Actual flow rates may be more than 10 times what BP is reporting, his calculations indicate

By Janet Raloff

Web edition: May 19, 2010

“It’s not rocket science.” That’s how a Purdue University mechanical engineer described his calculations of startling amounts of oil spewing into the Gulf of Mexico from fissures in heavily damaged piping at a BP drill site.

During a May 19 science briefing convened by the House Subcommittee on Energy and the Environment, Steve Wereley walked members of Congress through his use of particle image velocimetry to explain how he and other engineers track changes in video images of gases or liquids to estimate the volumes billowing before their eyes.

This technique has been around for a quarter century and has thousands of practitioners. So it’s “well-established,” Wereley said. And when done carefully – with good starting imagery – its accuracy can approach 99 percent, he observed.

Six days earlier, BP for the first time publicly released seafloor video of the oil spewing from pipes at the site of its Deepwater Horizon accident. As soon as engineers saw this video, Wereley and a few of his colleagues started mapping features in the roiling plumes and measuring how quickly those identified features sped downstream. Landmarks of known dimensions helped them calculate cross-sections of the plume and its density.

After probing a 30-second live-action snippet from the well’s damaged riser pipe, a conduit that had essentially served as a huge straw to carry oil from the seafloor to a floating platform 5,000 feet above, Wereley calculated the gusher’s flow rate. Then he projected the daily quantity emerging from the pipe’s wound – a staggering 70,000 barrels per day.

On May 18, BP released a few more video clips, this time showing a 1.2 centimeter diameter hole in another segment of piping. Wereley's preliminary calculations indicate that the jet of high-pressure oil shooting out of it unleashes somewhere in the neighborhood of another 25,000 barrels of oil each day. With 42 gallons in a barrel, “It seems incomprehensible that so much oil would be coming out of that hole,” he acknowledged. But this tiny breach is upstream of a plume shooting out of the riser pipe, he explained, “so its flow is at a considerably higher pressure.” 

An hour or so earlier, at a hearing before the House Transportation Committee, BP America president Lamar McKay was asked whether his company still subscribed to the view that the damaged well’s maximum release rate hovered around 5,000 barrels a day. “That is the best estimate,” he said. But estimates are hard to make, he noted, since there's no way to attach a flow meter to the top of the gashes in the damaged pipe.

But when Purdue’s Wereley was asked to hazard a reasonable estimate of the damaged well's oil-release rate, he concluded that BP's quantity was a pipedream. A far more likely figure, he offered, was 95,000 barrels a day, plus or minus 20 percent. At least four other independent engineers have pegged the figure at between 25,000 and 100,000 barrels a day, he reported. So all of these estimates from outside the industry “are considerably higher than BP’s,” he pointed out, “and there’s a good overlap between the outsider estimates.”

This would suggest BP’s number is an outlier, said subcommittee chairman Ed Markey (Dem.-Mass.). It is, Wereley assured him.

Is there any chance BP got the number right, Markey asked?

“I don’t see any possibility – any scenario – under which their number is accurate,” Wereley said. He could envision his own estimate dropping, if longer streams of video were made available and they showed large quantities of gas were being emitted, temporally edging out the oil. The big variable, he said is the gas-to-oil ratio emanating from the well. BP has those numbers but hasn’t shared them yet. And the oil giant also has not been sharing much video.

Earlier in the day, Rep. Markey said, he put in a formal request to BP asking that it begin making live streaming video from its wellhead available to the public.

That’s a good start, Wereley said. But the video he’s seen was “compressed” so that much of the fine detail in its data was missing. What proves critical for high-quality flow analyses, he emphasized, is “original unadulterated footage.”

Markey pledged to look into getting it.