May 28, 2007
Troubles Run Deep on Gulf Oil Platform
THUNDER HORSE PLATFORM, Gulf of Mexico -- The day after massive Hurricane Dennis churned through the Gulf of Mexico in July 2005, a commercial vessel traveling past BP PLC's hulking Thunder Horse oil platform radioed the bad news to its owner: The platform's top deck was listing into the water.
When a landing party from BP arrived at the platform two days later, they had to tie onto rails near the control tower to haul themselves up the platform's 30-degree incline.
"It looked like a ship that had been sunk," recalled Stan Bond, BP's head of subsea operations for the Gulf of Mexico.
The workers were surprised to learn that the platform, evacuated before Dennis hit, had not taken on water from a leak through its hull. Rather, an incorrectly plumbed, 6-inch length of pipe had allowed water to flow freely among several ballast tanks. That began a chain of events that caused the platform to tip into the drink.
Now BP is attempting to do what no oil company has done before: essentially rebuild the entire architecture of an oil field on the sea floor some 6,000 feet beneath the waves.
At $250 million, the job is costlier, and riskier, than putting the equipment on the gulf floor in the first place. On the frontier of oil exploration, the margin between riches and disaster can be as small as a 6-inch piece of pipe. Yet for BP, rebuilding the platform is critically important because the company desperately needs the oil flowing as reserves in formerly rich fields such as Prudhoe Bay in Alaska dwindle.
Politics have made oil from the Middle East, Africa, Russia and South America is increasingly out of reach. And new discoveries around the world are more rare and continue to shrink in size.
"We have passed the peak for world discoveries," said Robert Gillon, an analyst at oil-industry research firm John S. Herold Inc. "It's hard to see how the industry can do anything whatsoever to materially increase its oil reserves or production."
Against this backdrop, Thunder Horse, sitting atop a reserve that possibly holds 1.5 billion barrels, promises to deliver up to 250,000 gallons of oil a day, making it one of the gulf's biggest producers. For U.S. consumers now paying an average of $3.10 a gallon for gas, Thunder Horse would relieve some of the price pressure: Fully operational, it would boost total U.S. production by 5 percent.
For BP, the troubles at Thunder Horse have turned the oil platform into a dual symbol. Like Janus, the two-faced Roman god that glimpses both the past and the future, Thunder Horse stands as a reminder of BP's mistake-prone recent track record. Looking forward, though, it holds out the prospect of a lucrative, rewarding future.
The Thunder Horse mishap followed by nearly four months BP's worst-ever accident on U.S. soil, a refinery explosion in Texas City, Texas, that killed 15 people. Then, last spring, BP spilled 200,000 barrels of oil onto the Arctic tundra, the first of several pipe leaks that ultimately led BP to temporarily shut down half of North America's largest oil field.
Yet getting Thunder Horse on line will not be easy. The deep Gulf of Mexico is challenging in its own right. Removing and reassembling an oil field at such depths has never been attempted.
One daunting challenge: delicately lifting miles-long strings of steel pipe from the sea floor. If the pipes stress or twist too much, they might weaken and perhaps spring a leak one day, resulting in disaster.
"It will not be easy to pull off," Bond said. "You're trying to change things to make something good. You've got to make sure you don't change things and make them worse."
Exploring the 'Dead Sea'
BP's history in the deep waters of the gulf began inauspiciously. Though the company had drilled in the shallows since the 1980s, in 1991 it began focusing on finding new reserves under water greater than 2,500 feet deep.
The early effort did not go well. At the outset, BP drilled a series of dry holes. At a cost well beyond $100 million, BP was learning why others in the industry had given the gulf a derisive nickname: "The Dead Sea."
"Drilling a succession of dry holes, it was almost the definition of insanity," said Cindy Yeilding, a leading geologist for BP's Gulf of Mexico effort. Executives at BP's London headquarters agreed. For two years, they would not approve any additional deep-water gulf drilling.
But that thinking changed. After a reassessment, BP's oil explorers decided on a new strategy that focuses all the company's energy on seeking big reserves, dubbed "elephants." And the company put big resources behind the new approach: as much as $2.5 billion annually in recent years on gulf exploration.
That's nearly double the amount spent in BP's next-largest target, Azerbaijan, and roughly 20 percent of BP's total exploration and production budget.
The allocation makes sense for BP, because the gulf's deep waters today float above one of the hottest oil prospects on the planet, matching up with Angola and a small handful of lesser places at a time when new huge prospects are not on anyone's maps.
BP is able to move so aggressively in large part because of its $55 billion purchase of Chicago-based Amoco in 1999. Until then, the efforts of BP, like those of other oil giants, had been stymied in the deep Gulf of Mexico by thick layers of liquefied salt that sit like opaque blankets over much of the gulf's oil deposits.
But Amoco had world-class imaging technology, as well as data-mining capability and mathematical algorithms that could interpret the data it collected.
Combining Amoco's tools with some of its own, BP developed a unique exploration approach: It began placing sensing nodes on the floor of the gulf floor. Combining the sea-floor node data with information collected by the conventional method of towing sensors behind a large boat, BP could look through salt from several angles. Suddenly, the opaque blanket was lifted.
Exploration was not cheap, though. These days, it costs about $50 million to fully map a potential oil reserve. To drill an exploratory well costs an additional $100 million. Both those figures are huge jumps from what they would have been a decade ago, when shallower, less complex oil reserves were still available to tap. Back then, roughly $10 million would cover the cost of an exploratory hole.
Such costs and technology hurdles are what drove BP to adopt its "elephant hunt" strategy: Focusing only on the potentially biggest and most lucrative prospects, and ignoring the rest.
An aggressive lease-acquisition strategy, paying $300,000 and more a pop for rights to explore and pump oil from a 9-square-mile plot of the ocean floor, backed the effort. Taking advantage of a controversial Clinton administration program that drastically reduced royalties on deep-water gulf leases sold after 1995, BP stocked up.
And for good reason. Like most oil companies, BP has seen its exploration opportunities diminish over time. Its reserve replacement ratio, which measures whether a company adds new oil reserves at the same rate it depletes its existing resources, has fallen steadily in recent years.
BP's replacement ratio had a modern-day peak of 191 percent in 2001, meaning BP added almost twice as much in reserves as it sold. But that number dropped below full replacement in 2004 and 2005 before climbing above the break-even line again last year, to 113 percent.
By 2006, BP held leases on 650 tracts in Gulf of Mexico water deeper than 1,250 feet. After 15 years of effort, BP was vying with longtime deep-water player Chevron to become the largest leaseholder in the deep gulf.
A host of productive exploratory wells followed. Going by names like Atlantis, Neptune, Mad Dog and Holstein, they are among the gulf's richest finds.
One, at first called Crazy Horse, got a name change after descendants of the Native American warrior protested. Today it's called Thunder Horse.
The $250 million pipe
At a cost of $1 billion to build, and physically imposing with a top deck that rises 15 stories above the water's surface, the Thunder Horse platform appears to be invulnerable to the forces of nature and a wonder of technology. After all, more than 18 major parts on the platform have Serial No. 001 -- meaning they were invented just for this job.
It turns out Thunder Horse is vulnerable to both the power of nature and the shortcomings of modern technology.
The platform was designed to handle hurricanes as strong as Dennis. But the evacuation for the hurricane, combined with just the slightest shifting in Dennis' strong winds, set in motion an unlikely chain of events that caused the platform to tilt. That, in turn, has led to the delay that is costing BP billions in lost revenue -- and serving for the industry as an example of what can go wrong at the outer limits of technology.
The platform rests on four hollow, airtight legs that are as wide across as a two-bedroom apartment. Normally, the legs give the platform buoyancy, and horizontal connecting sections add stability.
After workers evacuated in advance of Hurricane Dennis, though, the misplumbed pipe allowed water to cascade through ballast and bilge tanks. The force of the flow forced open valves that in turn allowed the water to gather in the two port-side legs of the platform.
As Thunder Horse's top deck tilted toward the water, ballast pipes that normally pump water out began taking water in.
The support legs filled with water, and all manner of calamity set in. Some 30 car-size pumps and motors were ruined. A corroding process started that ran through the platform's 25 miles of electric cable and wiring like oil being sucked up by a wick.
"There's the $250 million pipe," said Sammy McDaniel, BP's head of Gulf of Mexico operations, a wry smile on his thin face as he showed a visitor the cleaned-up inside of one of Thunder Horse's large, hollow legs.
Neither McDaniel nor Bond had set foot on Thunder Horse before the mishap. On the first helicopter flight in, they agreed to work together, with McDaniel focusing mainly on the platform's operations and Bond zeroing in on the bottom of the ocean.
"We knew this one was going to be a bear," McDaniel said.
In the weeks after the landing party first boarded Thunder Horse, three days after the storm, the platform became a hive of frantic activity. With 150 workers living on a ship anchored nearby, working with lamps on their hard hats until electricity could be restored, McDaniel and Bond led a frantic cleanup and restart effort.
Work stopped only for hurricanes. After the devastating successive storms, Katrina and Rita, came through, the workers stayed off the platforms while trying to help their colleagues piece their lives back together.
BP's corporate brass told the public that it believed Thunder Horse could restart by late August 2006. Privately, Bond and McDaniel thought they could get the platform back in operation before the end of 2005. Rushing to meet the deadline, workers piled up nearly 4 million man-hours on the cleanup alone.
With start-up approaching, the recovery team in May of 2006 used water to pressure-test the subsea system of pumps, wellheads, piping and gathering centers that sprawl over an oil field on the ocean floor that covers an area nearly as wide across as the North Side of Chicago.
Then the unthinkable happened: The system leaked.
"We were this close," said McDaniel, holding a thumb and forefinger close together. "Then, 'Damn! What went wrong?'"
Sleuthing at 6,000 feet
Perhaps a valve was left open. Perhaps a coupling on a pumping station wasn't properly tightened. "We figured we would find out in our spare time," McDaniel said.
Two weeks passed, then a month. One pressure test held for eight hours, and then failed. The team injected ink into the piping network and sent a remotely operated, unmanned submarine 6,000 feet down into the water to photograph what was going on.
Outfitted with cameras and high-precision robotic arms, the sub was capable of spotting any problems, and fixing many potential mishaps.
As the robot's operators watched on a black-and-white video monitor inside a cramped control room, the camera focused on an image of a sea-floor metal structure, a manifold. The size of the container on the back of a semi-truck, the manifold is a key piece of equipment that ties together the lines from dozens of sea-floor wells, and then helps transfer oil up toward the platform.
Most of the huge manifold looked fine. But on the side, on one of the large pipes that snaked through the frame that formed a sort of exoskeleton for the structure, was a shocking sight: an inch-wide gash slashed through a weld. The leak was found.
Cause of the cracks
Perhaps it was just one bad weld, but McDaniel and Bond had to determine if there were any more. They directed the submarine to another manifold and found a second ruptured weld. Inspection of other welds in the subsea equipment turned up even more cracks.
Thunder Horse's oil reservoir is nearly 5 miles below the water's surface. At that depth, oil will gush from the drill pipes at a temperature of 275 degrees Fahrenheit, under a metal-crunching 17,400 pounds per square inch of pressure.
Those conditions can stress even the mixture of high-strength steel and alloy that make up the half-inch welds on the manifolds and pipes of the Thunder Horse oil fields. But the equipment had gone through severe tests -- at 125 percent of the worst stresses that the Thunder Horse field might exert.
There had to be something else.
"We're operating at the edge of what is known," said Kenny Lang, BP's head of Gulf of Mexico operations. "When you're at the edge, you're creating knowledge. And when you create knowledge, you sometimes stub your toe."
Now the hunt was on for a new spot of knowledge: What caused the problem?
Lang flew in a team of experts in subsea oil production, welding and metallurgy from around the world to Houston to determine the cause of the weld failures.
Meanwhile, he directed others to touch base with the manufacturers of every component built into the sea-floor manifolds. He asked for testing of the anti-corrosion materials and insulation that enshrouded the subsea pipes. He wanted no clue missed.
"Ultimately you say, 'What if I'm wrong about what caused this? We put our equipment back on the seabed, and it fails?' " Lang said. "You can't risk that."
Lang also wanted other oil companies to be aware of the dangers. Learning that Shell Oil Co. was due to submerge manifolds at depths similar to Thunder Horse in the fourth quarter of last year, he made certain Shell was notified of the possible risks.
Even as the investigation started, though, pressure mounted onboard Thunder Horse.
BP had commissioned the Balder, one of only two ships in the world capable of lifting the manifolds and other heavy equipment from the sea floor, to visit the platform in December. After that, the Balder wouldn't be available again for almost a year.
By late September 2006, the manifold investigation team delivered its verdict. The welds, indeed, were the problem thanks to an unforeseen chemical reaction.
While the manifolds sat idle for a year after the platform tilted, the crushing pressure at the bottom of the sea forced hydrogen atoms into the mix of steel and high-strength alloy that made up the welds. The hydrogen caused the metal to become brittle, and when water was forced through the piping during the restart testing, the welds failed.
Drilling toward Mardi Gras
In the meantime, Bond hadn't been waiting for a verdict. He knew he only had until the end of 2006 to have all the sea-floor equipment ready to be lifted. That meant sealing wellheads, cutting pipes and planning logistics. It also meant working around the schedules of the 280 people onboard Thunder Horse, some of whom continued drilling new holes even as the rest of the sea-floor operation stood idle.
Drilling, after all, is what Thunder Horse was built to do.
On a recent spring day, a team of workers operated the ship's drill rig, pulling up a drill bit that had gone more than 20,000 feet below the seabed floor. Nearly 3 million pounds of pipe stretched from the drill rig to the bottom of the hole.
Spinning furiously, with "mud" that is used to lubricate them spitting out of the hole, the drilling pipes came out in 95-foot sections. As each joint emerged, workers stood by as a huge, mechanized clamp twisted off the coupling that separated it from the long line still stuck in the ground.
Directed by operators using joysticks in an air-conditioned control studio, an overhead winch grabbed the newly freed section of pipe and hung it on a rack. The pipes knocked together, sounding like a supersize wind chime.
Nearby, in the main Thunder Horse control room, BP workers monitored huge computer displays that showed the pressure, temperature and fluid volumes in all of the oil platform's piping systems.
There was something eerily missing on the screens, though: Not an ounce of oil was anywhere to be found.
The rebuilding process
Today, Thunder Horse's crews have removed about three-quarters of the equipment that once nestled on the seabed. They are putting new insulation and anti-corrosion coatings on some, replacing other pieces entirely.
The most delicate operation -- pulling the pipe up from the seabed without bending it -- is necessary, Bond said, because it's the only way he can reassemble the equipment that's needed on the oil field. The deep-sea robots can cut the pipes at the point they connect to the equipment 6,000 feet below the surface. But robots can't weld.
So Bond must oversee an operation that pulls up the freed pipe and brings it within reach of the Thunder Horse deck. There workers can weld it back to the huge, heavy pieces of equipment. Then BP workers must carefully lower the joined pieces back down, all without causing any new problems.
No one says it will be easy. But everyone onboard says it must happen on time. They will need the Balder for some of work, and demand for that ship is so high that it only comes by every 18 months or so.
"We've just been going full speed for a long time, and there's no letting up," said McDaniel, the operations chief.
"What we want to do is prove to ourselves and the world that we're ready," he said. "We just need to get all this stuff under us, and begin operation."
Leading a reporter on a tour of the complex onboard systems that separate oil, water and gas, McDaniel pointed to a pipe from the platform that plunges deep into the ocean. By the time Thunder Horse goes into production, the pipe will connect to Mardi Gras -- a $1 billion pipeline BP is building that one day will carry half of all the oil pumped from the deep-water gulf."This is the top end of the Mardi Gras pipeline," McDaniel said. "When the oil leaves here, it's gone."
For BP, and for gas-hungry consumers across the U.S., it can't happen soon enough.
- - -
About this story
Chief Business Correspondent David Greising spent six months reporting on BP PLC's attempts to fix three major problem sites within its huge North American unit. Greising, along with Tribune photographer Bob Fila, reported from Deadhorse, Alaska, site of the Prudhoe Bay oil spill last summer; Texas City, Texas, the location of a deadly refinery explosion that killed 15 in 2005; and atop BP's Thunder Horse oil platform in the Gulf of Mexico, which nearly toppled into the sea during a hurricane in 2005. The Tribune is the first news organization to visit all three sites since the disasters and the only one ever to set foot on Thunder Horse in the gulf.
IN THE WEB EDITION: Read Part 1 of the series, plus the Tribune's David Greising describes his visit to BP's facilities and how the oil giant is struggling to rebuild in a video at chicagotribune.com/bp