deepwater decisions

It should hardly come as a surprise that a great many e-mails circulate around the interwebs regarding the Deepwater Horizon and the circumstances leading up the explosion. I recently received something that I assume is public record and that would be the letter from Congressmen Waxman and Stupak to BP CEO Tony Hayward dated June 14, 2010. If you want a copy, let me know.

Before I get into the letter, along with many legitimate articles and investigation pieces, there is a great deal of false information running around. Some of it takes the form of sloppy reporting and some takes the form of nonsensical rumor-mongering. (To be clear, the claim that SLB paid for a private helicopter to fly personnel off of Deepwater Horizon is incorrect). Ah internet, you will always be strange and mystical things to people who do not know much. That misinformation can spread so pervasively is the reality of the world, but it is hardly a new phenomenon, just a much faster phenomenon now.

The letter from Congressmen Waxman and Stupak addresses five key points:
1) The decision to use a well design with few barriers to gas flow
2) The failure to use a sufficient number of “centralizers” to prevent channeling during the cement process
3) The failure to run a cement bond log to evaluate the effectiveness of the cement job
4) The failure to circulate potentially gas-bearing drilling muds out of the well
5) The failure to secure the wellhead with a lockdown sleeve before allowing pressure on the seal from below

As I have written before, I do not have special or “inside” knowledge of Deepwater Horizon and the events surround the explosion of the rig. I only have my own experience and I most definitely do not represent an official opinion. I am simply slightly better informed than the general public on the matter.

I want to deconstruct the five points in the letter. I am not trying to suggest that mistakes were not made or that best practices were followed. I merely want to provide an explanation of why some of the decisions may have been made and give some more context. It is clear from the general media and reporting that the decisions are generally being portrayed as beyond reckless and driven by financial considerations. Well, yes. This is a business. Almost all decisions are driven by financial considerations, and even seemingly ethical or ‘soft’ factors can be given a price. However, I don’t want to go down that path of what a human life is worth right now. I just want to note that businesses are in the business of staying in business. Yes? But that does not mean BP did not weigh considerations beyond the most basic ones being seen on the surface.

Point 1 on using a design with few barriers is a reference to run a single string of casing (which was actually in two different sizes but that’s not important) instead of a liner. The single string of casing is a reference to the metal pipe used in the well. Casing (or liners) are run into a newly drilled section of the hole and then usually cemented in place. The cement isolates the various formations (which contain oil, gas, water, etc) from each other and also keeps the casing/liner in place. A single-string refers to having casing in one continuous length that runs from the bottom of the well all the way to the sea floor. A liner would only run from the bottom of the well to somewhere inside the previous casing/liner. On a final section like this one, a tieback section of casing would later be run from the top of the liner to the sea floor and that would also be cemented in place. The liner and subsequent tieback would take significantly more time to run than a single string of casing mostly owing to the time it would take to set the liner, wait on cement, run the tieback, cement the tieback, and wait on cement again. However, a liner carries with it additional operational risks. There is an increased risk of placing little to no cement in the open-hole annulus at all. This is due to possible failures that can occur at the top of the liner where darts and plugs (used to wipe the inside of the pipe and separate fluids that are pumped) can get stuck in the liner top assembly. This is not necessarily a high risk, but I have seen some operators express enough concern that they do not run darts and plugs, but instead another method of pipe-wiping and fluid separation that is generally less effective and carries its own set of risks. What is the risk of a liner failure where (almost) all your cement ends up inside the liner instead of on the outside? Generally, it’s a disaster since there is no isolation of the annulus and now you cannot even circulate the well.

Point 2 discusses centralizers. As their name implies, the centralizers are used to keep the liner centered in the hole. Without them, the liner will tend to lay against the “low” side of the hole. Even in a vertical well this will still occur (and there is also no such thing as a perfectly vertical well). When the liner is against one side of the well (or even if it is sufficiently out-of-center without touching the side), there will be little to no fluid flow on that narrow side. The mud, spacers, and cement slurries will all take the path of least resistance which is on the wide side. Once again, this could result in inadequate cement coverage and a lack of isolation. This makes it seem like we should be running a lot of centralizers all the time, yes? (A lot of centralizers could be about 1-2 centralizers per joint of casing/liner where a joint is usually about 40-45 ft long.) No. Centralizers work because they act like springs on the side of the casing/liner pushing it off the walls of the open hole. When running the casing into the well, centralizers create a drag force and lots of centralizers means lots of drag force. This will make it more difficult to get the casing to the bottom of the well and increases the chance the pipe will get stuck before it is at the bottom. Based on what I have read, HAL proposed 21 centralizers which seems like it would be a bit less than 1 per joint of casing/liner in the open hole. This is probably very reasonable for a well that was planned to be vertical. BP chose to run six centralizers out of concern for time (which equals money) but also at least expressed a concern about getting stuck. (As an aside that is probably worth a post on its own, getting operators to run enough centralizers is one of the more difficult things to convince a client to do. It is an example of the nature of the relationship between the operators and service companies).

Point 3 discusses the evaluation, or lack thereof, of the cement job. Ok, I’ll admit to not knowing why they would not run some sort of log to at least determine top of cement, but preferably a proper evaluation of the job. At this point, cost considerations are probably the most prominent factor in the decision.

Point 4 is on the lack of circulation prior to cementing operations. Circulation is important to both evaluate the mud and ensure it is not coming back with gas in it and also to cool the well as cementing is done with prejob tests at specific temperatures that generally assume a particular amount of circulation. For a well that is experiencing losses that cannot be easily stopped, operators will sometimes not fully circulate the well with one hole volume or a “bottoms up” as it’s often called. This is because losses consume mud and circulating usually increase the rate of loss. If losses are severe enough, it can make it difficult to maintain sufficient mud volume (which needs to be constantly made up at the rig) to both circulate the well and keep adequate emergency reserve if well control problems do occur. As a consequence, cementing operations sometimes occur when the well has not been fully circulated. (Cementing a well that is experiencing losses carries with it many risks, but sometimes it remains the “best” option instead of spending more time trying to stop losses.)

Point 5 is on the lockdown sleeve and I am not really familiar with this device. I will instead cheat and speak generally about hardware and tools and how much stuff gets put into a well. The more complicated something is, the more likely it is to fail. What are the risks of a sleeve? Maybe it sets at the wrong time or place. Perhaps it creates an additional flow restriction which would make loss problems more difficult to deal with. I don’t know and I’m not going to get worked up over it.

I want to drive home that problems are rarely simple black and white choices. In this business, many factors go into decisions and selection processes. This is what we need to remember when examining why failures occur and how complex decisions get made.