THE ROOT CAUSE OF THE MACONDO BLOWOUTA faulty understanding of fluid mechanics under static conditions, by George Baker, Technical Writer; Publisher, Mexico Energy Intelligence® & Energia.com and Ron Sweatman, Principal Advisor, Reservoir Development Services, Baker Hughes.
ON APRIL 20, 2010, THE LOSS OF CONTROL OF THE MACONDO WELL in federal waters in the U.S. portion of the Gulf of Mexico led to the loss of Deepwater Horizon (DWH) and the lives of eleven contractors and crew members. The resulting oil spill would cause the greatest ecological damage in the region since the blowout of Pemex’s well Ixtoc‐1 in 1979.
This report was prompted by the belief that much—if not most—of what happened that led up to the blowout on the Macondo well has not yet reached the public record. In the first five years following the event, with litigation and government investigations in progress, persons with a direct knowledge of facts were ordered by their employers not to speak about the incident. At a Houston conference in 2014, Kevin Lacy, former head of BP’s drilling operations in the Gulf of Mexico, told the audience that part of his testimony to government investigators is still under seal. In the light of the absence of a memoir by a DWH survivor or crew member of the nearby mud boat, it may be supposed that each person received an indemnity in exchange for his or her agreement keep silent about what happened.
At a prior Houston conference in April, Ronald Sweatman, an industry expert in cementing and fracking who served on the Macondo Task Force, drawing on a co‐authored paper, presented findings about the behavior of drilling fluids under static conditions that offered a fresh interpretation of the root cause of the accident. Owing to geothermally‐induced changes in the drilling fluids, a loss of density occurred that allowed formation fluids to penetrate the slurry within two hours after placement of the cement.
He observes that this important lesson from Macondo has not yet been absorbed by the drilling community.
This report, which draws on extensive discussions with industry sources, complemented by a reading of the Macondo literature, provides a brief, technical discussion of
critical events that took place in the Macondo wellbore on the day of the blowout.
INTRODUCTIONIN THE VERY REAL DRAMA of the explosions on Deepwater Horizon, the loss of eleven lives and an oil spill that took three months to control, the attention of investigators was focused on understanding the ￼reasons why the multiple barriers that were in place to protect the crew, rig and environment all failed. Attention was focused on the
blowout preventer (BOP), which mysteriously jammed and allowed hydrocarbons to pass through and up the riser to the rig floor.1
As for the first of the barriers, the cement that had been placed to seal the well against the pressure of the formation, it was an indisputable fact that the cement had failed. As BP’s initial accident report put it, “the annulus cement did not isolate the hydrocarbons.”i It was speculated that the foam cement had too much nitrogen, making the density of the cement inadequate to contain formation pressures. It was also speculated that the placement of fewer than the planned number of centralizers in the wellbore had made the cement barrier too thin in places. BP was criticized for not performing a Cement Bond Log (CBL) test to assess the evenness of the cement barrier.
Finger‐pointing and attendant litigation were everywhere: at Halliburton, which prepared the cement, at M‐I SWACO, which provided the mud engineers; at Cameron, the manufacturer of the BOP, at Transocean, the rig owner and at BP, the operator of the well. The operations offices of these companies became like military bunkers, with tight controls on the information that was released to the public and to investigators. Employees were told not to talk. Authors of technical papers were told not to mention Macondo by name.
Meanwhile, operators from across the U.S. Gulf of Mexico came to BP’s assistance, offering vessels, manpower and supplies. Onshore, at industry events, little was said. At the Offshore Technology Conference (OTC) held just two weeks after the explosion, an event to award BP for safety was cancelled. At a $50/plate luncheon event, a speaker from Shell spoke on the timely subject of “Challenges of Deepwater Operations.” His presentation, however, was about a project in the Philippines; nothing was said about the Macondo well that was flowing some 75 miles away.
An uncounted number of investigations were undertaken by government agencies, industry boards and company labs to determine how the sequences of mechanical and human failure could have happened. The results of one of these investigations is the topic of this report.
One team of investigators asked a new question: Suppose that the cement would have served as an effective seal, as planned, were it not for unanticipated conditions in the Macondo wellbore. Were there conditions in the Macondo wellbore that could have led to the failure of the cement as a seal?
This line of inquiry led to a new hypothesis about what happened in the wellbore on April 20.
BACKGROUNDLONG BEFORE 2010, it was known as a general fact of fluid mechanics that all fluids to some extent are compressible and can shrink or expand; that is, changes in pressure or temperature will result in changes in density. It was known that a drilling fluid called synthetic‐based mud (SBM) is highly compressible with a much higher coefficient of thermal expansion compared to water‐based mud.
The oil industry had been aware that during drilling operations with SBM (which was used as the drilling fluid on the Macondo well) shrinkage could occur; but there was no way to differentiate actual shrinkage from the static losses of drilling fluid that occurs when fluids have leaked into the formation that surrounds a well. In either case, the correct response would be to insert mud into the wellbore. If the mud crew fails to make this response during drilling or cementing operations, the risk of a blowout is created, as shrinkage brings with it a loss of hydrostatic pressure. During cementing, a loss of overbalanced pressures compromises the integrity of the cement during the curing phase of 14‐21 hours.
In 2010, the science that explains the need for an open fluid column during the curing phase was only partially understood. Missing then was a precise understanding of the effect that changes in wellbore temperature have on volume, hydrostatic pressure and density. In the
standards set forth today by the American Petroleum Institute (API), an explanation of the need to maintain an open fluid column during curing is only minimally presented; some readers may be left with the mistaken impression that hydrostatic pressure would be unaffected by isolating the wellbore from the riser.
DiscussionON MACONDO, the installation of the casing seal at the wellhead shortly after the placement of the cement created a closed system, which meant that the hydrostatic pressure would now respond to changing conditions in the wellbore. These changes would be undetectable by the crew.
The higher temperature of the rock formation affected the properties of the fluid: As the SBM
in the annulus cooled, a thermal‐induced shrinkage led to a loss of density. The effective
mud weight fell below the formation pressure of the five flowable zones (Fig. 1), resulting in a loss of overbalance pressure in the trapped fluid column (from the wellhead to the hydrocarbon‐bearing formations).
During first two hours, when the cement was still a slurry, this progressive loss of density became an undetected risk. The equivalent density that exerted back‐pressure on the
formation fluids became less than the pore pressures in the rock formations, creating an underbalanced condition (Fig. 2). Formation fluids, brine, oil and gas, penetrated the wet cement, creating flow channels into the weak zone below.
Over time, the flow channels grew larger, eventually building up sufficient pressure to penetrate the mechanical barriers. The influx of formation fluids came from below, through the shoe track, penetrating and passing through the float collars and up the riser.
During this process, the DWH crew received no data or warning that the cement would
not serve as a barrier. Today, there still is no commercial software product that would
enable the thermal modelling of the behaviour of drilling fluid in cases like this where geothermal gradients could affect fluid temperatures and pressures; nevertheless, there are measures that may be taken to cancel the effect of temperature.
For example, the fluid’s temperature vs. the geothermal gradient may be made more stable by a complete top‐to‐bottom conditioning of the mud, a precaution not taken on the DWH. In some cases, mud temperatures may be adjusted on the rig to lessen the downhole effects. In addition, the setting of the casing seal should be delayed until after the cement sets hard enough to create a pressure barrier.
OBSERVATIONSLOOKING BACK, neither drillers nor operators understood much about the changes that the formation could induce into the drilling fluid. They could not appreciate the danger that such changes could lead to an underbalancing of the well. It was this undetected change that occurred at the Macondo well that, combined with subsequent mechanical and human failures, led to the blowout and loss of the rig.
Some seven months after the Macondo blowout, in December 2010, the API published an update to its industry guidelines for isolating potential flow zones during well construction.2 RP 65 (Part 2) notes correctly that “temperature has the single greatest influence on cement slurry performance” (Sec. 5.6.4); but there is no explicit warning or explanation about what could go wrong. The text goes on to observe that “accurate estimates of cementing temperatures (both static and circulating) are essential to the success of the cement job”; but why they are essential is not explained.
As has been noted, temperature as a parameter that affects density must be considered along with hydrostatic pressure. Drillers are encouraged to carry out well procedures following the placement of cement “in such a way that they will not disturb the cement and
damage the seal or cause the seal to set improperly” (Sec. 5.10.2); but the primary lesson of Macondo‐‐that the cement seal will be compromised by prematurely installing the casing seal assembly‐‐is not mentioned.
The connection had not been made between a) the gradual adjustments of the temperature of the drilling fluid to the geothermal gradient and b) the effect that such adjustments could have on the integrity of the cement seal. This lack of understanding puts drillers and operators today at as much risk of another incident as were their peers on DWH in 2010.
Looking forward, the drilling industry, supported by its university educators and government regulators, will need training tools that will teach them to anticipate geothermally induced effects.
Commercial software is needed to allow for routine thermal modeling in which the undetectable changes in the fluid properties and the hydrostatic pressures are calculated as the fluid temperature equalizes with the surrounding formation temperature. With the aid of such software, profiles over time may be plotted that show formation‐induced changes on the properties of the drilling fluid. Such profiles would tell you if such changes are induced by the heating or cooling of the fluid. At Macondo, it was the cooling effect that led to the compromise of the integrity of the cement barrier.
In the meantime, the curricula of training schools, plus API standards, need to be updated to include instruction and guidance from this poorly understood risk to drilling and cementing operations.
1) Shortly after the cement was placed in the well, there were unanticipated reductions in the hydrostatic pressure in the wellbore caused by cooling in the annular column of drilling fluid. The initial overbalanced pressures across flowable formations were transformed into underbalances while the cement was still a slurry.
2) These changes became undetectable when the casing seal assembly was installed soon after the placement of the cement.
3) This understanding did not exist in the drilling community in 2010.
There is need for renewed research on how changes in wellbore temperature and pressure can affect the curing process of cement and its effectiveness as a seal of formation fluids.
Commercial software is needed that models hypothetical effects on density from such changes. It would be especially important in a wellbore where there is a narrow window between pore pressure and fracture pressure.
The Macondo blowout and its consequences offer a university of lessons on fluid mechanics, drilling, process safety, leadership and public oversight. If all goes well, the lessons to be learned will be incorporated into government regulations and the curricula of universities and industry‐run training schools. Eventually, the Macondo accident will no longer be needed as a point of historical reference.