zone.”
Rust: The Longest War Page 26
