Forgot your password?
typodupeerror
Supercomputing Earth Social Networks Technology

NSF Gives Supercomputer Time For 3-D Model of Spill 102

Posted by timothy
from the why-can't-we-just-add-a-vinegar-spill? dept.
CWmike writes "Scientists have embarked on a crash effort to use one the world's largest supercomputers to create 3-D models to simulate how BP's massive Gulf of Mexico oil spill will affect coastal areas. Acting within 24 hours of receiving a request from researchers, the National Science Foundation late last week made an emergency allocation of 1 million compute hours on a supercomputer at the Texas Advanced Computing Center to study how BP's gusher will affect coastlines. The computer model they are working on 'has the potential to advise and undergird many emergency management decisions that may be made along the way, particularly if a hurricane comes through the area,' said Rick Luettich, a professor of marine sciences and head of the Institute of Marine Sciences at the University of North Carolina in Chapel Hill, who is one of the researchers on this project. Meanwhile, geographic information systems vendor ESRI has added a social spin to GIS mapping of the BP oil spill."
This discussion has been archived. No new comments can be posted.

NSF Gives Supercomputer Time For 3-D Model of Spill

Comments Filter:
  • In Time? (Score:3, Interesting)

    by MozeeToby (1163751) on Wednesday May 26, 2010 @03:47PM (#32352260)

    Is it really possible to develop, test, and run a complex simulation of the gulf's currents and weather in time for it to be useful for the recovery? It seems to me like the kind of job that could take a couple years to get right, even assuming we know everything we need to make such a simulation. Maybe they want the simulation developed today so that it can be used to evaluate potential future disasters? To help quantify the risk involved in this kind of drilling?

    • Considering that BP doesn't even bother measuring how much oil is spilling I think the modellers will have some trouble coming up with anything else than cable news entertainment animation.
      • Re:In Time? (Score:5, Interesting)

        by Stephen Samuel (106962) <samuel AT bcgreen DOT com> on Wednesday May 26, 2010 @04:54PM (#32353144) Homepage Journal

        Considering that BP doesn't even bother measuring how much oil is spilling

        Well, I'm pretty sure that they have a rough idea, and it's way more than what they're saying in public. They just don't want to formalize their estimates, because then they'll have to report the numbers. I did some napkin (units(1)) calculations based on the volume of dispersant that they say they've been using and, if they're using at the suggested dilution, then at a minimum they're dealing with 60K~500K barrels per day.
        ((don't have the actual calculations on hand, right now, and I'm on a different computer, so I can't even just look at my command line history))

        And that's a minimum.... the volume that they're using may be limited by the supply chain.

        I'm guessing that, in internal conversations, they're duck-speaking their way around solid numbers... For example, they can talk about how much dispersant they'll need ( a number based on oil flow), but there's probably an unwritten rule about never mentioning the oil flow estimate that underlies that calculation, because it'll be seekable in disclosure requests, and they'll still be able to 'truthfully' claim that they've never talked about the actual oil flow.

        • by thegarbz (1787294)

          Well, I'm pretty sure that they have a rough idea, and it's way more than what they're saying in public.

          That's a pretty solid statement. They've gone out of their way not to actually mention any figures in their public statements. All the 5k barrels / day numbers come from people like the Coast Guard and NOAA. The only numbers BP reports are what worked, i.e. how much oil they are syphoning off with their riser insertion tube.

          Given the circumstances can you really blame them. :-)

          • by azrider (918631)

            They've gone out of their way not to actually mention any figures in their public statements. All the 5k barrels / day numbers come from people like the Coast Guard and NOAA.

            Except for the one where they admitted that the pipe they attached was drawing 5K barrels/day into the ship with no visible decrease in the leak rate. That one did come from BP.

    • Re:In Time? (Score:5, Funny)

      by omnichad (1198475) on Wednesday May 26, 2010 @03:50PM (#32352314) Homepage

      If my knowledge of B-movies is correct, there's already a GUI interface with 3D graphic modelling too (in Visual Basic, no less)! It's just a matter of typing a few parameters on a keyboard. How real supercomputers got mixed up in a cheap disaster movie, I'll never know.

      • Hey! My work resembles that remark. Especially the VB part. Other than being horrible, what's wrong with it?!
      • Re:In Time? (Score:5, Interesting)

        by Monkeedude1212 (1560403) on Wednesday May 26, 2010 @04:15PM (#32352596) Journal

        Well, You're actually not too far from the truth. Surprisingly enough, video game engines provide great simulators for these kinds of things.

        I know of a few cases where people have licensed the Unreal Engine for creating their own simulators, specifically I think it was to deal with some high collision testing. Since the Unreal team has started with some fluids physics, its not hard to see someone possibly using that engine to simulate the effects.

        Then, all thats needed is the writing of Entities, which is usually done in C++, not VB. But the idea is the same, you only need about a weeks worth (40 hours) to write the basic entities you'll need, and maybe a bit longer for the complicated ones. Depends on what you are simulating, but if say there are only 4 different types of matter they have to consider, water, oil, air, dirt, than there really isn't that much to add.

        So, really, the GUI would be already made, it'd probably be the Unreal Map Editor. The Code they would be writing would be C++, which most /.ers can at least recognize, and you won't need to make any models, since you are going to be using brushes for the landscape and liquids for the rest.

        Not that they had to us Unreal, they also could use the Source Engine, though I don't think source offers much in the lines of liquids. But there are actually many ways to go about this. In fact, some modelling and animation software lets you simulate water physics, one in particular I know but I can't recall the name right now. (A plugin for Maya or 3dsMax or Softimage? Bah).

        So really, the only LENGTHY part of development is duplicating the ocean currents, which if you have recent records on, is about as trivial as stitching their co-ordinates to the map you create, and making your entities flexible enough to handle the varied input.

        I'm not trying to belittle the task of simulating something like this, but with a team of 10 or so people I could see it being done rather quickly. You have to realize that simulations are never quite 100% like the real thing. They have to cut off at some approximation point to keep the simulation running, otherwise the computer would hang trying to figure out the googelth digit. So, really, because the gulf is so large, its not like they have to calculate everything down to the particle.

        • Re: (Score:3, Insightful)

          by jpmorgan (517966)

          Uh, no. Nobody uses the Unreal Engine or the Source Engine to model diffusion and fluid flow. This is a field I work in professionally and academically, and you'd be a laughing stock if you seriously suggested taking this route. There is an enormous difference between a game engine which is designed to make things look good, and an accurate physical simulator.

          Furthermore, the underlying software already exists, academically and commercially, with 6 figure licensing fees, which is good because developing a s

      • Actually, the GUI interface *does* exist! There are various packages to do that, e.g. Paraview and Ensight. One needs to output the numerical data in a format that these programs can understand (e.g. .vtp for Paraview that can be binary or even plain ASCII, with a really simple structure) and then it is possible to visualize the information in colorful animations, just like in the movies.
    • Re: (Score:3, Informative)

      by squidfood (149212)
      I don't know about the GOM specifically, but models of oceanography of most regions around the U.S. have been under development, availability, testing, and use for quite a while, and substantial toolkits exist, so it's just a matter of putting some fluid of the right density in the right place and getting present starting conditions. The crunch is really getting the computer time to run the models which is what the new money addresses.

      And yes, there's also a lot of graphical toolboxes [ucar.edu] so whipping together

      • I don't know what the oceanography codes can do, but for simulating the movement of a liquid in another liquid (oil in water) using CFD (as I would expect) one would probably need a surface tracking or capturing algorithm. Now, this can range from "pretty easy" to "freaking hard" depending on the accuracy needed. Maybe getting a rough idea of the direction in which the oil will move is not that complicated. Using COMSOL would give results pretty fast, but I the article does not say what they are using and I
    • by Machtyn (759119)
      Perhaps the Gulf of Mexico weather and current patterns have already been developed and modeled for the computer system. If so, they'll borrow/buy that work and enter their numbers and data.

      Simplistic, I know, but it could happen. Still, it took 30+ days for the oil to reach New Orleans. They've got a little bit of time to model this stuff out and give disaster areas a bit of warning. (About as effective as a tornado warning can be.)
    • Re:In Time? (Score:4, Informative)

      by Bill Barth (49178) <bbarth@@@gmail...com> on Wednesday May 26, 2010 @05:01PM (#32353246)
      The code in question (ADCIRC) has been used for years to do hurricane storm surge simulations. It's being continuously developed for work in the Gulf of Mexico and already includes contaminant transport effects. Also, as with all things scientific, "right" is a relative quantity. The better question is whether or not useful predictions can be made that are better than what's been done so far. I think the answer to that is a resounding "Yes!" Finally, I guarantee that this event will be used by modelers to refine and improve their codes for years to come. Recent hurricances (Ike, Rita, Gustav, etc.) have been used in the very same way.
    • by Nadaka (224565)

      NAVOCEANO is already running a highly complex simulation of the gulfs currents, tides and weather systems. It would only be a matter of adding the modeling of the oil.

  • Offtopic (Score:5, Insightful)

    by Pojut (1027544) on Wednesday May 26, 2010 @03:47PM (#32352270) Homepage

    Explain something to me: you people bitch about government getting involved with private industry, yet when BP screws up, you demand the government to take over. WTF?

    You want the government to take over...ok, what do you expect them to do? BP has the equipment, the government does not.

    You all want the government to step in...yet you don't want them to raise the liability cap. So...you want taxpayers to pay for cleaning up a private company's mess, then?

    Shifting gears...

    This is NOT a reason to stop offshore drilling. Offshore drilling is an essential part of our current energy use. What this is, however, is a good reason to reinforce laws surrounding safety and preparedness standards...and make sure they are fucking followed.

    • by omnichad (1198475)

      The government should manage it, the company should pay. That would keep the motives separate, anyway. Not that we really want to put something as slow as a government in charge. But it might prove to be better than a corporation.

      • by Pojut (1027544)

        That's what they are already doing, though [whitehouse.gov]...what else do you want from them?

        • Re: (Score:1, Insightful)

          by Anonymous Coward

          Do you happen to have a non-propaganda news source? America's answer to Pravda is hardly what most people would consider credible.

          • by Pojut (1027544)

            You could listen to hearings on C-SPAN, in which BP execs acknowledge the government's role in assisting them. Or you could just Google for the information yourself.

            No offense, but the fact that you are convinced that the White House is lying about having boots on the ground in the Gulf leads me to believe you likely get your info from a regular News Org, yes?

            C-SPAN all the way. Direct from the horse's mouth, no filter.

            • Re: (Score:2, Insightful)

              by Anonymous Coward

              Oh, I have no doubt that the WH has "boots on the ground," as you say. I also have no doubt that the WH also has "boots under the beds" of execs from just about every sector including energy, media, finance, and automotive.

              • by Pojut (1027544)

                I'm not denying that...in fact, I would agree with you.

                Regardless, the poster asked for proof that the administrationw as actually doing something, and I provided it. ::shrug::

        • Re: (Score:2, Informative)

          by Anonymous Coward

          Proper fucking booming would've been a good start.

          http://www.dailykos.com/story/2010/5/11/865387/-Fishgrease:-DKos-Booming-School [dailykos.com]

      • If you want it to be terribly inefficient and slow then, yes, the government should manage it. If you want the work done to a standard, then the government should have oversight, defined as regulatory supervision, and let BP and its contractors do the work.
      • by aaarrrgggh (9205)

        Who the hell in the government has the deep drilling experience to be able to understand the issues (and scale) at hand?! I would have hoped that the government would have watched over BP and made sure the booms were being done properly, but my guess is that the under-water problem is much more severe than anything you really see at the surface.

    • Re: (Score:3, Insightful)

      by Lithdren (605362)
      Odvious troll is odvious but I thought i'd add a comment. Its all togeather possible the people demanding the Goverment take over the situation are not the same people who want the goverment out of private industry. Its quite possible people are saying what they fell on one subject, while steering clear of the other one, that would contradict their original feeling. Imagine that!
      • by Pojut (1027544)

        While I don't post on them, I frequent the Fox News, Politico, and a few other political forums. From what I've seen on them (and from the few hard-right folks that I know in real life), they are many of the same people.

      • by timeOday (582209)

        Its all togeather possible the people demanding the Goverment take over the situation are not the same people who want the goverment out of private industry.

        Well, not in some cases: [washingtonpost.com]

        "Today is Day 36" since the well's drilling rig exploded, Sen. John Barrasso (R-Wyo.) said Tuesday during a hearing on offshore drilling liability. "The cloud of confusion over how much oil is spilling into the gulf is very concerning. And it's also very unclear who was in charge."

        Republicans are seeking to erode voter confi

    • by causality (777677)

      This is NOT a reason to stop offshore drilling. Offshore drilling is an essential part of our current energy use. What this is, however, is a good reason to reinforce laws surrounding safety and preparedness standards...and make sure they are fucking followed.

      If you haven't noticed, there are forces in our government that don't want us to be a strong, stable, well-managed, independent country. It would interfere with their desire to undermine our national sovereignty in order to accelerate globalism and a

    • Re: (Score:3, Insightful)

      by mcgrew (92797) *

      You want the government to take over...ok, what do you expect them to do? BP has the equipment, the government does not.

      First off, whoever modded you offtopic should have his moderator priveledges taken away.

      Now to the actual topic: Government shouldn't take over, but they'd damned well better show some oversight, and make sure that BP is following all laws to the letter. They may have the equipment, but they don't have the motivation -- their only motivation is to rake in profits, or the "accident" wouldn'

      • by Pojut (1027544)

        Why aren't corporate heads ever put in prison for negligent homicide when the company disobeys laws and people die?

        The insight shown in the rest of your post leads me to believe you don't actually need that question answered "out loud" :-)

        • by mcgrew (92797) *

          No, but a lot of people never even think of the question. Why isn't everyone asking it, loudly, to their lawmakers?

    • by Locke2005 (849178)
      Obviously you've never been rich. The way to get rich is to gamble with other people's money, and if you win, keep the profits, but if you lose, get somebody else to pay the losses. The history of the very rich is full of examples of this, e.g. the Hunt brothers. It used to be fashionable to leave the banks holding the bag when you gambled and lost; your corporation simply declared bankruptcy leaving you with no personal liability. Lately it seems to have become fashionable to leave the government holding t
    • Responsibility and liability are two different things, and worse, the Federal government already limited BP's liability through previous legislation.

      But the Responsibility for protecting the *coast* rests ultimately, on whoever actually wants the coast to be protected. The residents, for sure, many of the rest of us, as well. In that regard, if certain youtube [youtube.com] videos can be considered representative, the governments at all levels up to the federal have been as negligent as BP: It's been spewing for like a

    • > This is NOT a reason to stop offshore drilling. Offshore drilling is an essential part of our current energy use. What this is, however, is a good reason to reinforce laws surrounding safety and preparedness standards...and make sure they are fucking followed.

      Well, that sounds reasonable. But, then, as you pointed out:

      > You want the government to take over...ok, what do you expect them to do? BP has the equipment, the government does not.

      So.... ummm.... hmmm. I'm stumped. How exactly is regulator Jo

    • I read somewhere that BP’s oil spill disaster is a perfect example of the complete failure of self regulating capitalism. Currently, the US government allows oil companies to regulate themselves. The U.S. government is demanding more accountability from BP with regards to the the cleanup.
      Throughout history, engineers and scientists make mistakes and it has always been that everyone else pays for those mistakes. These guys have to be held more accountable for the SOCIAL AND ENVIRONMENTAL CONSEQUENCES

    • by drinkypoo (153816)

      This is NOT a reason to stop offshore drilling. Offshore drilling is an essential part of our current energy use. What this is, however, is a good reason to reinforce laws surrounding safety and preparedness standards...and make sure they are fucking followed.

      "This" is not a reason to stop offshore drilling; the very concept should not have been employed at our current technological level. And really, it should never have been employed. We have enough desert and sufficient technology to replace all our offshore drilling with biodiesel from algae. Once you realize that, you have to understand that the idea is simply immoral. It's about profit, plain and simple; there's money to be made doing it, so we will do it whether it is actually beneficial to the human race

  • How do they make a 3d model of the oil spill when they STILL have no idea how much oil is actually spilling?
    • Re:Um... (Score:5, Insightful)

      by Pojut (1027544) on Wednesday May 26, 2010 @03:51PM (#32352324) Homepage

      Rest assured...someone knows. It just isn't us. BP likely knows, which is (from their perspective) a good reason why we don't.

    • Re:Um... (Score:4, Informative)

      by confused one (671304) on Wednesday May 26, 2010 @04:02PM (#32352454)
      The models might actually be a good way to verify the estimates. Is it 5,000 barrels a day (not) or 30,000-70,000 barrels (within the realm of possibility). We know it's not 5,000 because they're pumping it out now at 5,000 (with the 4" pipe the inserted into the riser) and they haven't had much effect on the outflow.
      • They're pumping out 5K barrels of oil/water mixture, that's a pretty big difference.

        • Sure. I havent' dug into the details; but, as I understood it, they stuck a 4" pipe into a 21" riser which had oil and gas flowing out under some pressure. They're drawing oil up through the 4" pipe at a fairly high rate. I'd expect there to be some water intrusion, there is more than one breach in the riser. If the 4" pipe is inside the riser a fair way, (it's 5' long according to BP's site) I'd expect it to be picking up mostly oil.
  • "I don't think that they have any idea how this oil is predicted to move through the marshes and the nearshore zone," said Luettich.

    I understand that "nearshore" zones may be hard to predict: I wonder if that map [computerworld.com] (see white line) shows how close to shore they can predict...

    Also this may be used as a forecast model, but to me it seems like measuring and predicting a hurricane while ignoring storm surge...but I am not going to be critical with little knowledge on what data was available to the programmers.

    • by Bill Barth (49178)
      The model itself can get quite close to shore (much closer than the pixel-level resolution of that tiny map you linked to), includes wetting and drying of land regions, and has been used to predict (in both forecasting and hindcasting modes) hurricane storm surge.
      • Many of the computer models tracking the oil spill have resolutions of 500 meters to a kilometer, but the model being created on the Texas supercomputer can bring the detail down to a resolution of 50 to 40 meters, which is fine enough to show, for instance, simulations of currents moving up channels, said Dawson.

        I need to work my reading comprehension skills haha (I'm up to a third grade level now!) Also...

        The project is getting a "high priority," said another researcher, Clint Dawson, a professor of aerospace engineering and engineering mechanics at the University of Texas. "What our model can do that a lot of the other models can't do is actually track the oil spill up into marshes and the wetlands because we have fine-scale resolution in those areas."

        Tracking and predicting are two different sets of challenges in my opinion. I applaud the attempt to forecast and use models to hopefully mitigate damage, but I am still skeptical on how they will model the oil after it has already hit the marshes outside of tide patterns... I guess we will just have to wait and see :)

  • I'd be interested to see how the modeling matches up with the actual outcome.
  • by Anonymous Coward on Wednesday May 26, 2010 @04:03PM (#32352460)

    I caught this link a day or two ago. Take a look. [boston.com]

    • Someone mod this post up...those photos were amazing and really are worth viewing! I found the aerial pictures of the oil booms really interesting, thanks AC!

  • Compute Hours? (Score:5, Informative)

    by adbge (1693228) on Wednesday May 26, 2010 @04:06PM (#32352492)
    For those unfamiliar with a "compute hour", the supercomputer in question is capable of 63,000 compute hours per hour. To put this into perspective, the NSF seems to have allocated about 15 hours of supercomputer time to this project. 15 hours is, of course, not nearly as sensational as 1 million compute hours. ;)
    • by meatpan (931043)
      The 1 million hour quota seems even smaller when you consider the nature of the simulation. Computational fluid dynamics simulations rely on super-linear performing algorithms, so a quick & dirty modeling effort would need to drastically reduce the amount of data and number of variables considered.
    • by Bill Barth (49178)
      The simulation is currently running on about 4k cores which is about 244 hours or 10 days worth of simulation. Each of the simulations they're running is about 10 hours in length, so this is enough for about 24 forecasts.
    • by mapkinase (958129)

      "compute hour", also known as "CPU time" vs say, "wallclock time".

      • by Bill Barth (49178)
        I should have pointed out above that we measure in wall-clock time not CPU time. Most of these codes don't spend much time waiting on I/O, so the two numbers are usually close. We use wall-clock time because that is the time that the user monopolizes the nodes that are assigned to their job.
    • by jsepeta (412566)

      i wonder if there would be a better way of us DONATING our unused cycles to the project. because 15 hours kind of sucks.

      • by Bill Barth (49178)
        Unless we've all got InfiniBand between us, it's not really worth it. These simulations are tightly coupled across the nodes they run on making them very sensitive to the latency and bandwidth between them.
  • Tar Ball (Score:5, Funny)

    by c0d3r (156687) on Wednesday May 26, 2010 @04:07PM (#32352512) Homepage Journal

    They should make a big Tar Ball containing the build of the software they write that performs this analysis.

    • Re: (Score:3, Funny)

      Why bother? Once they open the tarball they'll just pipe the output fulfill their dependencies. ;)
    • I wouldn't be surprised if they are using off the shelf tools like FFTW on that supercomputer. It was just yesterday I was amazed at IBM featuring FFTW3 binaries and sources for BlueGene, just like some laptop support software from their website.

      I can't find the URL (which I saw on IBM) now but, as you see from here http://www.complang.tuwien.ac.at/skral/fftwgel.html [tuwien.ac.at] , it is just 2.7 mb ordinary tar.gz file, builds on PowerPC 440. Of course, number of PPC 440's it runs on is what matters :)

  • Hope this is more accurate than the ones they use on Wall Street or it's going to make it worse.
  • We need to get Bruce Willis to command a submarine down to the bottom of the ocean and seal the leak with a nuke!
    • Re: (Score:3, Interesting)

      by camperdave (969942)
      Well, maybe not a nuke. But how about a large rubberized canvas funnel connected to a hose to the surface?

      Oh, and why don't oil rigs have a large containment boom around them in the first place?
      • by timeOday (582209)

        how about a large rubberized canvas funnel connected to a hose to the surface?

        Basically that was one of the first things they tried [csmonitor.com]: "If successful, the containment box would begin funneling as much as 85 percent of the oil plume into a pipeline pumping the oil into a barge on the surface as early as Sunday." That was May 7.

        As for the boom, A) they're not working all that well and B) with the well a mile underwater, it could disperse over a huge area before reaching the surface.

    • Pretty sure you want Ed Harris for your underwater drilling problems...

  • This will probably be a big, complex simulation. It will be interesting to see how well it matches up with reality.

  • The effects of the spill will suck, and will suck for a long time (the halibut still have not returned to the area around the Exxon Valdez spill). How will knowing in advance just how badly it will suck help? Are they going to take some remediation efforts as a result of this simulation that they wouldn't have otherwise? They claim they are already doing everything they can... how can we expect them to do even more?
    • by Abcd1234 (188840)

      The effects of the spill will suck, and will suck for a long time (the halibut still have not returned to the area around the Exxon Valdez spill). How will knowing in advance just how badly it will suck help?

      By understanding how the oil will diffuse and spread in the ocean, they can direct cleanup efforts to areas that are likely to be hit hardest. There are, after all, a finite number of people, booms (yes, even fucking booms), etc, so optimizing their allocation is important.

  • NPR has posted the live-feed of the leak, and the topkill procedure on their website. so far its not working. i know, i know, off topic.
  • I'm curious to learn what numbers they're using for their analysis, since BP has been cagey and lying about the actual amount of oil flowing from the leak. For their model to predict accurately, they'll have to use better numbers than BP is fudging us with.

    http://homelandsecuritynewswire.com/bp-oil-leak-much-bigger-official-estimates [homelandse...wswire.com]

"Right now I feel that I've got my feet on the ground as far as my head is concerned." -- Baseball pitcher Bo Belinsky

Working...