Evidence That a Burning Oil Well Can Sometimes Be the Best Option
Well fires look all consuming, but proving they burn all the oil without leaving a spill behind required the efforts of Boots Coots plus a rocket scientist and a lot of high powered computer equipment.
An oil well blowout can burn like a rocket, which can be a good thing. Burning leaves little or no oil behind, according to a recent study (SPE 189610).
The study has been used to make a case for an oil spill response plan now before regulators to develop a field with more than 100 million bbl of reserves in the Beaufort Sea. Plan approval is one of the remaining barriers facing Hilcorp, the operator in a partnership with BP to develop the Liberty field.
If there is a blowout of a well on the gravel island that will be built for the project, the plan is to minimize the impact by igniting the well, and keeping it lit until a relief well can be drilled to plug it.
Spill plans based on burning have been an accepted option at other fields in that area, including the nearby Endicott field that was developed by BP and sold to Hilcorp in 2014. But unlike Liberty, Endicott is regulated by the state.
In the past, burning was favored because it was a better option than trying to clean up a spill in that harsh environment, which can range from difficult to impossible. Now getting a spill plan approved by federal regulators is considerably more complicated.
Regulations written since the Macondo spill in the US Gulf of Mexico require the operator to demonstrate that a spill can be cleaned up and a relief well can be drilled, said Mike Dunn, operations manager for Hilcorp’s North Slope asset management team.
The plan was challenged by environmental groups, whose critical descriptions of the project took the company’s 90,000 B/D worst-case discharge estimate and assumed all of the oil would be spilled.
To make its case, the partners hired Boots & Coots, one of the best-known oil well firefighting firms that is now part of Halliburton. They brought in a former rocket scientist from the US space agency, NASA, to develop a method to calculate how efficiently a well fire would burn the medium-weight crude that would likely surge out of the conventional field.
The Liberty spill plan is not a new one. “Hilcorp, as well as the previous operator of the leases [BP], have determined that the best remedy for a blowout is to light the well on fire,” Dunn said.
Studies dating back to the 1980s by the State of Alaska concluded that burning was a “viable response option” because it would result in a “large percentage of the oil burned,” according to the paper. A federal study the next year by the US Office of Technology Assessment found ignition of blowouts on gravel islands would immediately burn 95% of the oil another 3% could be removed by other methods.
Those studies were done back when the author of the paper, Steve Fitzgerald, was working on spacecraft propulsion systems for NASA. Now federal regulators are looking for hard numbers based on current science to answer the question: How much oil escapes unburned?
“We needed computational proof and a demonstration that computation was substantiated by visual evidence,” said John Garner, a senior technical advisor for Boots & Coots.
Past estimates were not based on the science or computing power available now. “There was no existing technology to answer that question,” said Fitzgerald, a senior development engineer for Intuitive Machines, a firm employing a team of technical experts from manned spacecraft.
Their job was to develop a reliable system to do so. “The complex physics grew out of spacecraft engineering. I just applied those to the Liberty case,” he said during a presentation at the recent SPE/IADC Drilling Conference.
If there were a blowout at the Liberty field, which is near Prudhoe Bay, it would be the first in that area. “The North Slope basin is a normally pressured basin, with a saltwater gradient, and there has never been a blowout from an oil reservoir while drilling the well. There have been over 5,000 wells drilled on the North Slope,” Dunn said.
If that occurred, the drilling pad will be equipped with an automated system to promptly ignite the well, built by Firefly, a Canadian company that has been selling it for more than 30 years. The system fires a projectile filled with a flammable gel, according to a Firefly presentation.
Winning approval for this plan will require hard numbers on the impact of burning based on experience and rigorous modeling of what would happen in a worst-case scenario.
The starting point was that rockets and uncontrolled wells send out a supersonic-speed jet of volatile gas and liquids from a nozzle. Estimating the efficiency of a car engine’s combustion of gasoline in a controlled environment is simple compared with a blowout where many kinds of hydrocarbons are burned in an unpredictable environment, Garner said.
The variables ranged from the temperature of the burning gas to the size of oil droplets. Larger drops ignite at a higher temperature than small ones.
The method was based on physics and lab studies, past blowout experience, modeling using computational fluid dynamics to model fluid flows, and other engineering modeling approaches.
Based on the research, the temperature in the 100-m long flame would be 3,500°F, more than three times the level where the mist of oil would evaporate and burn. Tiny droplets of oil (10–20 microns in diameter) would flow through burning gas and ignite higher in the column.
Observations of actual oil fires raised questions, like a video of a well in the Rumaila field that was set on fire during the war in Iraq, which appeared to show drops of oil falling out of the tilted casing.
Based on their research and simulation, “as a droplet falls out it is sucked back into the jet and evaporates quickly,” Fitzgerald said.
They also needed to be sure the flame was going to stay lit long enough for firefighters to plan and mobilize equipment to cap the well. Going by what they had observed in modeling and real-world situations, the intense flow from the well, which would produce 84 MMcf/D initially, would not be put out by strong winds or rain.
When they plugged all their data into the multiple models, they concluded the flame would burn at least 93% of what came from the well. That was based on assumptions Fitzgerald described as conservative, so the actual results could be better.
The 7% not consumed come from the heaviest crudes—asphaltenes—which would be “polymerized.” Those tiny particles are expected to slowly descend over a wide area around the well depending on the wind.
“The mathematical modeling that Steve [Fitzgerald] conducted proves the wellbore stays ignited with little to no liquid oil hitting the ice, water, or land around the well,” Dunn said.
The results were consistent with experience. In 13 cases similar to Liberty there was no observed hydrocarbon fallout from any of those wells, Fitzgerald said.
Now it is up to the regulators. With oil prices up this year, the prospects of the project long delayed for economic reasons are looking better.
Hilcorp is working with regulators on an environmental impact statement, which is likely to be completed before the oil spill response plan, which is being reviewed by the US Bureau of Safety and Environmental Enforcement (BSEE).
“We continue to submit and edit and respond to questions from BSEE on our oil spill response plan. We expect the review process to continue through the end of the year,” Dunn said. The plan has to be reviewed by and approved by the US Environmental Protection Agency and the US Coast Guard.
For Further Reading
SPE 189610 Predicting Hydrocarbon Burn Efficiency of Ignited Blowout for Oil Spill Source Control by M.D. Dunn, Hilcorp Alaska; S. Fitzgerald, Intuitive Machines; and J.B. Garner, Boots and Coots.