Fire on the Horizon (20 page)

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.

He knew better than most that Macondo had been troublesome, in large part because of its geology. On the one hand, it had relatively high pore pressure—the accumulated weight of all the rock and ocean upon the hydrocarbons trapped inside the sandstone.

To keep the oil and gas from pushing out of the sandstone and up the well, the drillers had to put at least an equal amount of pressure down the hole, in the form of mud. That’s why the drilling mud could be made heavier with additives if needed—the weight of the mud multiplied by the depth of the hole equaled the downward pressure available to hold back the upward pressure of the oil and gas trying to escape.

But there was another variable that made this simple equation complicated: Each section of rock a well encountered had a breaking point. There was just so much pressure the formation could take before it began to fracture. If it fractured, the well lost its integrity, and the mud would start pouring out of the well like water from a leaky cup. The fact that the pressure needed to keep the oil and gas down was very close to the pressure that would punch holes in the well walls was what made the bottom of Macondo so difficult.
Enough
pressure was perilously close to
too much
pressure.

In these circumstances, drilling was difficult. Attempting to get a successful cement job could be a nightmare. They were literally between a rock and a hard place.

As Gagliano began to build a cementing model for Macondo’s bottom-to-top long string, all the particulars conspired against him.

Even the basics of cementing, like cleaning out the well first, could be risky in Macondo’s case. Drilling mud tends to gel like fat in a refrigerator if it’s been stagnant for long. Halliburton’s “best practices” insisted on a full “bottoms-up” circulation of the mud, which meant that all the old mud would be pushed out and replaced with new mud, or at least thoroughly loosened up. This had the added advantage of pushing any trapped gas to the surface, where it could be safely disposed of. But in fragile Macondo, applying enough pump pressure to circulate bottoms-up risked pushing out a wall.

On the other hand, if the cleansing mud flow was too gentle or too brief, in deference to fragile well walls, thick clumps could be left behind, blocking the cement from filling the entire space and leaving channels through which dangerous amounts of oil and gas could rise unbidden through the well toward the surface.

Two factors idiosyncratic to the long string in Macondo made this last issue critical. One was that the space between the outside of the casing and the well wall was thinner than planned—a 7-inch pipe in an 8½-inch hole. BP had intended for the final segment of the well to be nearly 10 inches wide, but with all the problems they’d encountered, 8½ was the best they could do. They could have shrunk the pipe, but 7 inches was the minimum size that would make economic sense when it began to pump oil to refineries.

The problem for Gagliano was that the thin space added to the difficulty of forcing the cement to flow evenly and completely around the outside of the pipe. If the pipe weren’t perfectly centered in the hole, it would be kissing up so closely against the side that no cement would go there at all, choosing instead to flow in the path of least resistance, on the fat side of the space. This was called “channeling,” and it meant that there would be little voids in the cement, allowing hydrocarbons to sneak through the annulus to the top.

Making matters worse was the fact that the MMS required them to seal the annulus with cement to at least 500 feet above the highest oil or gas deposit. The top of the primary deposit was 277 feet above the hole bottom, but a thin hydrocarbon zone had been discovered another 262 feet above that, meaning the cement would have to extend at least 1,039 feet up the annulus. They had no choice but to pump heavy cement—heavier than the mud—up a long, thin space in a fragile hole. The only way to make that happen was to push
down more cement, harder, and for a longer time. Given Macondo’s tendency to crumble, it was a pretty challenging scenario.

Too challenging. When the computer spit up the results of the simulation, its judgment was harsh. The notation on the report read: “unlikely to be a successful cement job due to formation breakdown.”

Gagliano knew he had bad news in his hands of the urgent variety. Fortunately, he was only feet from the BP drilling engineers working on Macondo; his office was on the same floor of the same building on BP’s Westlake campus, west of downtown Houston. He printed out the report and strode out to show it to them. As he turned the corner in the hallway, there they were: Brett Cocales and Mark Hafle.

“Hey, I think we have a potential problem here,” he told them. And he showed them the printout, pointing out where it indicated the potential for gas flow from the deposit and channeling in the cement.

The BP engineering team huddled and quickly came up with a new analysis that cited the unfavorable cement modeling results and recommended that they now go with the previously discarded two-part tieback option because it provided multiple barriers against cement failure.

Within BP, the recommendation was rewritten yet again in a matter of hours, flipping the conclusion. In a well that had already gone so far over budget, the casing hanger and tieback option would “add an additional $7–$10 MM to the completion cost,” the new recommendation said.

It also noted that the same thing that made a long string vulnerable—it was a straight shot from bottom to top—would also make it more durable and less prone to leakage during long-term production use.

Gagliano was sent back to the drawing board to find a way to hold on to the long string. He and Cocales and another BP engineer, Greg Walz, worked into the evening to find a solution. They had already decided to use a special cement impregnated with nitrogen bubbles—it had a texture reminiscent of shaving cream—because it was considerably lighter than normal cement and wouldn’t put as much stress on the well when it was pumped in. Plus, the nitrogen bubbles stiffened the wet cement and made it more resistant to disappearing into a fractured formation—similar to how shaving cream sits in a sink, refusing to go down the drain. But foamed cement alone didn’t make the model work. Now they focused on the danger of channeling, finally realizing that it could be minimized if they could ensure that the pipe would remain almost exactly in the center of the hole from top to bottom. This would eliminate the stray sections where it kissed up against the wall and obstructed the cement from rising.

There were simple devices that could accomplish this. This was the oil industry, so of course centering devices would be called “centralizers.” They were like metal springs that fit around the pipe, projecting equally in all directions. As the pipe was lowered into the hole, the centralizers would push against the walls, guiding the pipe into the exact center of the hole. Gagliano’s initial model had used some centralizers, but by experimenting, they discovered he just hadn’t used enough. Finally they discovered that twenty-one centralizers spaced along the pipe would bring the model to an acceptably small risk of cement failure. And just to be sure, they recommended the use of an expensive testing procedure called a
cement bond log, in which a sophisticated sensor would be sent down in order to determine if the cement had formed a reliable seal with the exterior of the casing.

The cement bond log test would take nine to twelve hours of rig time and cost $128,000 on top of that. But if any channels had been left, it would detect them. The channeling was a serious problem, but correctable—it would take another time-consuming and expensive procedure, but it could be done. The casing in the troubled areas would be perforated and new cement would be pumped through the holes using a special tool. Then those sections would be recased.

If the engineers were pleased with themselves for coming up with a solution, it didn’t last long. When Brian Morel, the BP engineer who’d been dispatched to the Horizon to work on site with the rig’s BP company men, Sepulvado and Don Vidrine, heard about the idea of increasing the centralizers, he e-mailed back: “We have 6 centralizers, we can run them in a row, spread out, or any combination of the two. It’s a vertical hole, so hopefully the pipe stays centralized due to gravity. As far as changes, it’s too late to get any more product on the rig, our only option is to rearrange placement of these centralizers.”

They all understood what the “too late” part of that message meant. Morel apparently wasn’t about to advocate delaying the drilling of Macondo one more time just to wait on a handful of metal rings. They also understood the severe consequences if the computer prediction came true. It wasn’t his job—he was a well engineer, not a parts manager—but Walz got on the phone and not only hunted down the fifteen additional centralizers, but arranged for them to be flown out to the rig on an already scheduled flight the next morning along with personnel to install them. Walz sent a memo to his boss, BP’s Macondo well team leader, John Guide,
telling him what he’d done, adding, “We need to honor the modeling to be consistent with our previous decisions to go with the long string.”

Walz must have been worried that he was overstepping his bounds, and feeling defensive about asking for changes in a cement job so late in the process. In the next sentence he cowered a bit before the boss: “I do not like or want to disrupt your operations…. I know the planning has been lagging behind the operations and I have to turn that around.”

His attempt at diplomacy didn’t appear to have much effect. Guide messaged back that he thought the centralizers had a kind of fitting that could fall off in the hole and cause a whole set of new problems. It turned out Guide was mistaken about that, but he had other reasons to turn down the centralizers Walz had rustled up. “It will take 10 hrs to install them,” he wrote. “I do not like this and…I [am] very concerned about using them.”

Guide’s word turned out to be law.

Brett Cocales, one of the engineers who had worked into the night with Gagliano to come up with the twenty-one-centralizer solution, couldn’t completely mask his bitterness at the outcome. That same day, he e-mailed Morel: “Even if the hole is perfectly straight, a straight piece of pipe even in tension will not seek the perfect center of the hole unless it has something to centralize it.”

In what comes next, you could almost see him briskly brushing his hands:

“But, who cares, it’s done, end of story, will probably be fine and we’ll get a good cement job.”

BP applied to MMS for an amended permit to run the long string. The application was approved the same day.

As BP engineers in Houston wrestled with how to finish off Macondo’s exploration phase, a helicopter full of men in coveralls with the logo of ModuSpec, a risk assessment contractor, descended on the Horizon and began climbing and crawling over every part of it, bow to stern. In 2011, the Horizon was scheduled to go into drydock for the first time since it was launched in Korea. Transocean wanted an independent assessment of what needed tending. When the ModuSpec crew finished crawling through tight spaces and marking up their clipboards, they spent days poring over paperwork and interviewing crew members. The audit was completed on April 14. It identified hundreds of items of urgently needed maintenance, ranging from repairing an out-of-order thruster to complete rewiring of the pipe-racking system. It also noted this: The blowout preventer was significantly overdue for its five-year inspection.

CHAPTER THIRTEEN

UNEASY PARTINGS

April 14, 2010

Sandtown, Mississippi

Ivoryton, Connecticut

On Wednesday night, the night before Dale Burkeen, the Horizon’s starboard crane operator, left for his late April hitch on the Horizon, Janet Woodson told her husband she wanted to go see Dale off.

Janet, Dale’s little sister by eighteen months, hadn’t seen much of her brother this time, and she wanted to make up for that. Dale had spent his first few days off, as he usually did, just staying at home in his tiny central Mississippi community of Sandtown with Rhonda, the wife he adored, and six-year-old Timmy, whom he always called Bo.

Janet didn’t want to interrupt. He and Rhonda hadn’t seen each other in two weeks, after all. Dale was sometimes so gooey with love, even around company, sweety-pie-ing and baby-ing her, that Rhonda would get embarrassed and have to tell him to shut up. “Isn’t she just the purtiest little thing?” he’d say.

And that little boy meant the world to him. Dale would take
him just about everywhere and was always buying him things. She heard he’d just gone out and bought the boy a rifle with a scope in honor of his upcoming seventh birthday, relishing the idea that this would be the year Bo brought down his first buck.

Dale was a big playful boy himself—
very
big at six foot one, 280 pounds. Easter Sunday, he’d thrown a barbecue for the whole family, grilled up a mountain of burgers for their parents and all the cousins, nieces and nephews. But before they ate, Dale had insisted on an egg hunt. Only he didn’t want to be the one to hide the eggs—he wanted to hunt them with the kids. Classic Dale.