A look back: Berkeley Lab scientists raced to estimate oil flow from Deepwater Horizon macondo well

The first two weeks of June 2010 were a blur for six scientists fromthe U.S. Department of Energy's Lawrence Berkeley National Laboratory(Berkeley Lab). As the world focused on the ongoing crisis in the Gulfof Mexico after the blowout of BP's Deepwater Horizon Macondo well,the scientists dropped everything to estimate how much oil was flowingfrom the mangled wellhead.

The clock was ticking: Their work would help assess the environmentalimpact of the disaster, as well as develop ways to cap the well, whichhad been spewing unchecked since April 20.

They used some of the world's most sophisticated numerical modelingtools, developed at Berkeley Lab over the past two decades forapplications ranging from geothermal energy production toenvironmental hydrology.

Working quickly and amid abundant uncertainties, they estimated thatbetween 60,000 and 100,000 barrels of oil were flowing into the Gulfeach day. Their calculations were in line with a final estimatederived two months later based on much more information.

Their research is recounted in an article published in this week'sonline early edition of the Proceedings of the National Academy ofSciences.

"We were able to harness Berkeley Lab's expertise in multiphase flowand computational tools to quickly take on this urgent problem," saysCurt Oldenburg, a staff scientist in Berkeley Lab's Earth SciencesDivision and lead author of the article. Also on the team were fellowEarth Sciences Division scientists Barry Freifeld, Karsten Pruess,Lehua Pan, Stefan Finsterle, and George Moridis.

The scientists were part of a group established by the NationalIncident Commander in May 2010 to estimate the oil flow rate from thewellhead. One component of this effort comprised scientists from fiveDepartment of Energy national laboratories, including Berkeley Lab.

The Berkeley Lab team first developed a simplified conceptual model ofthe system despite a lack of knowledge about the flow path from thereservoir into the well, reservoir permeability, and pressure in theblowout preventer. They then developed a coupled model of thereservoir and wellbore using a numerical program, called TOUGH2, whichsimulates fluid and heat flow in porous and fractured media.

Their simulations painted a range of flow rates, from a low of 60,000barrels of oil per day to a high of 100,000 barrels of oil per day.Their initial estimates are in line with a final estimate establishedin August 2010 by the entire group and based on independent analysesand observations. It pegged the rate at 62,200 barrels of oil per dayupon initial blowout in April, tapering to 52,700 barrels per day justbefore the well was capped in mid-July.

The Berkeley Lab team's modeling approach also allowed them todetermine the role played by various uncertainties. For example, theyfound that the rate of oil flow greatly increased as the length of thewell in contact with the reservoir increases.

Surprisingly, they also determined that oil flow rate is relativelyinsensitive to the pressure at the bottom of the blowout preventer.Common sense dictates that as pressure drops at the bottom of theblowout preventer, the oil flow rate increases. Instead, thescientists found that the lower the pressure, the more natural gasexsolves from the oil. Natural gas interferes with oil flow andcounteracts the pressure that drives oil upward in the well.

Source: DOE/Lawrence Berkeley National Laboratory