A Sea in Flames Read online

  Right from the start—beginning with Hurricane Ida forcing the Marianas rig off the well location—various things didn’t proceed as planned, or struck people as risky.

  The Deepwater Horizon, built at a cost of $350 million, was new in February 2001. In September 2009, it had drilled the deepest oil well in history—over 35,000 feet deep—in the Gulf of Mexico’s Tiber Field.

  It was a world-class rig, but it was almost ten years old. The wonderful high-tech gadgets that were state of the art in 2001 did not always function as well in 2010. Equipment was getting dated. Old parts didn’t always work with new innovations. Manufacturers changed product lines. Sometimes they had to find a different company to make a part from scratch.

  The world has changed a lot since the rig was built. So has software. More 3-D, a lot more graphics. Drillers sit in a small room and use computer screens to watch key indicators. Depth of the bit, pressure on the pipe, flows in, flows out. But on this job, the software repeatedly hit glitches. Computers froze. Data didn’t update. Sometimes workers got what they called the “blue screen of death.” In March and April 2010, audits by maritime risk managers Lloyd’s Register Group identified more than two dozen components and systems on the rig in “bad” or “poor” condition, and found some workers dismayed about safety practices and fearing reprisals if they reported mistakes.

  Risk is part of life. And it’s part of drilling. Yet drilling culture has changed, with much greater emphasis on safety than in the past. Many people still working, however, came up the ranks in a risk-prone, cowboy “oil patch” culture. A friend of mine who worked the Gulf of Mexico oil field in the 1970s says, “It was clear to me that I was way underqualified for what I was doing. Safety didn’t get you promoted. They wanted speed. If we filled a supply boat with five thousand gallons of diesel fuel in twenty-five minutes, they’d rather you disconnect in a big hurry and spill fifty gallons across the deck than take an extra three minutes to do it safe and clean. I’d actually get yelled at for stuff like that. Another thing that was clear: if you could simply read or write, you could pretty much run the show. They actually gave oral exams to workers who couldn’t read. I was still a kid, but pretty soon I was put in charge of a supply boat because I could read and write. That was the culture then.” Another friend, now a tug captain, says, “Never in the four years I worked the rig did I hear anyone say, ‘Let’s wait for better sea conditions.’ We were always dragged into situations we didn’t want to be in, doing things I didn’t think were safe. Now it’s a lot better. It used to be the Wild West out there.”

  When you pump drilling fluid down the well, it comes out the bottom of the drill pipe and circulates up between the drill pipe and the wall of the well, and comes back to you. For every barrel of drilling fluid you push down, you’d better get a barrel back. If you get more—that’s really bad, because gas and oil are coming up in your fluid. If you get less—that’s really bad, too. Drillers call it “lost returns.” It means the returning fluid has lost some of its volume because fluid is leaking into the rock and sand of the well’s walls, sometimes badly. Sometimes there are fractures in the rock and the fluid’s going there. When it’s leaking like that, you can’t maintain the right pressure in the well to tamp down the pressure of oil and gas that wants to come up from below.

  In a March 2010 incident, the rig lost all of its drilling fluid, over 3,000 barrels, through leaks into the surrounding rock and sand formation into which they were drilling.

  BP’s onshore supervisor for this project, John Guide, later testified, “We got to a depth of 18,260 feet, and all of a sudden we just lost complete returns.”

  BP’s senior design engineer, Mark Hafle, was questioned on this point:

  Q: “Now, lost returns, what does that mean in plain everyday English?”

  Hafle: “While drilling that hole section we lost over 3,000 barrels of mud.”

  Three thousand barrels is a lot of barrels. At over $250 per barrel for synthetic oil-based mud, that’s $750,000.

  A high-risk pregnancy is one running a higher than normal risk for complications. A woman with a high-risk pregnancy needs closer monitoring, more visits with her primary health-care provider, and more careful tests to monitor the situation. If BP can be called the birth parent, this well was a high-risk pregnancy.

  Several times, the well slapped back with hazardous gas belches called “kicks,” another indication that the deep pool of hydrocarbons did not appreciate being roused from its long sleep.

  At around 12,000 feet, the drill bit got stuck in rock. The crew was forced to cut the pipe, abandon the high-tech bit, and perform a time-consuming and costly sidetrack procedure around it to continue with the well. The delays cost a week and led to a budget add-on of $27 million.

  The work had fallen forty-three days behind schedule, at roughly $1 million a day in costs. At a “safety meeting,” the crew was informed that they’d lost about $25 million in hardware and drilling fluid. Not really safety information. More pressure to hurry.

  High-risk pregnancy, added complications. On April 9, 2010, BP had finished drilling the last section of the well. The final section of the well bore extended to a depth of 18,360 feet below sea level, which was 1,192 feet below the casing that had previously been inserted into the well.

  At this point, BP had to implement an important well-design decision: how to secure the final 1,200 or so feet and, for eventual extraction of the petroleum, what kind of “production casing” workers would run inside the protective casing already in the well. One option involved hanging a steel tube called a “liner” from the bottom of the previous casing already in the well. The other option involved running one long string of steel casing from the seafloor all the way down to the bottom of the well. The single long string design would save both time (about three days) and money.

  BP chose the long string. A BP document called the long string the “best economic case.” And though officials insist that money was not a factor in their decisions, doing it differently would have cost $7 to $10 million more.

  BP’s David Sims later testified, “Cost is a factor in a lot of decisions but it is never put before safety. It’s not a deciding factor.”

  Sims was John Guide’s supervisor. Guide described the long string design as “a win-win situation,” adding that “it happened to be a good economic decision as well.”

  Guide insisted that none of these decisions were done for money.

  Q: “With every decision, didn’t BP reduce the cost of the project?”

  Guide: “All the decisions were based on long-term well-bore integrity.”

  Q: “I asked you about the cost of the project. Didn’t each of these decisions reduce the cost, to BP, of this project?”

  Guide: “Cost was not a factor.”

  Q: “I didn’t ask if it was a factor. I asked if it reduced the cost. It’s a fact question, sir. Did it not reduce the cost, in each case?”

  Guide: “All I was concerned about was long-term well-bore integrity.”

  Q: “I just want to know if doing all these decisions saved this company money.”

  Guide: “No, it did not.”

  Q: “All right; what didn’t save you money?”

  Silence.

  Q: “Which of these decisions that you made drove up the cost of the project, as opposed to saving BP money? Can you think of any?”

  Guide: “I’ve already answered the question.”

  Q: “What was the answer?”

  Guide: “These decisions were not based on saving BP money. They were based on long-term well-bore integrity.”

  Some people called the long string design the riskier of two options. Greg McCormack, director of the University of Texas at Austin’s Petroleum Extension Service, calls it “without a doubt a riskier way to go.”

  But others disagree. Each of the two possible well casing designs represented certain risk trade-offs. One called for cement around casing sections at various well depths, providin
g barriers to any oil flowing up in the space between the rock and casing. The other called for casing sections seamlessly connected from top to bottom with no outside barriers except the considerable bottom cement. Investigators would later focus lasers on this aspect of the well design for weeks after the well blew. The cost savings led many to believe that this was a cut corner that resulted in the blowout. Months later, however, it became clear that this decision was not a direct cause of the disaster.

  Final hours. In the eternal darkness of the deep sea, the well is dug, finished. All that’s needed: just seal the well and disconnect. The plan was for a different rig to come at some later date and pump the oil for sale.

  At BP’s onshore Houston office, John Guide is BP’s overall project manager for this well. He has been with BP for ten years, has overseen more than two dozen wells. Mark Hafle is BP’s senior design engineer. With twenty-three years at BP, Hafle created much of the design for this well. Brian Morel, a BP design engineer involved in many of the key meetings and procedures in the final days, splits his time between Houston and the rig. Out on the drilling rig itself, BP’s supervisors, titled “well site leaders,” are often called the “company men.” They oversee the contractors. Because a drilling rig operates twenty-four hours a day, BP has two well site leaders aboard, working twelve-hour shifts: Don Vidrine and Bob Kaluza. Vidrine is in his sixties. Kaluza in his fifties. Vidrine has been with the rig for a while. Kaluza is new.

  Because the Deepwater Horizon was both a drilling rig and a vessel, rig owner Transocean has two separate leadership roles. When moving, the rig is under the authority of the captain; when stationary at the well site, an offshore installation manager, or OIM, is in charge. Jimmy Harrell, OIM, managed the drilling. He’d been with Transocean since 1979 and on the Deepwater Horizon since 2003. Curt Kuchta was the Deepwater Horizon’s captain.

  Managers play an important part in the decision process, but the drilling team executes the plan. At the top of the drilling personnel chart are the “tool pusher,” who oversees all parts of the drilling process, and the driller, who sits in a high-tech, glass-paneled control room called the “driller’s shack” and leads the actual work. Many people work under the direction of the tool pusher and driller.

  On duty on the evening of April 20 were Transocean’s tool pusher Jason Anderson and driller Dewey Revette. Thirty-five years old, Jason had worked on the Horizon since it launched, in 2001, and was highly respected by his crewmates. At home before the explosion, Jason had been concerned about putting his affairs in order. He wrote a will and gave his wife, Shelley, instructions about things to do if anything were to “happen to him.” Jason told his father that BP was pushing the rig operators to speed up the drilling. In telephone calls from the rig before the explosion, Jason told Shelley he could not talk about his concerns because the “walls were too thin,” but that he would tell her about them later, when he got home. Jason had just been promoted to senior tool pusher. He had been due to leave for his new post aboard the Discoverer Spirit on April 14, but was persuaded to stay aboard the Deepwater Horizon for one more week. He was scheduled to be helicoptered to his new job at 7:00 A.M. on April 21. By then, he had died in the explosions and the rig was an inferno.

  At the closing of a well, it might seem you’d want the team members most familiar with the well and one another to be present. But approaching the critical juncture of closing up the well they’d been drilling for months, one of BP’s company men, with thirty-three years of experience as a well site leader, was sent off the rig to take his mandatory biannual well-control certification class. Just four days before the explosion, his replacement, Bob Kaluza, appeared on the rig. A Wall Street Journal article said of Kaluza, “His experience was largely in land drilling,” and he told investigators he was on the rig to “learn about deep water,” according to Coast Guard notes of an interview with him. We don’t have a better feel for Kaluza because he has exercised his Fifth Amendment right not to provide testimony that could incriminate himself.

  Transocean’s onshore manager responsible for the Horizon, Paul Johnson, testified that he was troubled by the timing in BP’s switch of well site leaders. “I raised my concerns,” he noted. “I challenged BP on the decision. We didn’t know who this gentleman was. I wasn’t making any assumptions on him, I just—I heard he come from a platform, so I was curious about his deepwater experience in a critical phase of the well. They informed me that Mr. Kaluza was a very experienced, very competent well site leader, and it wouldn’t be a concern.”

  Kaluza showed a tendency toward appropriate caution, but the simple fact that he was new seemed to get in the way. At a critical juncture, Kaluza was uncertain enough about a crucial procedure called a “negative pressure test” that he sought out Leo Lindner, a drilling fluid specialist with the company M-I Swaco. Lindner, who’d worked on the rig for over four years, was in charge of the different types of fluids used during the negative test, an important role.

  “Mr. Bob Kaluza called me to his office,” Lindner testified. “He wanted to go over the method. I briefly explained to him how the rig had been conducting their negative tests and he just wanted—.” Lindner interrupted himself to note, “Bob wasn’t the regular company man on the Horizon.”

  So, competence aside, there were working dynamics, team cohesion. It was a time for familiar faces and an almost literally well-oiled team. But it felt like BP was taking out its quarterback during the fourth quarter of a playoff game.

  And on the morning of April 20—the day the rig exploded—BP engineer Brian Morel departed the rig, creating space for visiting company VIPs. Wrote one industry analyst later, “Let’s face it; the timing of that VIP visit was terrible. It could not have been at a worse time.”

  A difficult pregnancy, new doctors, altered procedures: BP decided to turn this exploratory well into a production well. Usually, the purpose of an exploratory well is to learn about the geological formation and what the oil and gas–bearing production zone contains. Then the well is closed out. Engineers use the information to decide where to drill a production well, perhaps in a nearby spot. If you decide to turn an exploratory well into a production well, you obviously save a fair amount of drilling expense. But is an exploration crew going to be familiar with production technology? BP’s drilling and completion operations manager David Sims testified that the decision was not a major technical issue. Yet veteran well site leader Ronnie Sepulvado who’d been on the rig for eight and a half years had a different take: “We’re in the exploration group, so we hardly ever set production strings. We did maybe a handful of wells that was kept for production.”

  Added complications. The oil and gas—in the pay zone, or “production zone”—lay between 18,051 and 18,223 feet. The well was drilled to below the zone, to 18,360 feet below the sea surface, which allowed cement to be placed under the oil and gas reservoir as well as around it.

  Because this was an exploratory well, the idea was to find the oil, then seal the well shut so a different rig could later tap it for commercial production. Cement is the main barrier for preventing the pressurized oil and gas from entering the well. So it was crucial that the cement job at the bottom of the well absolutely seal off the oil and gas reservoir from the well casing. A bad cement job could let oil and gas into the well.

  The environment at that depth means cementing is not a matter of getting a few bags of concrete from the hardware store. Temperatures and pressures at the bottom of a well like this—it’s hotter than boiling, 240° Fahrenheit—make cementing a highly technical endeavor, requiring calculations and tests to select several chemical mixtures, which will be used in layers.

  Earlier heavy losses of drilling fluid told technicians that they could be into very loose rock and sand. If you are nervous about a soft zone, you also worry that when you insert cement to seal the well bottom, your cement may ooze into the loose stuff. This complicated the cementing deliberations. John Guide: “The biggest risk associated wit
h this cement job was losing circulation. That was the number one risk.”

  If you’re worried that the well walls may be so porous that they’ll suck in cement pumped under pressure, you might add some nitrogen gas to the cement mixture, to get it to form foamy bubbles; this would prevent the cement from leaking into the loose spots.

  From the well’s training resources document: “Foamed cement is more expensive than regular cement and it works better than regular cement in some applications. One of the advantages is that the bubbles stiffen the wet cement so that it is less prone to being lost into a zone or being invaded by fluids in a zone. A remote analogy is that when a sink is drained after washing dishes, the water flows out the drain while the soap bubbles remain in the sink.” But, the document notes, while foamed cement is good at sealing off shallow areas, “use of nitrogen foam is less common for deep high-temperature, high-pressure zones.”

  Halliburton cement specialist Jesse Gagliano first proposed including nitrified cement. After some back-and-forth, BP agreed. But because nitrified cement is usually used for shallower jobs, the depth created concern on the rig.

  Transocean offshore installation manager Jimmy Harrell: “That nitrogen, it could be a bad thing. If it gets in the riser, it will unload the riser on you.… Anything can go wrong.”

  There were three parts to the cement and three formulations. “Cap cement” topped the cement in the space between the casing and the oil-bearing rock and sand formation of the well’s sides. Below that, the nitrified “foamed cement” filled the rest of the narrow space outside the casing and along the formation. “Tail cement” filled the “shoe track” at the bottom and was used inside the lower part of the casing itself.