The Genesis of the Deepwater Horizon Blowout Full Report
Editor’s Note: Theories on the cause of the Deepwater Horizon disaster are a dime a dozen. Everybody has one. The problem is that all the theories widely reported in the media and on the internet are wrong. Even the most extensive analysis done on the disaster yet, BP’s own Deepwater Horizon Accident Investigation Report makes wrong assumptions and draws various erroneous conclusions by misinterpreting key events. In the attached report, Phil Rae proposes a theory that does the one thing none of the others has – it fits the facts. Studying the report is the only way to understand the nuances of the disaster and, especially, to prepare to avoid its re-occurrence.
In our book, ” The Energy Imperative” earlier this year, my co-authors (Leonard Kalfayan and Michael Economides) and I commented on the Deepwater Horizon disaster, which was unfolding at the time. It was rather natural, therefore, for me to take an active interest in the public and political hysteria that ensued. Despite the furor, the finger-pointing and investigative hearings, no satisfactory explanation emerged for several reported events that appeared to defy the laws of physics. Particularly puzzling was the lack of an adequate explanation for the pressure and flow anomalies during the so-called negative test, the critical and final check to ensure the well was secure and could be circulated to seawater. However, by careful analysis of the data available in the public domain, including official investigation transcripts, blogs and documents like the BP Deepwater Horizon Accident Investigation Report I have arrived at a theory that offers the only logical explanation. In fact, it provides not just an explanation for the anomalies but also a unique insight into the genesis of the blowout.
Many individual factors almost certainly came into play and contributed to the disaster but there was one final critical factor that set the whole train of events in motion, sealing the fate of the Deepwater Horizon that day. That critical step was the high rate displacement of drilling mud by an unusually large volume of high viscosity, heavyweight spacer, followed by seawater, in preparation for the negative test and, ultimately, well suspension. During the pumping and displacement of this heavyweight spacer, a breach in pressure integrity at the casing shoe resulted in the undetected loss of about 80 barrels of drilling mud into the probably uncemented annulus. This, at the time, undetected loss of mud resulted in under-displacement of the heavyweight spacer and led to otherwise inexplicable pressure and flow anomalies during the negative test, induced by U-tubing and other phenomena, including flow from the well.
The net effect of this undetected loss and under-displacement was to leave an estimated 80 barrels of heavyweight spacer in the well, along much of the drill pipe (DP)/casing annulus and critically across the blowout preventer (BOP) and kill line (see Figure A). It was this under-displaced spacer that provided the unexplained extra 700 psi of differential pressure and the anomalous flow volumes that were to persist throughout the negative test, confusing the crew. It was this same under-displaced spacer that caused U-tubing of fluids, due to density differentials, decoupling the fluid column in the kill line from surface and rendering the pressure gauge there ineffective (See Figure B). The partial evacuation of the kill line allowed heavyweight spacer to flow into the lower end of the kill line when the well was shut-in, rendering the kill line useless for pressure and flow monitoring for the duration of the negative test (See Figure C). The flow of spacer into the kill line left the line with a combined hydrostatic pressure that was in equilibrium with formation pressure acting in the wellbore. Hence, later in the test when the kill line was left open, there was no flow from it, despite a 1400 psi differential pressure between it and the drill pipe surface pressure (See Figure D).
The theory neatly explains all the key anomalies during the negative test and provides new interpretations to one or two key events, in particular the much reported and generally accepted belief of a leak in the annular BOP at the start of the negative test (See Figure C). At that time, the kill line and DP had been shut in after bleeding pressure down but, in the space of only 6 minutes, pressure on the DP side quickly rose to 1250 psi. Sealing pressure on the annular preventer was increased and, with a leak suspected, the riser was topped up with an estimated 50 to 60 barrels of mud. It should be noted that this was only an estimated volume – no measurement was made. In fact, according to this analysis, there actually never was any leak on the Hydril and no spacer had leaked beneath it. The spacer had been there all along and the mud that topped up the riser was simply replacing the mud that had been silently lost into the bottom of the well during the initial spacer displacement and had gone unnoticed until only now. However, this mistaken, but widely-accepted, belief in a leak, confirmed by the fluid that had apparently escaped from the riser through the Hydril and that required topping-up, conveniently provided some explanation for the rapid pressure rise on the DP. In fact, the pressure rise came about because the well was flowing.
Exactly why the well breached at the shoe, during the displacement of the mud with viscous spacer and seawater, remains a mystery. The volume of spacer, its viscosity, the high pump rate and the configuration of the tubulars may have combined, in some yet undetermined manner, to expose the well to forces that it could not handle. It is possible that momentum changes caused some transient overpressure at the bottom of the well or it may simply have been that the well’s pressure integrity, as determined in the positive test, was unreliable. Whatever the cause, on this occasion, pressures that the well had successfully contained previously under static conditions, proved too much. Something gave way and the well lost fluid through the shoe track. This loss went unnoticed because at that same time other operations were ongoing. According to the BP Report, the trip tank was being cleaned so flow rate data during this time was considered unreliable. The exact moment when integrity was lost is unknown but there are possible indications on the mud log data of abnormal changes in flow rates and pressures during this period.
One way or another, the well’s integrity had been compromised and, In effect, it was no longer secure. This breach at the shoe and the unnoticed and unsuspected loss of a significant volume of mud, at a time when the well had already demonstrated positive pressure integrity, was the event that was ultimately to cause disaster. Anomalies that would have raised concerns in open hole were explained away by other suggestions and ideas in a well that was lined with steel pipe, cemented and already positively pressure tested.
The above explanation and interpretation of the chain of events is the only one that matches the observed behavior of the well and all its components during the negative test and the only logical reason for the kill line and DP pressures and flows to behave in the way they did. The only way for a sufficient volume of kill weight fluid to get into the kill line was by its being void of lower density fluid (i.e. seawater), at some point. The U-tubing initially induced by the under-displaced heavyweight spacer falling in the DP/casing annulus simply by gravity, due to the large density differential, would certainly have voided portions of the initially seawater-filled and hydraulically-coupled kill line. The heavyweight spacer, in turn, was under-displaced because, while displacing it, the well had ruptured, for whatever reason, through the casing shoe track, with loss of mud into the uncemented, or badly cemented, production casing annulus. Interestingly, this posited loss of ~80 barrels of mud exactly matches estimates made by others of an unexplained ~80 barrel discrepancy in mud volumes based on analysis of mud log data. This rupture alone may have provided the flow path for hydrocarbons back into the well or could have helped establish a flow path via the casing hanger by providing extra unbalanced lift forces from beneath it. Since it’s beyond the scope of the analysis, I make no attempt to identify which of these flow paths was responsible for hydrocarbon ingress and ultimate loss of well control and, indeed, both may have contributed in the disaster.
The resulting blowout killed 11 men, caused the sinking of the Deepwater Horizon and spilled almost 5 million barrels of oil into the Gulf of Mexico, demonizing BP, the biggest US domestic oil and gas producer since 2001, and precipitating a public backlash against the entire oil and gas industry. An unsupportable and, in fact, unsustainable government moratorium on deep water drilling has only recently been lifted in tacit recognition of the reality that deepwater production now accounts for 30% of US domestic oil production, and, of necessity, that share will continue to grow. Undoubtedly, new safeguards and tighter legislation will be introduced to minimize the risk of any similar catastrophe occurring in future but some of these measures will increase costs without necessarily improving safety. As the Deepwater Horizon saga confirms, things can still go wrong, not necessarily because of efforts to cut corners or reduce costs, as the media and politicians alike have repeatedly trumpeted, but because components can fail unexpectedly and human beings can make mistakes. Air travel provides us with ample evidence of this. In the past 10 years, there have been, on average, 29 accidents and 775 deaths every year in commercial civil aviation and the main causes are pilot error (~50%) and mechanical failure (~22%).
In the E&P industry, further improvements can and must be made in training people and it should be mandatory that critical procedures like the negative test are supported by clear and concise documentation that fully explains the rationale for the procedure and the criteria for acceptance. Anomalous results should mandate a temporary halt in operations to allow consultation with experts rather than relying on anecdotal explanations and unsupported hypotheses. Well construction plans should still maintain the operational flexibility to make essential changes when faced with unexpected pressures or other events. However, critical decisions e.g. to run a long string versus a liner or to use foamed cement rather than something more conventional, should always involve consultation with relevant technical experts and a clear understanding of the implications of such decisions. Despite the outcry, the Deepwater Horizon disaster will not put an end to deepwater drilling and production but, hopefully, the lessons learned from it will help avoid such accidents in the future.
Go here to read the full report.
Phil Rae is the co-author of The Energy Imperative. He is also an internationally recognized expert on well completions with extensive industry experience.