Big Oil’s New Favorite Toy: Supercomputers
Thursday, April 12, 2018
Xukai Shen, a geophysicist working at BP BP 0.65% Plc, had a hunch he could solve a riddle that had vexed the company: whether there was a lot of oil hidden beneath a salt dome 7,000 feet underwater in the Gulf of Mexico. So he asked to use the company’s supercomputer exclusively for two weeks to check it out.
Using an algorithm, the 33-year-old with a Stanford PhD harnessed the computer’s massive power last year to produce a clearer seismic image of what lay beneath. The result: a potentially massive oil find. With a clearer picture of the area, BP estimated 200 million barrels of crude lay hidden in the Atlantis oil field, a region the company had been plumbing for decades.
“Basically we found a field within a field,” said Ahmed Hashmi, BP’s head of technology for exploration and production, during a recent tour of the company’s Houston supercomputer, as the machine hummed nearby.
BP is now in love with beefy computer power—and it’s far from the only one in the oil patch. Italy’s Eni E 0.69% SpA has built a computing facility the size of a soccer field outside of Milan, crediting its help in all its most recent oil and gas discoveries. France’s Total SA recently upgraded its Pangea supercomputer, nearly tripling its computing power.
While big oil companies were early adapters of supercomputers, some have poured hundreds of millions into upgrades, and now possess some of the most powerful commercially owned computers on the planet.
“We’re going all in,” said Bernard Looney, BP’s head of exploration and production. “We’re only scratching the surface today of what’s possible.”
BP is in the middle of a five-year, $100 million investment in its Houston supercomputer. It’s built a 15,000-square-foot room in a 3-story, flood-proof building to house the titan, which currently takes up about 50% of the space and has the computing power of around 50,000 iPhone 7s.
BP claimed it was the most powerful commercial research computer in the world in December. Within a month, however, it was overtaken by Eni’s supercomputer. BP said it has room to expand its computer further.
Not everyone can take BP’s build-your-own approach. Like high-cost deepwater oil projects, in-house supercomputing remains largely the domain of only the world’s biggest oil companies. But smaller players are finding creative ways to take advantage of technological advances.
“It costs me hundreds of thousands of dollars versus tens of millions,” said Benjamin Williams, Devon’s chief information officer. “You have to decide, am I going to use this giant capacity enough to justify the investment versus the premium I may pay from a cloud provider.”
Mr. Williams said many high performance computing centers are idle 80% of the time, while Devon only pays for supercomputing when it needs it. Devon has a small innovation lab in its Oklahoma City headquarters, where its computer and data scientists can experiment on high-end computers and virtual reality platforms. The company developed three-dimensional visualization tools there that its geologists use to “get inside” Devon’s oil reservoirs.
Exxon Mobil Corp. , the world’s biggest non-state-backed oil company, also outsources its demand for computing power to analyze seismic data and map the rocks it’s planning to drill. Even BP uses outside computers for regular seismic analysis, reserving its giant facility in Texas for cutting-edge research.
How much the digital revolution really changes the century-old oil and gas industry remains to be seen. Some executives and analysts are already warning that companies have unrealistic expectations about the speed of adoption.
John Gibson, who heads the digital innovation team at energy investment bank Tudor Pickering Holt & Co, said most energy companies gutted their research and development funding during the past commodity price crashes, and hasn’t recovered.
BP says it’s already reaping the benefits of experiments with advanced technology.
In Alaska, the company said it crunched 40 years of data on its operations, weather patterns and pipeline corrosion and found ways to maintain its 1,300 miles of pipelines in the state more efficiently by reducing on-site inspections.
The physical inspections—as many as 100,000 locations a year—only found issues 2.5% of the time, and were difficult to perform in a state where temperatures can get so cold that workers can only be in the elements for minutes at a time. With the analysis, BP has managed to reduce inspections 25% and better predict corrosion, Mr. Looney said.