A drilling perspective on what went wrong at BP on 4-20-10, page 1

TL;DR? I'll post a 2-3 paragraph summary soon.

I interned and worked at an oil drilling company for about 8 years now, this is a paper I wrote on the spill. It's a child of in-office debate we had about what really happened. Some of my co-workers knew people on the platform, so we had alot of insider info to make use of.

I finished this a little less than a week after the spill. I sent it to WSJ when I finished it and they quickly changed their report from depicting a methane bubble to bad cement. They didn't cite me for it, but they did use my terminology, explanations, and cite "those with knowledge in drilling."

There is alot of information that came out after this, alot of which supports the argument, and some of which is simply exclusive. I will comment on them both below the essay.

As to why these things happened, that is open to interpretation. Personally I think someone was stoned (4/20??) and made a mistake, all the while they were already running a very dangerous program which I don't think you need worry about being repeated elsewhere, especially after this incident. It's important to note that cutting these corners were not as cost-effective as BP planned they would be, they are reaping what they sow. All the while other big drilling companies (don't worry about us small ones we cannot afford to make even the smallest mistake) look on and say "let's make sure we never get into THAT mess." Point is, please don't turn on the hand which feeds you oil, such as me and other oil engineers.

The Deep-Water Horizon Incident

On April 20, 2010, the Deep-Water Horizon oilrig crew was given a five-minute warning to evacuate their oil rig when they discovered that a blow-out was imminent. The rig crew detected a substantial "kick", or a blowback of drilling fluid, in the well that they were working on. Regularly the rig crew would simply equalize the pressure inside and outside of the well by pumping fluids down into it. However, this warning came very late and their pumps would not have time to equalize pressure before the oil and gas hit the derrick. Drilling fluids, gas, and oil would be forced upwards out of the pipe; any sand or debris in the mixture would then spark with the rig's metals causing a tragic explosion and even a loss of life.

The deep-water Horizon is one of the most high-tech oil derricks in the world, and run by some of the best contractors for the job. There is no doubt amongst those in the drilling industry that this catastrophe should not have happened; that they should have had more than a five-minute warning. Consequently, that they should have been able to equalize the pressure stopping the oil and gas from reaching the derrick. Like any disaster on this scale, there is no one reason as to why it was so catastrophic. Likewise, there are many reasons as to why the Horizon received only five minutes of lee-way to equalize the pressure. The question is, "what happened?" Although an official report is months from being written and released, we have received enough critical clues to deduct to a high probability what some of the general problems were.

The well that blew out at Horizon was not yet a completed well. It had one more stage to go through in order to be producing oil. At any stage before a well is finalized, the rig crew has the well's piping filled with drilling fluid. This drilling fluid can be altered to increase or decrease its weight. When a crew receives a "kick" it means oil and gas is being forced into the well and pushing drilling fluid out the top of the well. To counter a kick, the rig crew increases the weight of the drilling fluid, equalizing the pressure between the inside and outside of the well. Kicks are relatively common in the drilling business, and equalizing the pressures is standard procedure, which results in no blowout when the procedure is completed. In Horizon's case, they had a five-minute window to equalize rather than a more common 12-48 hours to control the well. This is not nearly enough time to increase the density of drilling fluid. Horizon could not complete the procedure of equalizing the pressures with the drilling fluid method.

The possible culprit most media outlets are blaming the disaster on is a methane bubble (msnbc.com, “Oil leak”). According to the ideal gas law, specifically Boyle’s law (Britannica, “Boyle’s Law”), a gas bubble increases in volume as the pressure decreases. A gas bubble that is traveling up from the depths of the earth undergoes less and less pressure until it reaches the atmosphere. Thus, any gas bubble traveling up a drill pipe expands as its depth decreases. This expansion displaces more and more drilling fluid as it travels up. The drilling fluid that is displaced comes out of the top of the well thus notifying Horizon's crew of the gas bubble. However, Horizon's blown out well was practically horizontal at its end, and only vertical towards the surface. This type of well is called “slanted hole”, and is common in the drilling industry. What this means for the gas bubble is that it would not be expanding until the last stretch of pipe, and the crew would not be notified of the gas bubble until then. This is one possibility as to why the crew only had a five-minute warning. However, this possibility is a fringe one. It is largely based on the unlikely scenario where no liquid got into the drill pipe with the gas. Oil and gas are generally together in the rock formation, and methane gas is even liquid at these intense pressures. If any of these fluids at all were to get into the pipe with the bubble, it would send a Kick relatively instantaneously. The crew would then be notified hours before the methane bubble reached its final stretch, or even before it became gaseous, giving the rig crew plenty of time to bleed out the Kick. But even a methane bubble does not fully explain the short window of time the crew was given. Regardless if a methane bubble was the possible culprit, there must have been other reasons as to why they received only a five-minute warning.

Horizon's plan before the warning was given was to move on to another well and come back to it days, maybe weeks, later to 'perforate' and finalize the well which now is known to be faulty. At this stage, before perforation, the piping would completely surrounded by cement. For any oil or gas to get into the well it must first go through the cement. We also know that the crew had finished cementing the last string in the faulty well, when they received the warning of the blowout (Walt Cunningham). This is the biggest clue as to what happened at the Horizon incident. The faulty well had not yet been perforated, but it had been cemented. There is no doubt that the cement was compromised in Horizon's well, it is only a matter of why. Getting through the cement is quite a task if the cement is properly set. However, cementing at these variable and intense pressures under the earth is very complicated, and it is relatively common for the cement to not set right and for the crew to be required to re-cement. The most likely reason the cement had been compromised is that it didn't set properly. But this does not explain why the rig crew did not know that the cement was compromised. The rig crew by law must make sure the cement is properly set before proceeding.

A finished well is typically made up of telescoping "strings" of steel piping. First, a big hole is drilled, and piping is put down in it and cemented in place. This is the first "string" of the well. After that, a second slightly smaller pipe slides down inside the bigger pipe and cemented/hung from the bottom of the bigger pipe. This is the second "string" of the well. This process is repeated until the well has typically somewhere between 2-4 strings.

Small gaps are left over in between the larger and smaller pipe, in order to make sliding the smaller pipe down inside easier. One of the many jobs of the cement is to act as a sealant in this gap. In a well with cement that didn't set right this gap is not going to be sealed. Thus, any oil and gas formation which the bad cement intersected, would travel through the bad cement like a sponge, possibly for thousands of feet, until it reaches this gap. The oil and gas would then flow inside the well pipe of the previous string. Before perforation, drilling fluid is the only gauge of whether or not oil and gas are getting into the well pipe. This leak would have virtually no effect on the drilling fluid until it reaches three quarters of the way up the well. This is one reason that the rig crew had such a shortened window to bleed out the Kick.

Another possible reason that the cement was compromised is that the cement was set properly, but was then put under tremendous stress which cracked the cement. To analyze this possibility, we must first take a look at how the engineers know the cement is set properly. Engineers have ‘loggers’ that they lower down into the pipe that checks for conductivity as it descends. After the cement is poured and given time to set, a log is always taken in order to ensure that the cement is properly set before proceeding further. According to the engineer who analyzed the log of the blown-out well at Horizon, the log showed that the cement was set properly. This tells us that the cement was fine, but then some event unfolded which stressed the cement too much and cracked it. Thus causing oil and gas to flow along the outside of the lower pipe, between the steel casing and open hole, through the cement sealing or “lap” between it and the pipe above it, and into the next string up. The drilling fluid wouldn't notify the crew before the oil reached a few thousand feet from the surface, having already picked up tremendous speed while it was undetected. This is another reason that the crew had only a five-minute warning, but this still does not entirely explain it.

Two rig crewmembers of Horizon have recently come out to give a huge clue as to why the cement was compromised. Horizon's initial plan was to drill a group of wells and cement them. The crew would then come back later to perforate the wells and extract the oil. The crew also decided to re-use a portion of their drilling fluid from each well on the successive well. It is not uncommon for drilling rigs to do this, as the drilling fluid is very expensive and this is one of the many safe ways of preserving it. This is even truer for a rig that is drilling in such deep depths such as horizon; they can save thousands of feet worth of drilling fluid. The drilling fluid that is taken out would be replaced with sea water, which is much lighter but is still dense enough to keep the cement from cracking if not too much is put in. However, in order to leave some of the drilling fluid in the well, a cement plug needs to be made which prevents the seawater from mixing with the drilling fluid in the lower portion of the well. Not only is the plug necessary to make sure the drilling fluid is preserved from contamination with seawater. The cement plug also serves as a safety precaution, without the plug, the seawater could potentially fill the entire well. This would create a vast density difference between the inside and the outside of the well and could compromise potentially all of the cement in the well. In addition to cracking the cement, the sea water would act as a vacuum sucking up oil and gas out of any compromised cement and into the well. Mr. Probert, who was on Horizon's rig, said in a statement, that "Prior to the point in the well construction plan that the Halliburton personnel would have set the final cement plug, the catastrophic incident occurred. As a result, the final cement plug was never set." (Wsj.com, “Link rig blast to ‘Plug’”) Horizon's cement plug was never set. This fully explains the extremely short 5-minute window from the warning until the blowout.

However, the Horizon had a state-of-the-art blowout preventer that did not activate. If it had activated successfully, no blowout would have happened. All wells, by law, must have a blowout preventer attached to them as a fail-safe, in case a blowout was to happen. In order to understand why, we must first take a look at how blowout preventers work. A typical blowout preventer has three stages, each one being more destructive to the well. The first mechanism is non-destructive to the well. This is a simple bag that blows up which blocks any passage way for fluid to come up between the well and the derrick. The crew then pumps heavy mud into the well through the drill pipe (a smaller pipe inside the well during drilling) and bleeds out lighter mud in attempt to suppress the kick. If that fails, various steel doors shut, serving the same purpose as the bag, only with more strength. The next, last resort, is a shear-ram which cuts any drill pipe (not steel casing) which could be in the way and acts like the rock, which would normally be over the oil in the rock formation.

If there were some uneven amount of stress pushing in on the outside of the steel casing under the blowout preventer, Hooke's law states that the casing would deform and become slightly longer (Britannica, “Hooke’s Law”). Over Horizon's 18,000 feet of well, even the slightest deviation in stress could stretch it's casing a few feet. However, Horizon's well after being filled with sea water would have been under substantial forces of compression, such that the casing could stretch tens of feet. If the steel casing were to stretch far enough up to get in between the shear-ram of the blowout preventer, then the shear ram would not be able to fully close. When the shear ram closes on steel casing it would only dent the steel casing, not making the full seal necessary to cap the well and prevent it from blowing out. Most shear rams are designed to shear drill pipe, which is made of much weaker material than the steel casing. Very few blowout preventers are designed to cut through steel casing. The ones that are, still, are not designed to cut it in the specific scenario of the steel casing stretching up into the shear ram’s way. However, the combination of the length of Horizon’s well, and the substantial stress the well is under are both firsts in the drilling industry. In addition, Hooke’s law is a very hard law to calculate and account for due to the negligible resulting effects of the deformation. So engineers designing blowout preventers most likely did not predict that steel casing could ever reach the shear ram. They also probably did not change their design, which has been working for 50+ years, for deep-water drilling which is still in its infancy.

A recent article published on Yahoo states that the problems were due to “a leaky cement job, a loose hydraulic fitting, a dead battery.” (Yahoo.com, “Oil Spill: Smaller Tube”) We know that the cement was compromised as a result of the plug never being set. Yahoo is blaming this on a lack of over-sight from the Mineral Management Service. But the MMS tests blowout preventers on an 8-14 day interval (mms.gov, “Testing Requirements”). The likelihood of the hydraulics failing independently and also in the same week as the Kick is very remote. However, the expansion of the steel casing could rip the hydraulic seal, packers, and locks where the casing is hung at the BOP. Preventing the shear rams from activating in the first place. It is most probable that the hydraulic failure was dependent on some event which happened during the Kick, and not an independent incident. Most likely this event is the stress that the hydraulic seal, packers, and locks are under when the steel casing expands by Hooke’s law.

The incident at Horizon was a perfect storm of problems, which led to one big disaster. If one of the problems was not there, the blowout could have been avoided entirely. The short window the crew was given for bleeding out the kick was the most catastrophic problem. The short window of warning was mostly caused by the failure to set a cement plug before taking the heavy drilling fluid out. The intense pressure difference this caused between the inside and outside of the well cracked all of the cement, all the way down to the production zone and including the seal between the connecting strings of casing. The oil and gas then entered through the string's cement seal into the casing, all while the sea water acted as a vacuum sucking up oil and gas. The same force that cracked the cement also stretched the steel casing far up into the blowout preventer's shear ram, preventing the shear ram from closing and stopping the blowout. It's most probable that there was a miscommunication between the operator and the rig crew; that the operator giving the command to fill the well did not know that the plug had not been set. It would only be speculative to say why the operator did not know, but there are many possibilities. The incident is likely not one person's fault. Drilling personnel around the world are now reminded of what the consequences are when they do not take the usual proper precautions. In addition, new and improved technology for blowout preventers and well designs for this hostile environment should become available and advance the prevention of any future blowouts.

Gold, Russel, Steven Power, et all. “Two Oil Firms Link Rig Blast to ‘Plug’.” The Wall Street Journal. May 11, 2010. Web. May 11, 2010. .

"Hooke’s law." Encyclopædia Britannica. 2010. Encyclopædia Britannica Online. 11 May. 2010 .

"Boyle’s law." Encyclopædia Britannica. 2010. Encyclopædia Britannica Online. 13 May. 2010 .
Donn, Jeff, H. Josef Hebert, et. All. “BP's next try to stem oil gusher: Smaller tube.” Yahoo! News. 13 May. 2010. Web. 13 May. 2010.
Burdeau , Cain, Vicki Smith, et all. “Deep-sea Ice Crystals Stymie Gulf Oil Leak Fix.” Msnbc.com. 3 May. 2010. Web. 7 May. 2010. www.msnbc.msn.com...
“MMS Issues Final Rule on Preventer Testing Requirements.” MMS. 10 May. 2010. < www.mms.gov...>

The new information

Some new information since the paper was written, the casing was not centered in the well! This is an excerpt from an e-mail i sent to a friend. I'll put it all there in case people want a compacted version of the above.

To give you a quick idea, the well wasn't perforated yet so NO oil should have gotten inside. There was 14,000 feet of steel casing surrounded by cement, literally nothing should get in at this point. However...well, for one, I should say...They did not do a complete cement job. When you cement a casing you have to centralize it in the open hole so that the casing is in the center of the well. Haliburton (the cement company) recommended 18-24 centralizers in the well. BP used 6!!, that's atleast 1/3 of the centralizers needed! The casing was not centered inside the hole. One could get away with this, realistically. But it leaves you no margin for error whatsoever. The casing must be centered or one side of the casing could be resting up against the rock formation and have no cement on that side of the casing at all.

On to the catalyst...what was the straw the broke the camels back? Being a huge venture, BP drills maybe 20 wells at a time all phased together. For example they drill 20 open holes one by one, then they cement them one by one, then they perforate them one by one. This is common procedure for a big company. But BP in particular was also taking some of the heavy drilling fluid out from the preceding well to use in the next well. This is also semi-common procedure. Drilling fluid is very expensive and it can save you a ton of money if you do it safely and smartly. The drilling fluid has many purposes, but the most relevant purpose is that it equalizes the pressure inside and outside the well. Obviously you cannot take all of the drilling fluid out of the well or you could compromise the cement and the steel casing under the enormous pressure of the earth. So you put a cement plug around 2/3s up the well, and take out only the top 1/3 of the drilling fluid in the pipe and replace it with sea water. British petroleums green and "safe" rig crew skipped a step in this process. They did not set the cement plug, the filled the entire well with sea water. Although this is really stupid and they were already cutting corners, you have to note that this plug WAS in their game plan but just had a mis-communication or something. Albeit a man-slaughtering and costly mis-communication.

But even still, this doesn't explain the blow-out. There is a tri-redundant blow-out preventer at the top of all wells in construction by law. And this BOP must be tested every 8-14 days by law, so the BOP was working 8-14 days prior. So what was it? It was an unforseen design error in the BOP. Drilling in 3000 feet of water is a rather new invention, but we were using the same old technology. Not in the sense that we just picked them up off our 200 ft wells and put them down there, I mean like design technology. They had beefed up versions of old technology. The good news is that it is easily fixable with a new design of BOPs. The problem was that this 14,000 ft of casing was under so much pressure, that the casing stretched maybe 20-30 feet. This could do one of two things, it could either push steel casing up into the dead-man valve which is not designed to cut such strong material (it might dent it but not cut it). Or it could break seals inside the BOP preventing it from ever even trying to cut it. I've heard reports of hydraulic leaks, so it must be the latter.

Why wasn't Hooke's Law taken into account in the BOP?

This could easily be the first well in history that needed it to be taken into account, and I just don't think that anyone ever thought of putting hooke's law into account. I have never heard of any well being completely filled with sea water.

This well was 14,000 feet long, and with ALOT of water on top of it too. Our wells down here are in 100 feet of water (less than 30 times that of Horizon) and our wells only go to 5000 feet at the most, and thats not true depth thats length.

Somewhere between 30-50 years of proven BOP technology, the idea of starting to put these BOPs on the ocean floor (it would not have happened had it been at the derrick), and gradually getting deeper and deeper wells. Had all made this problem kind of sneak up on the engineers IMO.

The leak would definitely be inside the BOP, there are a few seals inbetween the actual casing inside the BOP that are essential to the deadman valve and it's hydraulics working properly. An abrupt stretching or even throwing of the casing could easily break one of these seals. One of my co-workers actually believes the pipe was thrown not stretched to break the BOP. Which is an insteresting plausibility, sometimes blowouts can throw piping a thousand feet or more into the air. You basically have a 14,000 foot long barreled rifle when a blowout occurs. All the energy is very focused in one direction. If you look hard enough I'm sure you could find pictures of the drill pipe a fwe thousand feet in the air after a blowout.

But the hydraulics were definitely bad for one reason or another, the only question is how. They are tested once every week or two by law, and so it must have been working recently. And it is most likely that the leak was caused by some point in the catastrophe.

edit on 1-11-2010 by venik88 because: (no reason given)