BP Deepwater Horizon Accident Investigation Report

my name is mark blye I'm the group head of safety and operations at BP and I led the company's internal investigation into the events surrounding the tragic accident aboard the Deepwater Horizon on April 20th 2010 as you may know the Deepwater Horizon drilling vessel was stationed in the Gulf of Mexico on the Macondo well the rig crew was in the process of finishing work on a deepwater exploratory well when a series of events culminated in a blowout of the well explosions and fire on the rig the tragic loss of 11 lives a number of serious injuries and one of the largest oil spills ever seen this video summarizes the findings and analysis of this investigation team as I review these it's important to keep in mind the goals of the investigation our team was assembled immediately after April 20th in order to identify the critical factors underlying the accident our mandate was to determine what happened so that BP could take whatever steps necessary to prevent a similar accident from ever happening again our purpose was not to apportion blame or liability but rather to learn recommend areas for improvement and share our lessons with others while I believe that we had sufficient evidence to reach the findings summarized here our work was limited by a lack of direct access to some witnesses and to certain physical evidence I will be presenting our findings factually but it is important to note that our purpose was not to create a record that would serve as the definitive or final word on what happened the investigation team consisted of more than 50 internal and external engineering specialists with expertise in various aspects of deepwater drilling and exploration we worked over a four-month period examining all available evidence this visual portrayal of our key findings is best viewed in the context of the report which describes more fully the basis for findings the findings were based on witness accounts analysis or modeling as well as engineering drawings real-time data from the Deepwater Horizon physical evidence and testimony gathered from governmental inquiries the team's report was delivered to the board and VP's top management they are sharing it publicly in order to keep the public informed to assist with the government's various investigations and to ensure that BP and the industry can learn from the accident and take appropriate actions going forward to determine what led to the tragedy it's important to first understand the events that occurred aboard the Deepwater Horizon leading up to the accident this short animation depicts the best understanding the team was able to reach about the chronology of key events as we understand them today this is an animated reconstruction of the tragic accident aboard the Deepwater Horizon on April 20th 2010 specifically this narrative has been developed to illustrate what we determined with the key events related to the accident the animation contained in this narrative is representative and for illustrative purposes only and has not been drawn to scale the Deepwater Horizon was a dynamically positioned drilling vessel whose construction was finished in 2001 it was designed to drill in water depths up to 10,000 feet the vessel arrived on the Macondo location on January 31st 2010 to finish the exploratory well that had been started earlier by its sister vessel the riser and the Bo P were installed successfully for casing strings were installed over a two and a half month period our narrative begins on April 9th 2010 the final whole section of the well was drilled to a total depth of 18,000 360 feet the long string production casing was installed during the afternoon of April 19th the cement job started cement is used as a barrier to seal the well bore from the reservoir sand nitrified foam cement was pumped into the annulus and non foam cement was pumped into the shoe track at this point the pump pressure was bled off and it appeared that the check valves in the flow collar were holding it's now the of April 20th the wellhead seal assembly was installed and tested a positive pressure test which is the first integrity test of the well was conducted by closing the blind shear rams and applying 2700 psi of pressure the tests confirmed integrity of the blind shear Rams the seal assembly casing and top wiper plug the shoe track which also plays a key role in isolating hydrocarbons is not tested because of the presence of the top wiper plug it's now early afternoon the drill pipe was run to eight thousand three hundred and sixty-seven feet in preparation for the negative test which was the second test of the wells integrity the tests purpose was to place the well in a controlled under balanced state to test all mechanical barriers it was conducted by displacing some of the mud in the well with a spacer followed by sea water after displacement the upper annular preventer was closed an attempt to bleed the system down to zero psi was made but fluid in the riser was leaking past the annular preventer it's now 1708 the hydraulic closing pressure for the annular preventer was increased to 1,900 PSI in order to create a tighter seal against the drill pipe the riser was filled with an estimated 50 barrels of mud to replace the volume that had leaked past the annular at seventeen twenty seven the drill pipe valve was opened and the pressure reduced to zero by bleeding off fifteen to twenty three barrels of sea water it's now 1735 the well site leader advised the rig crew that the negative test procedure needed to be conducted on the kill line to meet permit requirements the drill pipe valve was closed in the testing reconfigured for flow to be monitored on the kill line at 1752 the kill line valve was opened based on eyewitness accounts somewhere between 3 and 15 barrels of seawater were bled off the kill line valve was closed and the drill pipe pressure gradually increased it's now 1842 seawater was pumped into the kill line to confirm that it was full the Joline was routed to the mini triptank and less than one barrel was bled from the kill line the flow stopped the kill line was monitored for 30 minutes and showed no flow but the drill pipe pressure remained at 1,400 psi a discussion about the source of the 1,400 psi took place it was explained as a phenomenon called the bladder effect the test was deemed successful it's now 20 hundred hours as part of the normal operations to temporarily abandon the well the crew began to displace the remaining drilling fluid with seawater to start this operation the annular preventer was opened and the well returned to an over balanced condition preventing further influx into the wellbore the displacement continued as planned and the well went under balanced at 20 52 this means the pressure in the well dropped below the reservoir pressure the well started to flow it's now 20 101 the crew was emptying the trip tank which likely masked any flow indications on the flow meter at this point with constant pump rate pressure should have declined as a heavier mud was being replaced with lighter seawater instead drill pipe pressure increased by a hundred psi indicating a problem with the well analysis of real-time data shows that a thirty nine barrel gain was taken by 2108 it was then time for the sheen test on the spacer which is done in order to check that there is no free oil in the fluid that will be discharged to the sea the pumps were shut down when the spacer reached the surface machine test was performed and the spacer was determined to be suitable for discharging overboard the drill pipe pressure increased by 246 psi in five and a half minutes while the pumps were off the fluid returned from the riser was routed to the overboard depth line during this period our modeling suggests that the well was flowing in an estimated nine barrels per minute displacement of the mud with seawater resumed at 2114 it's now 2131 all mud pumps were shut down an estimated 300 barrel gain has been taken over the next five minutes the drill pipe pressure increased by 556 psi rig personnel discussed the differential pressure on the drill pipe in the next two minutes it appears an attempt was made to bleed the drill pipe possibly to investigate the differential pressure over the next 20 minutes a series of critical well control events occurred ending in the order to abandon ship it's now 20 140 mud overflowed onto the rig floor a minute later the mud shot up through the derrick the diverter was closed and flow is diverted to the mud gas separator at about the same time the rig crew appeared to close an annular preventer and drill pipe pressure increased steadily from 338 psi to 1200 psi over a 5-minute period mud and hydrocarbons discharged onto the rig and overboard activation of the annular preventer appears to have failed to seal the annulus it's now 2147 first gas alarm sounded a cloud of gas spread setting off other gas alarms there was a roaring noise and the vessel vibrated the drill pipe pressure rapidly increased to 50 730 psi which is thought to be the result of the Bo P sealing around the drill pipe likely after activation of the variable bore Rams or the annular preventer fully sealing one minute later the combustible gas cloud reached the aft starboard quadrant of the main deck and probably entered the air intakes for the engine room gas also dispersed under the deck and into the engine room enclosures the main power generation engines went into overspeed within another minute electrical power was lost and there were two explosions the first occurred at about the same time as the power loss the second occurred an estimated 10 seconds later the second explosion appears to have taken place under the main deck in the aft starboard quadrant of the rig The Associated blast pressure wave traveled forward and caused extensive damage to the star side potentially damaging the MUX cables which allow the vessel to communicate with the Bo P the hydraulic line which powers the Bo P may also have been damaged the fire and explosions may also have caused damage to the surface drill pipe equipment resulting in hydrocarbons flowing to the surface through the drill pipe the valve on the drill pipe was left open exposing surface components to high shut in pressure while the B opiate closed around the drill pipe it did not seal the well and the subsequent loss of power may have resulted in the rig drifting off station the drifting rig may have broken the Bo P seal around the drill pipe and allowed hydrocarbons to continue to flow up the riser one or a combination of these events ultimately allowed the fire to be continually fed it's now 21:52 over the next five minutes personnel attempted to shut in the well and disconnect the lower marine riser package from the Bo P stack the emergency disconnect sequence for the Bo P was activated from the bridge while lights changed on the control panel no flow was observed on the flow meters it appears that the emergency disconnect system did not function the lower marine riser package failed to unlatch from the Bo p and the Bo P did not seal the well with the Deepwater Horizon unable to disconnect from the well the order to abandon ship was given at 2200 hours this animated chronology has shown what we believe happened now we'll turn to an analysis identifying the contributing factors as you'll see we concluded that there was no single action or inaction that caused the accident instead we found that there were eight interrelated and contributing factors that led to this tragedy the first two findings of the investigation relate to hydrocarbons entering the well we believe that the evidence shows that both the annulus cement barrier and the chute track barriers did not isolate the hydrocarbons allowing them to flow up the well bore the next three findings stem from evidence that the rig crew did not observe recognize that hydrocarbons had entered the well before appropriate steps could be taken to contain them the negative pressure test conducted in order to confirm well integrity was incorrectly accepted an influx of hydrocarbons was not recognized until it was too late to prevent the hydrocarbons from reaching the surface because they had already entered the riser and well control response actions failed to seal and regain control of the well our 6th and 7th findings relate to evidence that hydrocarbons came up through the riser and were dispersed across the vessel creating the conditions for them to ignite we found that the well flow is diverted to the rigs mud gas separator causing gas to be vented onto the vessel rather than being diverted overboard then the vessels fire and gas system did not prevent hydrocarbon ignition our eighth and final finding was drawn from evidence relating to the blowout preventer with the loss of well control and subsequent explosions and fires the Bo P did not seal the well each of these findings is explained in greater detail throughout the remainder of this presentation our first key finding relates to the cement job BP relied on the expertise of Halliburton a cementing services provider to recommend a cement slurry appropriate for this well the BP team was responsible for assuring Halliburton's work and accepting its proposal it was clear from the event that the cement did not isolate the hydrocarbons the challenge facing the investigation team was to determine what happened without access to actual samples of the cement slurry used in the Macondo well therefore we had an independent laboratory analyzed the slurry design and also create and test samples based on that design through our investigation of the cement job we noted the following critical items first the high percentage of nitrogen in the foam cement made it difficult to create a stable bone slurry second there were no fluid loss additives used in the entire cement slurry third a relatively small amount of cement was pumped in relation to the displacement volume increasing the likelihood of cement contamination and fourth Halliburton did not conduct comprehensive lab tests that could have identified potential problems with the cement the independent laboratory constructed representative cement samples based on the slurry design and demonstrated that cement stability could not be achieved therefore we've concluded that the foam slurry likely experienced nitrogen breakout resulting in nitrogen migration slurry contamination and incorrect cement density our investigation of the cement job also included an analysis of BP's decision to use six centralizers and a long string production casing in the well we concluded that the long string design itself was sound and consistent with other wells in the area BP had 15 additional centralizers on the rig but they were incorrectly thought to be the wrong type although the six centralizers were placed across the main hydrocarbon production zones the decision to use six rather than 21 centralizers increase the possibility of channeling above the main hydrocarbon zone this matters only if the flow came up the annulus and through the seal assembly we believe that BP and Halliburton working together should have better identified and address the issues underlying the cement job improved technical assurance risk management and management of change by BP personnel could have raised awareness and led to better decisions regarding acceptance and implementation of the cement proposal similarly with improved engineering rigor and testing Halliburton could have better identified the reliability of the foam cement slurry and communicated the risks to the BP well team the investigation team's next task was to determine the actual point of entry for hydrocarbons into the we examined and analyzed every plausible entry point as a result our second key finding is that the initial hydrocarbon entry occurred at the shoe track and that it's mechanical barriers therefore did not isolate the hydrocarbons our conclusion is based in part on hydrostatic pressure calculations well flow modeling and an analysis of data recorded during the successful static kill of the Macondo well the shoe track has two types of mechanical barriers the cement in the shoe track and the double check valves in the float collar the investigation team believes that the shoe track cement failed to act as barrier the team has not established whether this failure was attributable to the design of the cement contamination of the cement by mud in the wellbore commingling of the cement with nitrogen due to nitrogen breakout from the phone cement swapping of the shoe track cement with the mud in the bottom of the well or some combination of these factors hydrocarbons were able to bypass the float collar check valves because they were damaged they never properly converted or they simply failed to seal the evidence shows that the flow came up through the shoe track at the bottom of the production casing therefore the decision to use SiC centralizers likely did not contribute to the outcome the third key finding is that the negative pressure test was accepted even though well integrity had not been established the negative test was conducted to confirm the integrity of the annulus cement shoe track casing and wellhead seal assembly looking back the investigation team believes there were several issues associated with this negative test first at the start of the test the annular preventer allowed approximately 50 barrels of spacer to pass into the Bo P this may have inhibited the pressure readings on the kill line second the bleed off volumes we examined were in excess of what should have been expected indicating communication with or through failed barriers and third Rik personnel found that drill pipe pressure had increased to 1,400 psi with no flow on the kill line the investigation team believes this reading should have resulted in further inquiry into the wells integrity instead the abnormal pressure was attributed to a phenomenon called the bladder effect ultimately the negative test was incorrectly deemed a success by BP and Transocean rig personnel a more detailed procedure would have helped BP and Transocean personnel to implement and interpret the negative tests the investigations fourth key finding is that the influx of hydrocarbons into the well was not recognized until it had entered the riser well control principles rely on early influx detection and effective response the investigation team examined real-time data from the rig conducted dynamic flow modeling and reviewed witness accounts in order to reconstruct the key indications of flow that we believe were not observed or recognized it appears that there were a number of indicators of hydrocarbon influx that went unobserved or unrecognized first from 2058 to 2108 the drill pipe pressure should have decreased instead drill pipe pressure increased by a hundred psi at a constant pump rate starting at 20 101 by 20 108 a 39 barrel gain had been taken second from 2108 to 21 14 the drill pipe pressure increased by 246 psi with the pumps off pressure should not have changed during this period and third by 2131 an estimated 300 barrel gain had been taken over the next five minutes the drill pipe pressure increased by 556 psi again with the pumps off these indicators should have been observable but it appears the first well control actions were not taken until about 21 41 by which time an estimated 1000 barrel gain had been taken the investigation team has not reached a conclusion as to why these indicators were not observed or recognized we note however that simultaneous operations were occurring on board the vessel at this time including mud pit cleaning and mud pit transfers which may have masked the indications of a hydrocarbon influx the fifth key finding is that well control response actions did not regain control of the well upon recognizing an influx the well must be shut in and if necessary hydrocarbons diverted safely away from the vessel the investigation team believes that the response on April 20th was not typical of a rig crew that was aware that significant well control event was occurring for example in retrospect there was about a 40-minute gap between the first indication of hydrocarbon influx and the first well control response roughly eight minutes later the first explosion occurred when action was taken the diverter was closed and the flow was routed to the mud gas separator also at that point in time it appears that an annular preventer was activated but apparently did not immediately seal the Bo P sealed about six minutes later the investigation team believes that the time for effective emergency response was significantly shortened because initial indications that hydrocarbon had entered the well went at observed or unrecognized the sixth key finding is that the diversion of hydrocarbons to the mud gas separator resulted in gas venting onto the rig in a well controllable dog of verdure can be closed and the fluids directed to either the overboard diverter lines or the mud gas separator based on testimony and interviews and an analysis of photographs and well flow data we believe the flow was diverted to the mud gas separator the mud gas separator is designed to remove small amounts of entrained gas but not the significant volumes experienced here large volumes would generally be expected to be sent over board via the 14-inch diverter line as gas dispersed across the vessel potential ignition sources were enveloped in a flammable vapor cloud the seventh key finding is that the fire and gas system on the Deepwater Horizon did not prevent hydrocarbon ignition gas dispersion modeling conducted by the investigation team shows that rapid spread of gas across the vessel covered the main deck and moon pool areas overwhelming secondary protective devices the investigation team believes that the gas likely entered the engine rooms through the main deck air intakes resulting in engines going into overspeed these engines were one potential source of ignition the eighth and final key finding is that the blowout preventer did not seal the well the investigation focused on the Bo Pease emergency systems and why they fail to seal the well after the initial explosion the investigation team identified issues that we believe interfered with the operation of the Bo P first the initial explosions and fire damaged the control cables called MUX cables and the hydraulic lines the MUX cables are critical to Bo P operation because they provide electronic community and electrical power to the Bo P control pods this is likely why the emergency disconnect system failed to activate the blind shear Ram when it was triggered by personnel on the bridge of the Deepwater Horizon second we believe based on testing down at the direction of Transocean after both control pods had been retrieved at the automatic mode function or a MF sequence did not complete the AMF should have activated automatically without crew intervention when the MUX cables and hydraulic line were damaged at least one operational control pod was required to activate the AMF however based on testing there was a defective solenoid valve in one control pod and insufficient charge on the batteries in the other control pod we believe that the ROV intervention to simulate auto shear conditions likely activated the blind shear Rams but they failed to seal the well and hydrocarbons continued to flow the investigation team has not determined the reasons for the failure of the blind shear Rams to seal presence of non shareable pipe across the blind shear Ram insufficient hydraulic power subsea and prevailing flow conditions are among the potential causes of failure the investigation team also found indications of potential weaknesses in the testing regime and maintenance management system for the Bo P the Bo P is a complex device and more may be known when it is fully examined in summary as I noted at the outset the investigation team concluded that there is no single action or inaction that caused the Deepwater Horizon accident eight safety barriers were breached the accident was the culmination of a complex and interlinked series of mechanical failures human judgments engineering design operational implementation and team communication these involved a number of companies including BP this presentation of the report's findings is intended as a high-level overview the full report is available on bp's website BP senior management has accepted all of the team's recommendations which address changes that should be made in drilling and well operations as well as contractor and service provider oversight finally in the course of conducting the investigation the team learned of many brave acts by people on the night of the accident I would like to close by recognizing those individuals for the lives they undoubtedly saved and by extending our sympathies to the families of the eleven man who died that night and to the community subsequently affected by the Spill

Channel: bp

Views: 465,779

Keywords: BP, British Petroleum, Gulf of Mexico oil spill, Macondo Well, BP internal investigation, BP investigation, Gulf oil well, BP report, BP probe

Id: zE_uHq36DLU

Length: 28min 56sec (1736 seconds)

Published: Wed Sep 08 2010