|
Transocean Investigation Key Findings
This summarizes the key findings of the investigation team based on its extensive review of available information
concerning the Macondo well incident.
As operator of the Macondo well, BP directed all aspects of its development. It chose the drilling location,
designed the drilling program that included all operational procedures, set the target well depth, and created the
temporary abandonment procedure for securing the well before departure of the drilling rig.
As drilling depths increased at Macondo, the window for safe drilling between the fracture gradient and the
pore-pressure gradient became increasingly narrow. Maintaining the appropriate equivalent circulating density
(ECD) became difficult, and BP experienced several kicks and losses of fluid to the formation. BP’s knowledge
of the narrowing window for safe operations guided key decisions during the final stage of operations.
BP’s changes from the original well plan in the final phase included:
• Reducing the target depth of the well
• Considering changes to the well casing
• Using a lower circulating rate than the parameters specified to convert the float collar
• Reducing cement density with nitrogen foam
• Using a lesser quantity of cement than that specified in BP procedures
• Deciding not to perform a complete bottoms-up circulation before cementing
Although aimed at protecting the formation and allowing operations to continue toward completion of the well,
these decisions set the stage for the well control incident.
Running Production Casing
BP chose a long-string production casing design that required the development of a minimal and
technically complex cement program to avoid damaging the formation during cementing, leaving little
margin for error within normal field accuracy. BP and Halliburton then increased the risk by failing to
adequately test the cement program.
The investigation confirmed that the operator’s long-string casing design met the loading conditions that were
experienced prior to and during the well-control incident. The use of this design, however, drove other plan
departures that ultimately increased risk and contributed to the incident. Primarily, cementing the casing required
a complex, small-volume, foamed cement program to prevent over-pressuring the formation. The plan allowed
little room for normal field margin of error; it required exact calculation of annular volume and precise execution
in order to produce an effective barrier to the reservoirs.
The operator had other abandonment alternatives. BP could have either installed a liner and tie-back or
deferred the casing installation until the future completion operations began. Either approach would have
placed additional and/or different barriers in the well prior to the negative pressure test and displacement.A
Deferring installation until future completion operations would have allowed additional time for detailed planning
and verification of the design.
Converting the Float Collar
BP deviated significantly from its plan to convert the float collar, but proceeded despite observations
of anomalies. The investigation team found it possible that the float collar did not convert and thus left
a clear path for hydrocarbons to flow from the formations to the rig.
BP’s planned procedure to convert the float collar called for slowly increasing fluid circulation rates to 5–8 barrels
per minute (bpm) and to generate pressure of 500–700 psi at the float, consistent with the float manufacturer’s
guidelines. However, because of the increasingly narrow window to avoid fracturing the formation, BP deviated
from its planned conversion procedure.
BP made nine attempts to convert the float collar over the course of two hours. BP never circulated at a rate
of more than 2 bpm, but it did increase the pressure applied on each successive attempt, finally achieving
circulation at a pressure of 3,142 psi — almost five times that planned — and a flow rate of 1 bpm, less than ¼
of that planned. BP took this result as an indication that the float collar had converted even though the resulting
circulating pressure was lower than BP’s model had predicted. The BP well site leader expressed concern about
the issue, took steps to investigate it, and discussed the question of whether the float collar had converted with
the Halliburton cementing engineer and BP’s shore team. Halliburton and BP proceeded, apparently having
concluded that the float collar had converted.
The investigation team found it likely that debris in the wellbore may have plugged the shoe-track assembly
and float collar and blocked circulation during the first eight attempts to convert the float collar. The increase to
3,142 psi may have cleared debris from the system without converting the float collar. If the float collar failed to
convert, the cement program may have been further compromised.
Cementing
The precipitating cause of the Macondo incident was the failure of the cement in the shoe track and
across the producing formations. This failed barrier allowed hydrocarbons to flow into the well.
The cement failed as a result of a number of factors that stemmed from BP’s ECD-driven management decisions
between April 12 and 20, 2010. These factors include the complexity of the cement program; inadequate testing
of the cement; likely cement contamination during the operation; and inadequate testing of the cement after it
had been pumped.
Complexity of Cement Job
BP required a cement program that would exert minimal pressure on the formations. To minimize pressure,
Halliburton devised a plan that called for pumping a small volume of cement, much of it nitrified, at a low rate.
While this plan would help BP avoid losing cement into the formations, it required precise execution, left little
room for error, and increased the risk of cement contamination.
Cement Program Tests
Despite the inherent risks of cementing the long-string production casing in the conditions at Macondo, BP did
not carry out a number of critical tests (e.g., fully testing setting time and cement compatibility with drilling fluid)
before or after pumping the cement. Post-incident testing by both CSI Technologies and Chevron demonstrated
that the nitrified cement slurry used at Macondo likely failed.
Cement Contamination
Contrary to best practices, BP decided not to perform full circulation — or a “bottoms-up” — to condition the
drilling fluid in the well before the cement job. A full bottoms-up circulation would have required approximately
2,750 barrels of clean mud to be pumped into the well over about 11.5 hours to keep the ECD under the
maximum limit. Instead, BP decided to circulate only 346 barrels to reduce the chance of fracturing the
formation, increasing the likelihood that debris remained in the wellbore after circulation.
The failure to run a full bottoms-up, coupled with the fact that drilling mud in the well had not been circulated for
more than three days, suggests that cement in the annulus could have channeled and become contaminated.
This could have delayed or prevented the cement from setting and developing the required compressive
strength. Pre-job testing of the cement and spacer/mud/cement compatibility was not sufficient to rule out
contamination.
Post-Cement Program Review
Testing of the adequacy of the cement program could have identified areas of concern, but was not done. After
approving the cement program, BP proceeded with its temporary abandonment plan.
Temporary Abandonment Procedure
BP’s final temporary abandonment plan contained unnecessary risks that were not subjected to formal
risk analysis.
BP engineers generated at least five different temporary abandonment plans for the Macondo well between April
12 and April 20, 2010. The plans varied considerably, as did the level of risk they introduced. The abandonment
procedure ultimately implemented at Macondo never received the required MMS approval. Further, it was
not developed and delivered to the Deepwater Horizon until the morning of April 20, 2010, after the rig had
commenced temporary abandonment operations. The investigation team found no evidence that BP personnel
on the rig or onshore subjected any of the successive temporary abandonment plans or changes to a formal
risk assessment process.
The safest of the five versions (that dated April 14) provided that the surface cement plug be set in mud rather
than seawater and that a negative pressure test be conducted before the drilling mud was displaced with
seawater. The plan that was finally implemented lacked both of these features.
The most significant deficiency in the final plan was the cumulative lack of barriers to flow. The final plan
required displacing the drilling mud to a depth of 8,367 ft. (approximately 3,300 ft. below the mudline), which
was much greater than the normal displacement depth of between zero and 1,000 ft below the mudline. In
addition, the plan removed the mud before testing the cement barrier with a negative pressure test and before
setting the surface cement plug. As a result, no secondary cement barrier was in place during the negative
pressure test and displacement.
Displacement
The initial displacement was planned incorrectly, and the execution did not meet the objective of
allowing for a valid negative pressure test.
The final temporary abandonment plan required displacing the casing annulus below the annular blowout
preventer (BOP) with seawater to achieve the desired negative pressure test conditions. However, post-incident
analysis determined that this objective was not achieved because of calculation errors in the final displacement
procedure, lower pump efficiencies which may have been caused by the unconventional spacer materials,
potential downhole losses, and the movement of spacer below the closed annular. These factors resulted in a
large volume of spacer in the annulus during the negative pressure test that went unidentified due to inadequate
fluid volume tracking and lack of procedures to identify the appropriate pressure readings for a satisfactory
initial test configuration.
With heavy spacer in the annulus below the closed annular BOP, a valid negative pressure test could not be
achieved by monitoring the kill line, which was the method BP decided to use.
Negative Pressure Test
The results of the negative pressure test were misinterpreted. After the test, BP decided to proceed with
the final displacement.
A negative pressure test is necessary to confirm that the cement will block flow from the reservoir into the
well after mud is replaced with seawater. There is no established industry standard or MMS procedure for
performing a negative pressure test, and procedures vary from well to well. At Macondo, BP was responsible
for overseeing the test and determining if the test was successful.
Post-incident analyses confirmed that the test failed. Anomalous pressure observed on the drill pipe during the
test should have alerted all of those monitoring the well to the fact that the cement barrier was not effective, that
pressure was being transmitted past the cement and float equipment, and that the well was in communication
with the formations.
Central to the misinterpretation of the test results was BP’s decision to monitor the kill line instead of the drill
pipe when conducting the test. Had the drill crew continued monitoring flow from the drill pipe, as they had
been doing previously, those monitoring the well would have detected flow indicating that the well was in
communication with the formation
Test and Final Displacement
Post-incident analysis indicated a change in flow path from the well during the final displacement
masked influxes into the wellbore.
Following its approval of the negative pressure test, BP directed the drill crew to proceed with displacing the
riser with seawater and, when the spacer was expected at the surface, to stop operations for a sheen test.
Replacing the heavier drilling mud with lighter-weight seawater during final displacement eliminated the
remaining hydrostatic barrier to flow, leaving the inadequate cement barrier at the bottom of the well as the
primary barrier. According to post-incident calculations, the well became underbalanced to one or more of the
exposed formations sometime between 8:38 p.m. and 8:52 p.m., but there was no clear indication of an influx
at that time.
Just before the pumps were shut down for the sheen test, the trip tank was dumped into the flowline to send
oil-based mud back to the mud pits. Based on post-incident analysis, the resulting increase in flow from the trip
tank across the flow sensors masked an influx into the well.
More than one individual on the rig indicated the well was not flowing when the mud pumps were shut down
for the sheen test. It is possible that no flow was seen because the flow path had been changed to overboard
to dispose of the spacer before the visual confirmations. Post-incident data analysis shows that hydrocarbons
flowed into the well during the sheen test, but the overboard discharge may have masked the flow.
Although the compliance engineer concluded and reported that the sheen test was successful, post-incident
analysis indicates that the spacer had not reached the surface at the time the test was conducted. This report
gave the driller an erroneous confirmation that the displacement was on track when, in fact, it was not.
Pump operations following the sheen test masked an underlying trend of increasing pressure which resulted
from an influx into the well. It is not known what data the drill crew was monitoring or why they did not detect
an anomaly until approximately 9:30 p.m. At that time, the drill crew acted upon a differential pressure anomaly
between the kill line and drill pipe. Actions taken indicate behaviors consistent with a belief that the well was
secure and a plug existed in the well. At 9:42 p.m., the pressure trend provided a conventional influx indication
with a drop in pressure. At that time a flow check was completed on the trip tank and well control action followed.
Sheen Activation of the BOP
The BOP functioned and closed but was overcome by well conditions.
The Deepwater Horizon BOP and electro-hydraulic/multiplex (MUX) control system were fully operational at
the time of the incident, and the equipment functioned. The equipment was maintained in accordance with
Transocean requirements, and all modifications that had been made to the BOP either maintained or improved
the performance of the device. Minor leaks identified pre-incident did not adversely affect the functionality of the
BOP for well control.
Upon detecting flow, the drill crew shut in the well by (1) closing the upper annular BOP; (2) closing the diverter
packer and diverting the flow to the mud-gas separator; and, (3) closing the upper and middle VBRs, which
initially sealed the well.
However, because of the high flow rate of hydrocarbons from the well, the annular BOP element did not seal
and the concentrated flow eroded the drill pipe just above the annular. The closing of the VBRs isolated the
annular space and temporarily stopped the influx, but increased pressure inside the drill pipe until it ruptured atthe point of erosion above the upper annular. The ruptured drill pipe allowed hydrocarbons to again flow into the
riser. When the Deepwater Horizon lost power and drifted off location, the drill pipe parted fully.
The explosions and fire disabled the communication link between the BOP and the rig, preventing activation of
the BOP emergency disconnect system (EDS) from the toolpusher control panel.
The automatic mode function (AMF) operated as designed to close the blind shear rams following the explosion.
However, high pressure bowed the drill pipe partially outside of the BSR shearing blades, trapping it between
the ram blocks and preventing the BSR’s from completely shearing the pipe, fully closing, and sealing the well.
Muster and Evacuation
All personnel who survived the explosions made their own way or were assisted to the forward lifeboat
muster station and successfully evacuated the rig. Despite the obstacles and challenges, the muster
and evacuation plans and training facilitated the evacuation of all 115 survivors.
The Macondo incident created extremely challenging conditions for everyone onboard. The explosions and fire
happened in the evening, when many off-tour crew were asleep or in their cabins. The blast damage blocked
some normal muster points. Some crew were injured and could not evacuate without assistance. It appears
that, under the stress of the emergency, four persons evacuated independently rather than pursuant to the
procedure in which they were trained.
Despite these obstacles and challenges, the muster and evacuation plans and training facilitated the evacuation
of all 115 survivors to the Damon B. Bankston supply vessel nearby. One hundred people evacuated in the
forward lifeboats, seven evacuated in one of the forward life rafts, and eight jumped from the forward end of the
rig into the ocean and were recovered by the Bankston fast rescue craft (FRC). After the survivors reached the
Bankston, the 17 most seriously injured survivors were airlifted by USCG helicopters to hospitals for treatment.
In addition to the heroic actions of many of the crew, assistance from the crew of the Bankston was critical in
the evacuation and rescue effort.
|
The Transocean report concludes that the Macondo incident was the result of a succession of interrelated well design, construction, and temporary abandonment decisions that compromised the integrity of the well and compounded the risk of its failure.
