When it became clear that the blowout preventer on BP's ill-fated well in the Gulf of Mexico could not be activated, BP began drilling two relief wells nearby. These were intended to provide a conduit for injecting dense mud and cement into the out-of-control well, thereby plugging it. At press time, the relief-well operation was scheduled to be completed in early August.
Drilling relief wells in response to a blowout is a standard tactic in the oil and gas industry. But that doesn't mean that such wells are easy to engineer. It takes a rather sophisticated geophysical sensing system, specialized simulation software, and some careful calculations. In particular, guiding the drilling of relief wells is a notoriously tricky business.
The fundamental problem is that a relief well must intersect with the target well at great depth—almost 4 kilometers below the seafloor in the case of the BP well. "The deeper you are, the more back pressure you can apply," says Elmo Blount, former manager of Mobil's Dallas-based drilling technology group. And it's the back pressure—mostly from the weight of the column of mud injected into the blown-out well—that will stop the flow. But the target is so slender—about 25 centimeters in diameter for BP's well—that hitting it from such a distance is like threading a needle from across the room. With your eyes closed.
The orientation of oil and gas wells is carefully measured as they are drilled, so BP knows where the problematic well is positioned throughout most of its depth. But small uncertainties in the orientation measurements add up. At the depth of the intended intersection point, the location of the well is known only roughly. So the relief well must be very carefully guided to its target as it is being drilled, using some form of geophysical sounding.
In years past, that was done using a magnetometer to detect the steel lining (the casing) of the target well. The problem with that approach is that you have to get pretty close to the casing—within about 10 meters—before you can sense it. The more modern approach has a much longer range and is able to provide the direction to the target well. "I think of it as electromagnetic sensing," BP spokesman Kent Wells told reporters in June. He was referring to a technique developed in the 1980s by Arthur F. Kuckes, who is now an emeritus professor of applied and engineering physics at Cornell University. His system is built and marketed through Vector Magnetics, based in Ithaca, N.Y. The radius of sensitivity with Kuckes's technique is almost 10 times as great as that from the earlier system of passive magnetic sensing.
Representatives of Vector Magnetics declined to be interviewed, but patent filings make it clear that Kuckes's method isn't truly electromagnetic in the usual sense of the word. That is, it doesn't depend on time-varying magnetic fields to induce electrical currents. Rather, the strategy is to inject current into the conductive steel casing of the target well more directly, using electrodes lowered into the relief well. The surrounding rock conducts electricity sufficiently to carry current away from these electrodes, and much of that current is then channeled into the conductive casing of the nearby target well. That concentrated current, in turn, creates a magnetic field whose presence and orientation can be measured several tens of meters away. All you need is an appropriate magnetometer at the bottom of the relief well.
Even with such guidance, drilling a relief well is a slow process. "We're going to drill a couple hundred feet, test and see where we think the well is, drill another 100 to 200 feet, do it again....We're sort of homing in on exactly where the well is," Wells told reporters.
Blount, who helped control blowouts in Sumatra, Canada, and Venezuela using relief wells, emphasizes how much precision is required: "If you hit it at an angle, it'll skid to one side—you're working with inches."
Complicating matters is the requirement that drillers make a good connection between the relief and target wells. Without that, BP might have trouble pumping mud fast enough into the blown-out well to stem the flow of oil and gas. The problem is that the hydrocarbons spewing upward will mix with the mud that is injected, and if there isn't enough mud going in, the density of the mixture will be too low to provide the back pressure needed to stop the flow.
Even if the targeting of the relief well is done perfectly and the connection between the relief and target wells is a good one, there is yet another complication: Too much back pressure could cause the rock formation that surrounds the well to fracture. Then the injected mud would leak into the surrounding rock, and oil and gas would continue to flow up the well. The delicate balancing act comes down to something as seemingly mundane as the density of the mud being pumped. The specialists BP has hired will likely use simulation software called Olga-Well-Kill to determine the optimal mud density and pumping rates. This software was developed at the Institute for Energy Technology, in Norway, after a 1989 blowout at an oil well in the North Sea.
As this article went to press, the outcome of BP's efforts to direct a relief well to its target and the challenges of killing that well afterward were not known. All that was clear was that this ongoing disaster would take considerable technical skill to bring under control.
This article originally appeared in print as "Targeted Relief".