Fire on the Horizon

by Tom Shroder

  After all, the Horizon had been drilling wells for a decade now, and nothing catastrophic had ever happened.

  When his three weeks were up, Jason headed back to New Orleans and Port Fourchon, Louisiana, for that helicopter ride to the rig. At Macondo, things were starting to get interesting.

  Up to this point, the hole had been passing mostly through sand and rock. The little pockets of gas it had hit along the way were mere soap bubbles compared to where they were headed—a fifty-foot-deep lake of liquid oil and natural gas permeating spongelike sandstone. Squeezed on all sides by billions of tons of rock exerting pressures of around 25,000 pounds per square inch, once pierced, it would explode to the surface as if a giant had stomped on a tube of toothpaste.

  And the whole point of well construction was to build in such a way that the enormous force of upwelling oil and gas could be controlled, neatly shuffled into pipelines and sent harmlessly away to refineries. Job one was to ensure that this pipe they were sinking down, under enormous pressures deep inside the earth, did not leak. That would have been easier if a well were a single seamless cylinder of thick steel from top to bottom. But Macondo, and all deep wells, had to be made of pipes with seams and screw-together ends, and with smaller pipes hanging from larger ones—each connection a prime invitation to a leak. The first defense was to make sure that every individual section of the well was sealed up tight with cement.

  Aboveground—pouring the foundation for a new house, say—cementing sounds simple enough. A plywood form is nailed together outlining the perimeter of the foundation, a cement truck backs up, dangling its cement chute over the mold, and pours down liquid cement until the form is filled. When the cement is fully cured, workers break off the plywood and the foundation is complete. What the homeowner does not see is more complicated. The humidity, ground temperature, and location all determine the amount and types of material—rocks, chemicals, silica—that go into the slurry to ensure the right drying time and strength of the hardened product.

  Cementing a well is even more complex and much more difficult. You can’t just pour cement in at the top and let gravity pull it down. You first have to get the correct mixture of chemical additives figured out, then find a way to pump the cement so that it flows out the bottom of the pipe and is forced up the thin annulus space between the pipe and the well wall, which is where the seal needs to be. As you pump the cement, it has to be untouched by any contaminants—like drilling mud or seawater, one or the other of which fills the hole at all times. Any mixing of those things with the cement will destroy its chemical integrity and render the cement job worthless. And you need to be certain, without being able to see it, that the cement went exactly where you meant it to go, and rose to an exact height in the annulus. To recap: The goal is to pump cement into spaces a couple of inches wide and completely submerged in water or mud without letting the water or mud contaminate the cement, using only tools that can be lowered through a narrow opening and operated from thousands of feet above.

  It sounded like a trick even David Blaine wouldn’t dare attempt. But the rig had its own magicians. They had nothing up their sleeves, but they did put a “shoe” on the bottom of the casing. The shoe had rounded edges that helped guide the casing as it was lowered to the bottom without scraping the sides of the hole or getting caught on ledges. A high-pressure nozzle that could pump mud or cement was lowered down into the casing pipe. To make sure that the pipe was clean and free of debris, mud was pumped through the pipe and out ports in the shoe. It hit the bare bottom of the hole, which forced it back up. The shoe, and the high-pressure flow coming out the holes in the shoe, prevented the mud from flowing back up the pipe. It had nowhere else to go but up the space between the outside of the casing and the bare wall of the hole until it came back out the top, carrying any debris with it.

  After the mud had been circulated through the pipe, cleaning it out, the rig measured and pumped something called spacer, a water base with additives to make it friendly to cement. You could count on rig terminology to tell it plain: Spacer was there to put a space between the mud now filling the well and the cement soon to follow. It pushed the oil-based mud ahead of it up the annulus, readying the space for cement. A device called a plug was inserted into the pipe. The plug was like a cork within the tube. It had two parts, a top half and a bottom half, each of which had polymer edges that made a tight seal against the interior of the pipe.

  When cement pumping began, an instrument called a “plug dart” was shot down a tube and hit a trigger in the top of the cement plug that unlocked the bottom half of the plug. The force of the cement flow through an opening in the top half of the plug pushed the bottom plug down the pipe. The plug in turn pushed the spacer ahead of it, out the holes in the shoe at the bottom of the pipe, where it made the U-turn at the well bottom and pushed up the annulus.

  Back in Houston, engineers had run a computer model of the cement job, which had calculated the exact amount and chemical composition of cement and forcing pressure needed to ensure that the annulus was filled to the desired height and density. The cementer on the rig only needed to make sure the right amounts of additives were mixed in at the right times and count the number of pump strokes it took to push that amount of cement down the hole. When that number was reached on the stroke counter, another dart was sent down the pipe. This one sealed the hole in the top plug and released it to move down the pipe behind the full load of cement. Now they began to pump mud again. The high-pressure flow of the mud forced the top plug down behind the bottom plug, with the premeasured amount of still-uncontaminated cement between them. It was like a cement motorcade with armed guards in the front and the back.

  As the bottom plug neared the shoe, it hit a barrier, called a float collar. The motorcade slammed to a halt. But as the pumping continued, the pressure increased to the point where a rubber membrane in the bottom plug ruptured, allowing the cement to flow through a valve that opened like a trapdoor in the float collar. The mud pressure continued to press down on the top plug, which pushed the remainder of the cement out the trapdoor and through the holes in the shoe. Just like the mud and spacer before it, the cement made the U-turn from the bottom and up the annulus, displacing the spacer.

  Finally the top plug bumped down on the bottom plug, both snug against the float collar, and the pumping stopped. Gravity pulled at the cement now in the annulus, creating a slight backflow through the holes in the shoe, reclosing the trapdoor in the float collar. By measuring the “returns,” the amount of mud that got pushed out of the well by the inflow of cement, the cementing team could see that it was the full amount, and therefore could assume that all the cement had been pushed where it was needed. Now all they had to do was wait on the cement to harden. Even that had its own acronym: WOC, for Waiting on Cement. It was no joke. If they didn’t wait long enough before they began to work the well again, the cement job, however perfect when it went in, could be ruined by stress before it had properly hardened. They just had to sit and wait, not the easiest thing to do on a rig, especially one so badly behind schedule.

  CHAPTER TWELVE

  A LONG STRING

  Early April 2010

  Macondo Prospect

  As he drove to the airport to catch the flight to New Orleans, on his way back to the Horizon for the last time, Jason Anderson looked over at his wife and asked, If something happened to me, would you and the kids stay in the house?

  He’d been in a strange mood these past few weeks at home in Midfield. You could have called it morbid. Shelley didn’t get it. He should have been celebrating—his promotion had finally come through. He was going to be senior toolpusher, the person who directed the entire drilling operation on the Discoverer Spirit. His only superior on the drill floor would be the OIM. In fact, it was common for senior toolpushers to become an OIM. But instead of enjoying the news, he’d been worrying about things on the order of: Did Shelley know how to drain the RV water system and get it winterized? Would the kids
learn the right way to handle a hunting rifle if he weren’t around to teach them?

  After years of putting it off, he sat down with Shelley and drew up a will.

  Shelley had lots of experience running the house and tending to the kids for long stretches without a husband’s support. She was used to it, still proud of the work Jason did. This time was tougher, though, thinking about all the things he’d been saying, and the fact that he shouldn’t have been leaving in the first place.

  The plan had been for Jason to spend time with Shelley and the kids before moving on to the Spirit. But the top Horizon managers had called to say they needed him on the Horizon, one last time, to help ensure that the new drill crew leadership was ready to take control. It was no surprise. Only three people were stationed in the driller’s shack, the toolpusher, driller, and assistant driller, and like the Three Musketeers, the team would have to merge into a flawless unit when the well barked. His trinity was already missing a man. Barney Ray, the assistant driller, had been promoted to driller aboard the new semi-submersible Champion just the previous hitch. Jason had supported Ray’s promotion and gave his thumbs-up to the selection of Roy Wyatt Kemp, a derrickman, as his replacement. Jason was confident in his own ability to watch over and mentor the new recruit. But the promotions, like dominos, meant the rest of the drill crew would also be promoted to fill the gap…the entire chain moved into jobs above their level of previous experience. This was not a time to replace the leader, especially on this job.

  His previous hitch had been the worst yet. The new section of the well proved even more troublesome than the first. They had it listed on the well plan at six to ten days and three thousand feet. Instead, it took nineteen days and they got only half the depth. It had been a vicious cycle: feeling like they had to hurry, then pushing too hard and blowing holes in the well. They’d been putting so much sticky well patching material in there to stop the mud loss, that the drill pile got stuck. They spent forever, using every trick they knew, but they couldn’t fish it out. The well services contractor Schlumberger came out to run sophisticated instruments down the drill pipe to determine where exactly the pipe was hung up. The measuring instruments got stuck, too, and they weren’t cheap. All of it was costing BP, big time. They ended up having to do a sidetrack—to cut the pipe above the obstruction and drill a new path around it.

  Through it all, Jason had felt uncomfortable with the pressure: the pressure readings down the hole and the pressure from the company men and from the beach to hurry up and get it done. By the end of his hitch, the date had come and gone when the Horizon had been scheduled to start another well. They were nowhere near the end of Macondo and had already overstayed their welcome.

  The final section of the well gave neither Jason nor the Horizon any relief. The geology here was maddeningly inconsistent, and the mud weight needed to contain pressure in some areas was too much for others, pushing out the well wall and causing mud losses all the way down to the pay zone. They had to keep stopping to run in the well patching material, then going forward only to start losing mud again when the well walls gave way.

  The primary objective was a 123-foot layer of sandstone, the top 53 feet of which was filled with hydrocarbons. Sandstone is most spectacularly familiar as the rock in the walls of the Grand Canyon. It looks solid enough, but in fact it is filled with microscopic pores. Under the high pressures of the deep earth, oil and liquefied natural gas can saturate it like a sponge.

  They hit the primary oil deposit where the geologists had told them it would be. They knew they were there when the mud coming back up the riser and through the mud processing equipment began to be saturated with gas. But immediately below the deposit, as they began to drill toward the secondary objective, the mud almost stopped coming back at all. It was as if the bottom of the well had just dropped out—taking about three thousand barrels of mud with it. BP was now literally pouring money, about $1 million, into a bottomless pit. Even more money was tossed on top of what had already vanished, in the form of the well patching “lost circulation material,” to seal the hole and stop the losses.

  BP execs had had enough. Macondo had beaten them into an early retreat and it became clear that steel and cement were the only remedy. On April 9, they chose to stop drilling at 18,360 feet beneath the rig, or 13,293 feet below the wellhead. This was just far enough below the main deposit to allow them to do the cementing they needed to complete the well, but it was more than a thousand feet short of the secondary objective, another possible pay zone the Horizon had been meant to explore. It was no longer worth it. Everyone just wanted to get the hell out of Macondo.

  All along, the primary purpose of this well had never been to begin siphoning the oil and shipping it to a refinery. From the start, the Macondo well was meant to determine just how lucrative the oil deposit would be. There had never really been any question of finding no oil. The seismographic studies had been unambiguous. The oil was there. The question was, how much, and more important, how quickly could it be extracted. Would the oil truly be worth the exceedingly high cost of pumping it out of the ground and into an oil refinery?

  Building a major production platform to suck Macondo dry over a period of years—a rig that would be owned, crewed, and operated by BP—would cost a multiple of what R&B Falcon had spent on the Horizon. The BP flagship production unit, Thunder Horse PDQ, was built in 2004 for $5 billion. It was a semi-submersible like the Horizon, only with about 10,000 square feet more deck area and permanently moored rather than dynamically positioned. The design phase alone for a new platform would be six to twelve months, followed by another year or two to build it. The better part of yet another year would be required to ship it from Korea and install it in the Gulf with permanent moorings and connect it to the pipelines that snake all over the bottom of the Gulf, linking offshore wells directly to the shore-based refineries. The end product would be a multibillion-dollar octopus floating on the surface, its tentacles stretched into the mud deep below.

  Bottom line: Macondo sat atop an oil field with real commercial potential, but after rushing so desperately and enduring so much adversity to complete the well, it would be cemented top and bottom, then abandoned for years before it would produce any oil.

  The goal in completing the well was to make sure that it was (1) absolutely secure sitting untended for years, and (2) able to be transformed into an efficiently producing well, one that could be put on line fast and with minimum expense and trouble when the time came.

  Like everything about this well, the completion promised to be difficult. The final section of the well was the portion most exposed to oil and gas deposits. So the fact that significant weakness in the formation had been encountered right at the bottom, causing 3,000 barrels of mud to escape into fractured rock, did not augur well.

  BP’s engineers were considering two plans to deal with the situation. One involved a two-step process. First, they would hang another stretch of pipe that would run from a hanger built into the end of the casing already in the well to the bottom of the hole. When that was cemented in place, they would insert a steel tube connecting the top of the hanger back up to the wellhead. This was called a “liner tieback” and it was, finally, the tube through which the oil would actually flow when the well was producing.

  The other option involved running a single string of steel casing from the bottom of the hole all the way to the wellhead—which was called a “long string.”

  The long string was BP’s preference from the beginning. It was the simpler option. And less expensive. Back in March, a BP engineer named Brian Morel noted in an interoffice e-mail that the long casing string “saves a lot of time…at least 3 days.” He followed that up with another e-mail to Sarah Dobbs, the BP completions engineer, and Mark Hafle, another BP drilling engineer, that “not running the tieback…saves a good deal of time/money.”

  That was important. The well was already more than $50 million over budget and it was the engineers’ responsibili
ty to find ways to stop the bleeding. They had a personal stake, too. BP employees were graded annually on how much money they save the company, under the evaluation category “Every Dollar Counts and Simplification.” To prove their case, they compiled itemized lists of all the changes they instigated that reduced BP’s costs.

  This created a strong incentive to find the cheaper option, and the long string was definitely cheaper.

  But it would also be riskier. With the single pipe going all the way to the top, any problems with the cement job at the bottom of the well would create an open path for oil and natural gas all the way to the wellhead, leaving the seal assembly at the top of the well—never designed to resist heavy pressure—as the only barrier, like a 160-pound placekicker left all alone to stop a 200-pound opponent from returning a kickoff for a touchdown.

  The two-part option would create four barriers to gas flow. The first and most formidable barrier would be the cement job sealing off the bottom section of the well from the oil and gas deposit. If that failed, rising oil and gas would still have to break through a seal installed at the top of the final string of casing, on the hanger where it attached to the existing casing in the well. And that seal would be backstopped by yet another cement job right on top of it, where the “tieback casing” that ran all the way up the hole would be cemented at its base. Finally, there would be the relatively weaker seal at the wellhead.

  Without those two extra barriers, only the cement at the bottom would stand between millions of gallons of oil and gas in the deposit and the vulnerable seals at the wellhead. Everything would be riding on achieving a quality cement job.

  On the morning of April 15, that fact was making Jesse Gagliano sweat.

  Gagliano was thirty-nine, a sharp-faced man with close-cropped brown hair that was graying at the temples. He was a technical sales advisor for the cement contractor Halliburton, but he had worked out of the BP offices for the four years in which he’d been assigned to assist BP with Deepwater Horizon projects. He’d been working on Macondo from the beginning, lending his expertise to the BP well team, designing the cement jobs and running tests and simulations to ensure that they would be successful.