Top Kill: what BP is trying to do

A post by Chris RowanI’ve been finding media coverage of operation ‘Top Kill’ – BP’s latest attempt to seal the Deepwater Horizon well leak in the Gulf of Mexico – rather confusing, so I’ve attempted to think through what pumping mud into the well is trying to achieve, and what we should expect to see if it is succeeding. I freely admit I’m not an expert, so I welcome any corrections and clarifications from people who are in the comments.
Update 30th June: ‘Top Kill’ has been abandoned – it seems that they just couldn’t get mud into the well fast enough.
The oil in the formation that the Deepwater Horizon drilled into is under high pressure: the high confining pressure provided by more than 5km of overlying rock has allowed high hydrostatic pressures to build up. Now that there is a weak point (the well) this pressure is pushing the oil upwards to the sea bed and out into the ocean (left in the figure below). To stop this, BP is trying to inject dense, heavy drilling mud into the well. If they can pump enough mud into the well, its weight will exert a downward pressure that will cancel out the upward pressure pushing the oil up the borehole, and no more fluid will escape (right in the figure below). At this point, the well can be permanently sealed by concreting it over.


Latest footage from the sea floor indicates we’re not at that point yet.


It looks like the stuff currently emerging from the leaks is probably drilling mud, rather than oil and gas – or at least mud mixed in with the oil and gas. As the mud starts to be injected into the well, the oil pushes back against it, and there is not yet enough mud at the top of the borehole for its cumulative weight to counteract the upwards pressure. The mud is therefore being forced back out of the well again.


So, if the operation continues according to plan, what should we expect to see? Assuming that mud can be pumped into the well at a faster rate than it is pushed back out again – and BP has enough drilling mud on hand to continue the operation – it will gradually push further and further down the well against the oil. More mud in the well means more weight for the oil to push against, which should reduce the amount of mud that can be pushed back out of the well. So the volume of leaking fluid should theoretically be decreasing as pumping continues. Lets hope so.

Categories: environment, geohazards, geology

Comments (36)

  1. Why does it have the bizarre moniker “Top Kill”?

  2. Chris Rowan says:

    It’s very Hollywood, isn’t it? You almost expect to find Bruce Willis piloting an ROV…

  3. It appears that BP may also be trying the “junk shot” along with the mud. At least that’s what this CNN story is reporting this morning:

  4. Chris Rowan says:

    Interesting. It sounds from that like they are having trouble getting any mud to stay in the hole – it’s all leaking back out too fast. Clogging some of the leaks with the junk shot might therefore help.

  5. Eric Palm says:

    Actually it not just the weight of the mud that they are counting on. Drill mud is highly engineered with lots of “platelets” that want to cling to the walls of the hole. In normal drilling operation the mud is forced out of the drill bit and into the walls of the hole that the drill bit has created. The excess mud carries the debris back up the hole and out. You can see that the mud has to seal off the walls of the hole or the hole will collapse on the drill pipe. So the mud itself is engineered to plug the hole. Having the “junk” will allow more rigid surface area for the mud to cling to and make it more effective.
    In addition, since they are pumping mud into the pipe through the blowout preventer the mud will naturally go both directions “up” and “down” seeing mud come out with the oil is expected. The hope is that Eventually the whole top part gets plugged and they can force the rest of the mud down the hole more effectively.

  6. BrianR says:

    As you note, when the hole gets to the point that downward and upward forces are more-or-less equal they will then go in and cement it up to plug it. My speculation (I’m no engineer) is that finding that equilibrium is not trivial and also takes some time (days?).

  7. timothy j slagle says:

    If they had a number of these 3 inch inlets to the well head, why were they not using them to channel out oil to temp pipes going to surface ships, and relieve presser on the leaking pipe. By decreasing presser on the broken pipe it increase the possibility of plugging or partial plugging the pipe from the outside.
    They had already inserted a 6 inch pipe in the hole of the broken pipe.

  8. Renato Fonseca says:

    Firstly, I´m not an engineer either, but a petroleum geologist (explorationist). The pressure that the injected mud will have to overcome is NOT equivalent to the WEIGHT of the overlying rocks. That would be enormous. It is equivalent to the weight of the overlying WATER column, i.e. hydrostatic pressure, plus or minus some component, due to undercompaction, gas column, etc.
    [Renato: Thanks for the catch. Have tweaked text a bit. – aj]
    Eric, remember that the well is CASED, so the mud will have no contact with the well walls, per se. Therefore, it is the weight of the mud that they will have to count on.

  9. Renato Fonseca says:

    If they did what you suggest, the oil would have to choose between two ways out: a broken pipe, maybe some 20 inches in diameter and the 3 inches inlet. Which one would you choose?

  10. Lockwood says:

    There’s a decent article on the nature of the mud in today’s NYT; as Eric P. says, it’s a carefully designed material, with a variety of formulations for different situations and applications. One thing the article missed is that a major component of most muds is ground barite (baryos is Greek for weight) to increase the material’s density.
    Another resource I found a few days ago is the blog The Oil Drum. In particular, this post from two days ago/a> gives some insight into how to interpret some of the images we’re getting from the ocean floor.

  11. Silver Fox says:

    Kill is from “kill the well,” Top is presumably referencing killing it from the top down – though am not sure about that last.
    Also, The Oil Drum has live video and live comment thread (plus two older threads).

  12. CherryBomb says:

    The Oil Drum is a sensible place to hear from people who know what they are talking about. It has been getting a little cluttered lately, though.
    The best common-sense analogy I can think of to a top kill is to imagine holding two garden hoses pointed at each other, and cranking up the pressure in one of them, hoping to force the water back down the other one. When the BP exec said this had a 60% to 70% chance of working, I just shook my head.

  13. Word is, Top Kill has been suspended for a second time. It simply isn’t working. I could never see any reason why it would.
    And twice in two days, Thad Allen went to the press with premature and over-optimistic claims that it was working. Given the importance of the position he holds, shouldn’t Obama have someone in it who holds some grasp of reality, however tenuous?

  14. mike hurt says:

    Suppose the mud pressure was sufficient to cause the mud to push the oil 100 meters (not very far) below the surface of the sea bed. In that case wouldn’t there be just oil and gas below the 100 meter mark and just mud above that point? If that is the case then I would expect (absent any concrete plugs) that if the mud pressure at the sea-bed surface were suddenly removed (i.e. the pumping of the mud was halted, as has been done twice) would result in the oil pressure pushing the 100 meters of mud right back out of the well. And the oil would then start flowing as before the effort w/ the mud.
    Is this what happened when they twice stopped the mud pumping?

  15. mike hurt says:

    I think another possibility is as follows:
    1) BP could not get the mud pressure at the sea-bed surface high enough to force the oil down the well at all- the oil and gas just kept flowing all the way up, through the BOP and into the sea.
    2) BP realized this and stopped pumping mud.
    3) They started the junk shot. I gather that this is simultaneously pumping mud and junk into the BOP. The junk is supposed to block the openings in the BOP to the sea (e.g. the ruptured pipe and oil flow channels within the BOP) so that these openings don’t bleed off the pressure of the mud.
    4) Then the increased mud pressure might be enough to force the oil back down the well hole.
    If anyone has firm knowledge which might confirm/correct my thoughts, please help.

  16. Mandobob says:

    The problem mostly has to do with the poor condition of the riser pipe at the top of the BOP. Earlier in the accident, it appears that the leak at the top of the BOP (where the riser is bent over) was from one location. By the time the top kill video feed was made available, it appears that the top BOP riser leak is now from 3 separate leaks. While the well has been flowing unchecked, abrasive material (formation material, such as sand, and possibly larger size clastics) along with likely broken cement has been eroding this portion of the riser, weakening it along already stressed and cracked pipe. Now that there is evidence of significantly weakened pipe at the top of the BOP, BP’s engineers are not very optimistic that adding “junk” to attempt to plug these leaks and provide some “back pressure” to the borehole. Given the likely pressure at the top of the BOP (around 8,000 psi based on lithostatic pressure [0.45 to 0.6 psi/foot X 18,000 feet TD for the well]. This assumes that the leak originated at TD it cold be higher in the bore
    hole), the damaged riser at the top of the BOP would need to hold that pressure. I would say that it is unlikely “junk” will be used and the attempt is to “overwhelm” the discharge via volume of kill mud. We will just have to see.

  17. Renato Fonseca says:

    Could you explain why are you considering lithostatic pressure, and not HYDROSTATIC pressure, in your assumptions?

  18. CherryBomb says:

    @ mike hurt
    You are just about spot on. It has been 20 years since I did this sort of thing, but the technology has not really changed that much, and the laws of physics are more or less the same.
    What they are trying to do is throw “junk” in the kill line, hoping that it will clog some of the passageways in the BOP where pressure is escaping. In my garden hose analogy, they are trying to prevent the water from simply spraying sideways. The leaks in the riser pipe are pretty much irrelevant, except as a way to monitor what is getting out. The problem is that they can only inject junk which will not clog the kill line itself, which is about 3″ in diameter, I think.
    Normally (I know, this is somewhat of a relative term when talking about blowouts), you would have the well shut in by using one of the BOP’s, and then using pumps to push the oil and gas fluids back down the well by brute force.

  19. CherryBomb says:

    One possibility I have not seen mentioned so far is to fish out the drill string and run it back in the hole with a perforation tool, to attempt a bottom kill. Such an operation would make all my inner organs scrunch up in terror, but it might work.

  20. mandobob says:

    I should do my research before posting. While I was not watching on Friday BP did introduce “junk” to what appears to be, based on the video feeds, of limited success. I would venture a guess that overcoming the well pressure without any back pressure doesn’t appear to be a success. Over on The Oil Drum, it has been reported that BP has stopped work on the 2nd relief well in preparation for removal of the BOP on that well so it can be used on plan “C” or “D”. This would involve removal of the riser and drill pipe at the top of the BOP (via cutting) and prep of the top of the BOP for a 2nd BOP placed to be secured on the top of BOP 1. Earlier today the video feed showed some sawing and other ops suggesting that that BP is pepping the well head and BOP for something?
    Renato Fonseca – lithostatic pressure overwhelms all hydrostatic pressure at deep subsea floor depths.

  21. Renato Fonseca says:

    As far as I know, you are not correct when you consider the fluid pressure to be equivalent to the lithostatic pressure at any given depth. You should consider hydrostatic pressure, give or take other components, such as under compaction, gas column, diagenetic reactions, etc.

  22. CherryBomb says:

    In general, reservoir pressures increase with depth, but not in a way that can be calculated in a straight line. Lithostatic pressure could be taken as a maximum possible pressure, but it could be anywhere between that and the hydrostatic pressure for a reservoir.
    What I suspect gave them fits on this well was multiple porous formations that did not increase in pressure with depth in such a way that they could be controlled with one column of mud with a constant density. You always have to deal with this to some extent, but you can tolerate a little slop +/- with the pressures, as long as it does not get out of hand. If it gets out of hand, you can have a situation where your column of mud is being invaded by fluids on one level, and actually flowing into the formation on another. This should be avoided, if possible.

  23. wouldn’t it make more sense to just learn from the disaster in 1979 and start drilling those relief wells right away instead of messing around for months while the gulf gets destroyed, AGAIN

  24. The NY Times is now reporting that Steven Chu advised against Top KIll, saying it was unlikely to succeed and might well make matters worse. So, if as Obama says, we’re really in charge, then why the heck is BP being allowed to carry out operations against the opposition of the US Secretary of Energy?
    Could it be that BP has more influence with the administration than its own Sec. Energy?

  25. Scott says:

    We have controlled explosives technology that should be sufficient to either shear off the pipe, or (at lesser force) slam it shut, regardless of the ambient pressure. Since the ruptured pipe is underwater so we aren’t worried about igniting the oil/gas mixture, why can’t they just use explosives to nip the pipe closed?

  26. Scott:
    The overpressure at the BOP is tremendous. Think thousands of p.s.i.
    I’ve seen schemes to collapse the shaft much deeper and over a much longer length, using a small nuke. I suspect those might work, but they’d still require drilling a wide shaft to insert the nuke, and constructing a SUV to deliver it, and realistically, there isn’t the political will to use a nuke, even for peaceful purposes. It would probably take as long to do this as drill a relief well.
    Nope, I think the Gulf is screwed.

  27. elijahbel says:

    It’s time to bring in the other oil producers equipment and expertise. How obvious can this be. Put down the corporate sign, and the political rhetoric, and get the best of the best to get this resolved. If were Obama, I would have had NASA’s best engaged as well.
    Free Business Cards

  28. Scott says:

    I found one measure of pressure generated by explosives (google “Castmex-3”) in a water filled bore hole used in an open pit mine. The reported pressure wave varied from 10 to 80 kilobars. Eighty kilobars converts to about 1,160,000 psi. The static pressure of water at 5,000 feet is on the order of 2,500 psi. The pressure of the escaping oil/gas mixture is reported at ~20,000 to 30,000 psi. From what little I found in 10 minutes, the normal industrial explosives reported would be enough to overcome that pressure (~7x to 38x). I have no knowledge of explosives, but the one-two punch of the explosive, followed by the sudden collapse of the bubble ought to be enough to close the pipe. A shaped charge generates pressures from 20 to 100 GigaPascals, or 14.5 million psi. That should be easily enough to cleanly cut the pipe at any depth, no matter what the pressure of the escaping oil.
    Of course, I could be wrong. I only know what I can google. 🙂

  29. CherryBomb says:

    One hopeful note (in a kind of sick way), is that by now, there is probably a ginormous HOLE in the producing formation caused by spalling of reservoir rock due to the pressure differential. The relief wells will not have to puncture the casing, since this well is blowing out through the cement outside the production casing. They just have to hit this hole, which I will take a wild guess and say it is 50 meters in diameter by now.

  30. Mike K says:

    This is not just a statics problem. There is a column of oil and gas that is, I believe, some 18,000 feet long that is moving at, by my rough calculation, about 20 feet/second (assuming a leakage rate of 12,000 barrels per day)in this 21 inch diameter pipe. Assuming that the oil and gas mixture has an average density one=half that of sea water, this column has the same momentum as a 140 ton vehicle moving at 60 miles per hour. It has to be brought to rest before a simple hydrostatic pressure balance can hold the oil in check.

  31. Sman says:

    @ Scott:
    I too am not a petroleum geologist, or engineer. But, as I understand it, using explosives could lead to communication with overlying reservoirs or the surface. If there are wells in some of those overlying reservoirs, that could lead to one hell of an increase in pressure, which could lead to blow outs, there???

  32. Jayanta says:

    I posted the following before I saw Gerard Harbison’s post. The nuke still seems to be the fastest option, as a tactical nuke is small in size, and digging a hole half a mile or a mile deep should take a lot less time than drilling a relief well.
    Previously posted:
    We now have the government saying that the leak will continue till August. Or maybe even until the entire well has emptied itself.
    What exactly is the situation? A pipe has been drilled into the well, and now is gushing out oil into the ocean bed.
    Solution: Dig a shaft, say about a mile deep next to the pipe (shouldn’t take more than a few days to do). Detonate a very powerful bomb, maybe even a nuclear bomb. The oil will no longer have a clear path to the surface and the leak will stop.

  33. Jayanta says:

    Sman probably has a point. The shock will travel very well through liquids.

  34. Ran in Michigan says:

    I think the problem with a nuke device under the sea floor and at that depth, is the possibility of creating a sink hole with a thousand leaks as the floor rises and the ceiling of the reservoir falls during the blast. The mass of the water MOVING back down on even 3 foot rise (bump) in the sea floor might be a greater force than the nuke itself.

  35. Rich R says:

    The initial analysis and follow ups seem sound to me. The top kill failed because they couldn’t force the heavy mud down the hole far enough so that the weight of the static column of mud plus the hydrostatic pressure of 5,000 feet of water was greater than the lithostatic pressure within the oil and gas reservoir. As Mike K observed there are also dynamic momentum, energy, and forces involved.
    Explosive solutions would likely worsen this problem by opening up a wider pathway for the oil and gas mixture to escape more rapidly. This would vastly increase the volume of heavy mud needed when they finally finish the rescue wells. I can imagine even worse scenarios as a result of a big blast such as a weakened well casing allowing the massive oil flow to erode and widen the well hole.
    Bottom kill would seem to work because the mud enters the original well column close to the oil reservoir about 3 miles below the sea floor. As this tall column fills up the weight of the mud counteracts the lithostatic force slowing and then stopping the runaway well.
    In the meantime, I hope federal government and BP work cooperatively to mitigate the amount of oil escaping and fouling the environment.
    Finally, we need to use less oil (and coal too) to preserve supplies and to prevent future environmental disasters.
    Caveat: I am an engineer, but not a petroleum engineer, so my comments are somewhat educated guesses.

  36. Paul Strickland says:

    Why not “quickly” manufacture a long tapered “pole” with a top diameter sufficient to stop the flow. Let a robot guide the tip into the broken pipe and then lower it into the pipe. Add whatever weight is required to overcome the oil flow. It may be a mile down. It may be a big pipe. It may be leaking oil and gases, but isn’t it just a big round pipe that needs pluging?