After several months of fruitless subsea efforts, BP appears to have finally plugged the leaking well in the Gulf of Mexico.
BP faced heartbreaking setbacks in its early attempts to stem the oil gushing from Macondo field, which included a failed 'top kill' operation. But the company had a multi-strand strategy, and knowledge gained from the failed early efforts helped it formulate more successful plans, which eventually led to the installation of a temporary cap on 15 July. Prior to capping the well, the company also succeeded in capturing over 700,000 barrels of oil on the seabed.
The capping operation presented enormous challenges, not least because an enormous fleet of ships was clustered on the surface. Things were even more congested on the seabed with a dozen remote operated vehicles (ROVs) working in close proximity in the pitch-dark, ice-cold conditions on the seafloor, in water a mile deep. The ROVs needed to navigate around a work site cluttered with pipes, cables, tools and debris, close to the open mouth of the well, which was belching vast quantities of oil and gas. So how did they arrive at this point?
Early containment efforts
The blow-out on 20 April resulted in the loss of 11 lives on the Deepwater Horizon drilling rig. The rig subsequently sank. Surveys carried out shortly after the disaster found that the drilling riser was still connected to the top of the well, and had collapsed under its own weight. It snaked around, and turned in on itself, so that its open-end lay just 600ft from the wellhead, with oil leaking at three points. A sharp kink in the riser above the blow-out preventer (BOP) was one source of leaking oil. This kink also throttled the flow along the mile-long pipe, thus reducing the size of the oil spill.
The first effort to contain the leak involved placing a 40ft-high open-bottomed chamber over the end of the riser, so that that oil and gas could be piped up to the drill ship Discoverer Enterprise. This operation failed, since hydrate crystals formed and blocked the outlet.
Hydrate is a regular problem for deepwater production, since the ice-like crystals block pipes, valves and fittings. Hydrate forms when natural gas mixes with water in the conditions of low temperature and high pressure, which are found on the ocean floor. BP had taken steps to inhibit hydrate, with chemical and hot water, but these measures proved inadequate with the first containment effort, since the large volume of the chamber permitted gas to mix too freely with seawater.
So the next solution was to collect oil by inserting a 4in-diameter probe into the open-end of the riser - this would allow the oil to be collected before it had a chance to mix with seawater. But to ensure that this was successful, it was first necessary to seal the broken end of the drill pipe, which stuck out of the end of the riser. This task involved cutting the drill pipe with a circular saw, then fitting a valve on the cleanly cut pipe-end.
The probe was inserted 5ft into the annulus between the 20in riser pipe and the 6in drill pipe, whose end had been sealed. Rubber baffles were provided on the riser insertion tube to channel the oil into the probe, and the fluids were piped up to a vessel on the surface.
This system was operating by 16 May and collected around 3,000 barrels of oil per day. It went on to recover a total of 22,000 barrels. It was put out of action for a short period when a ROV knocked the tube out of position.
Fitting the 'LMRP' cap
The riser insertion tool became redundant when BP made a decision to cut the riser, and collect oil with a cap from a position immediately above the BOP. Oil would be collected by the cap, before it could mix with seawater, and would pass up a pipe to be processed on the Discoverer Enterprise.
This plan involved removing a section of the broken riser, so two cuts needed to be made through the 20in-diameter pipe - a considerable challenge in 5,000ft of water. The first cut would be downstream of the BOP, and the second would be above an assembly known as the lower marine riser package.
The first cut was carried out with an 18ft-long tool, which weighed 46,000lbs and was described as 'like giant pruning shears'. To cut through the inch-thick steel pipe wall, the hydraulic tool was capable of exerting 3,000t of shear force.
The next cut, immediately above the LMRP, was due to be carried out with a wire-saw. This would provide a cleanly cut face as a base for the collection cap. But the saw blade became stuck halfway through. After freeing the saw, it was withdrawn to the surface, with oil flowing uncontrollably from the part-cut riser. To save time, a decision was made to use the super-shears to complete the job. As a result, the base for the cap was more jagged than planned.
The LMRP cap was provided with four 'chimney vents' to keep it stable as it was lowered though the billowing plume of oil and gas gushing from the well. The intention was to close the vents gradually to increase flow of oil to the surface. Various measures were taken to inhibit hydrate formation.
The cap was lowered to the seabed on 3 June by the drill ship Enterprise, and, once optimised, successfully collected around 15,000 barrels of oil per day, which was close to the maximum capacity that could be handled by the drill ship. As it was unable to handle the entire spill, clouds of oil and gas could be seen on live video feeds billowing out, obscuring the cap from view.
The billowing clouds were visibly reduced on 16 June, when a second containment system was connected up to pipe fluids from the BOP to the semi-submersible Q4000. This system used equipment put in place for a failed 'well kill' operation, and it captured an additional 8,000 barrels of oil per day.
Despite this success, concern was growing about the need for additional containment capacity, and the need to hurricane proof systems so that they could be rapidly connected and disconnected if a hurricane threatened. So the unified command gave instructions that BP should bring in more ships and provide greater system redundancy.
In late June, the subsea team was waiting impatiently to connect another production vessel, called Helix Producer, to a buoyant free-standing riser, but work was held up by bad weather. The free-standing riser would permit easier disconnection, should the vessel need to evacuate the area. This vessel was intended to provide a third leg to the containment operation, with capacity to recover an additional 25,000 barrels of oil per day. Work was also underway for a fourth containment system, involving another free-standing riser, which would bring total recovery capacity up to 80,000 barrels per day.
Their work had been delayed by Hurricane Alex, which passed far to the south but caused rough seas at the spill site. But then a seven-day period of calm weather was forecast from 10 July. This weather window was extremely fortunate, because it was sufficient to allow the subsea team to implement an even bolder plan in which they would unbolt the flange beneath the ill-fitting cap, and fit a new capping stack, to stem the flow.
BP engineers had been working on the design of various capping stacks at the Cameron works. The chosen cap was built up from standard components, and had all the functions of a BOP, but it was small and lighter than conventional equipment. The stack stood 18ft-high and weighed 150,000lbs, and was designed to latch directly onto a transition spool. In addition to the three rams, the stack incorporated a choke valve which could be used to throttle the well flow.
But installation of this capping stack was a formidable challenge. Preparatory work had been carried out to straighten a flexible joint, which was put three degrees out of vertical when the riser collapsed. A hydraulic power package and distribution system needed to be provided on the seabed to carry out this work, then a dozen hydraulic jacks with inclination instruments needed to be fitted around the flex-joint. Once the flange was level, chocks were inserted to hold the flex joint in the vertical position.
Admiral Allen finally gave the go-ahead for the capping operation on 9 July. This was a bold decision given that removal of the LMRP cap would allow uncontrolled leakage to the ocean. The main challenge for the subsea team was to loosen six large bolts which secured the flange. This was carried out with a hydraulically-powered wrench. But there were other important tasks, such as the need to clean the flange face, to ensure a pressure tight seal.
Great patience was needed to carry out this work because the ROVs were being buffeted about by turbulent currents produced by the escaping wellfluids. But the subsea team persevered throughout the evening of the 10 July - by midnight (local time) all six bolts had been removed and much of the other work had been completed.
It was recognised that the flange surfaces needed to be prised apart. So a flange removal tool, with a gripping mechanism, had been devised to remove the pipe stub (see illustration). This was deployed from another drill ship, called the Discoverer Inspiration.
Installing transition spool
Immediately after removal of the pipe stub, another difficult operation was needed to install a 'flanged transition spool'. This spool stood 12ft-high and weighed around 15,000lbs. Installation required enormous skill because the spool was swinging about at the end of a mile-long cable, and was being lowered directly into the plume oil and gas billowing out of the well. Further complications arose because broken drill pipe stuck out of the open flange.
Special features were built into the spool to guide it into position and ensure that the bolt holes were aligned. Cables were also used to help guide the spool into position. Once it was placed in position on 11 July, six pre-fitted bolts were quickly tightened and checks were made to ensure that the flange was sealed.
Capping the well
With the transition spool in place, installation of the 'capping stack' was relatively straightforward. The stack was lowered to the seabed by the Inspiration and was latched on to the transition spool on 12 July.
Great caution was needed in closing the valves as there was a risk that fluids could leak out underground if the well casing had been damaged in the blow-out. So the plan was to gradually choke the well flow, while monitoring pressures at various points, to test the integrity of the well. Extra tests, including regular seismic surveys, were later added at the behest of the scientific community.
Well integrity test
A well integrity test was started on 15 July when the choke valve was closed. This was a landmark event, since the flow of oil into the Gulf was halted for the first time since the blow-out on 20 April. While there were no signs of leakage, the pressure at the well head was lower than expected, causing some concern. But the pressure rose slowly, and experts formed the view that this was consistent in a well that had been flowing for some time at an excessive rate. Some gas bubbles were also detected on the seabed, but analysis showed that this was a natural seepage.
The apparent success of the capping operation, at the time of writing, could permit a 'static kill' operation to take place in the first week of August. If carried out in tandem with a 'bottom kill', via a relief well, this could lead to permanent plugging of the troublesome Macondo well by late August.