Bad Math Causes Explosion at CERN Collider

javipas writes "The Large Hadron Collider at the CERN has suffered a big explosion deep inside that has caused a leak of hellium gas and the quick evacuation of everyone working there. The reason: a mathematical mistake that affected the design of the giant superconductive magnets made by Fermilab. Now the company will have to repair and upgrade the 24 magnets that are installed on the 27 km. circunference of one of the most important research centers on Earth." This story might seem strangely familiar to you.

Comments

Not a Dupe

[Anonymous Coward]

Isn't this the same story as from a week or two ago?

While it may have the same message to you (Big Ass Magnet Fails on Fermilab's Collider at CERN), it's actually the result of an investigation.

From one of the articles in your link:

Fermilab will appoint an external review committee to analyze how this problem occurred and determine root causes and lessons learned.

The old story was that stuff blew up. The new story is why it blew up so we don't make the same mistakes. Turns out it, was just bad math. It wasn't that we didn't understand some physics, it wasn't the gods being mad, it was just plan old avoidable bad math.

A somber and depressing article for the /. community considering how many people have been posting about the huge leaps in physics this machine was supposed to bring us ... hopefully another country will come up with something similar to keep this research rolling while CERN awaits repairs.

Re:Not a Dupe

[bockelboy]

I work on the LHC (although not a particle physicist, I talk to ones every day).

If you had to explain it at an 8th grade level (as newspapers aim for these days), you'd say "bad math". If you are on a nerd site like Slashdot, I'd hope we wouldn't need to make that simplification. The story is a dupe. It is still the same as before - the assymmetrical load was not put into the requirements for the magnets and overlooked during four internal and *external* reviews. CERN had all the right data, and they overlooked that specific test too.

There is a committee reviewing the case, and their findings will be released April 24 (tentatively). FNAL's goal is to have this not delay turn-on at all, although it'll cost some amount of money to fix. They hope the repairs can be made in-ground. The absolute worst-case would be if they have to take the magnets up to the surface to fix them; that will certainly cause a time delay.

Right now, they suspect it's an additional cost, but not a delay for the November turn-on. That picture could get worse, but we won't know until around May.

Lots of the world's top particle physicists have been on this project for many years; any country capable of doign "other" research is certainly already heavily involved with the LHC. The only possible project which will benefit from the delay is the Tevatron at FNAL, but we're probably 18 months from running the LCG at Tevatron levels (it will take *at least* a year to begin to get all the bugs worked out and tunings done to a multi-billion dollar system).

One delay will be noise compared to the amount of effort needed to prove the existence of the Higgs.

Re:Not a Dupe

[bockelboy]

I guess I should cover my ass and point out that none of that post is priveleged information; beyond my meandering speculations, you can read the press release yourself:

The goal at CERN and Fermilab is now to redesign and repair the inner triplet magnets and, if necessary, the DFBX without affecting the LHC start-up schedule. Teams at CERN and Fermilab have identified potential repairs that could be carried out expeditiously without removing undamaged triplet magnets from the tunnel.. All three of the pressure-tested triplet magnets at Point 5, plus the associated DFBX, will be removed from the tunnel for inspection and, if necessary, repair. CERN will manage the redesign and repair effort and has scheduled a review for April 24-25 to validate the selected method. Fermilab will take part in the review. Repair of the triplet magnets would begin after validation by the reviewers. The immediate goal is to have a repaired triplet in another sector of the accelerator ready to participate in a pressure test scheduled for June 1.
Primary sources are always better than some guy commenting.

Some real information

[gvc]

Here's Fermilab's statment. Of course they are an interested party, but at least their statement contains information, unlike the snide popular press article.

http://user.web.cern.ch/user/QuickLinks/Announceme nts/2007/LHCInnerTriplet.html

Re:NPR Story missed this one

[Waffle Iron]

They mentioned how a particle zooming around in it would have the force of a bus

Not really. The most powerful cosmic ray particles ever observed, which have are millions of times more energy than anything we can create, each have approximately the force of a thrown baseball. Perhaps *all* of the particles in the ring together have the energy of a moving bus.

suffocation

[SuperBanana]

What's so bad about that?

What's bad is that it displaces all the oxygen in the area. This was a common cause of occupational deaths in MRI rooms- not flying metal objects attracted to the magnet (though a very small number of people have been killed by oxygen tanks and such.) An MRI repair tech was killed because of a slow helium leak that lowered the oxygen percentage enough that he passed out. That's why most if not all MRI facilities have gas monitors that monitor oxygen, nitrogen, and helium levels (liquid nitrogen is also used.)

MRI machines have vents for this sort of thing. Also because if the magnet quenches, a LARGE amount of liquid helium will boil off; all the electrical energy used to generate the field, which is constantly running in the magnet, turns very quickly into thermal energy. If the vent wasn't there, the room would pressurize, preventing one from opening doors (even an outward opening door- enough force would make it impossible to overcome friction on the bolt.) Magnet quenches are done only in situations where someone's life is in immediate danger (say, they're trapped by a ferrous object and about to bleed out) because of the danger (and the fact that there's a 1:4 chance of destroying the multi-million-dollar magnet and boiling off thousands of gallons of very expensive liquid hydrogen.)

It's been reported in vent failures when a magnet quenched that it rained oxygen; liquid helium is substantially colder than liquid oxygen. Shit happens: vent valves fail, birds nest in stuff, someone says "hey, what's that big empty pipe for" 6 rooms over and cuts it/blocks it off, etc. I think the MRI tech was killed because of a leaking o-ring.

Are they just afraid no one will take them seriously if they sound like the chipmunks when they report their findings?

Picture one guy yelling "Run, run! We'll all suffocate!" in a chipmunk voice, and everyone else laughing at how funny he sounds, and passing out. And dying.

I mean, it's not like it's spraying O2 in the direction of the pilot light of their oven.

Oxygen spraying in the direction of a pilot light in an oven will do nothing except make the pilot light burn at a higher temperature. It will not cause an explosion, because there's nothing else combustible in the oven, unless it's REALLY greasy.

What is not a joking matter is smoking in high-oxygen environments or fires in spacecraft, because they do have lots of flammable stuff, like wire insulation (which is fire-resistant, not necessarily fire-proof.)

Re:suffocation

[Hoi Polloi]

What is not a joking matter is smoking in high-oxygen environments or fires in spacecraft, because they do have lots of flammable stuff, like wire insulation (which is fire-resistant, not necessarily fire-proof.)

That sounds very familiar.

Super Conducting Super Collider

[kad77]

Thanks to democrats (neo-luddites?), the US doesn't already have this technology. Know your (recent) history.

http://www.rootsweb.com/~txecm/super_collider.htm

http://motls.blogspot.com/2006/03/ssc-and-clinton- administration.html

Re:Proton?

[Anonymous Coward]

> the Higgs boson particle is being sought in order to prove large portions of string theory.

Actually, it is parts of the standard model that it is proving. String theory would probably just be tweaked a little if something unexpected happened with the Higgs boson and the standard model.

Re:Something smells fishy

[The_Wilschon]

Fermilab and CERN are only competitors under a quite loose definition of the word. ATLAS and CMS are competitors, CDF and D0 are competitors, Fermilab and CERN are not really. Actually, most of the people working at CERN either also work at or have worked at Fermilab (or one of the other accelerator labs). Most of the people at Fermilab are anticipating working at CERN in the next few years. I myself have been working at Fermilab for the last few years, but I am starting work at CERN this summer.

CERN is a continuation of what Fermilab has been working on, not a rival.

The CAPTCHA is "footstep". Appropriate. CERN is following in Fermilab's footsteps (and then going quite a bit further).

Re:Forgetting...

[CaptnMArk]

Actually, it wasn't a crowbar until it was bent around the magnet.

Oblig. YouTube link to the video

[DeadCatX2]

Man, Goble rocks. He drives a car with the license plate UNIX.

Here's a YouTube link to the video. I don't think barbecue is the right word...try incinerate.

http://www.youtube.com/watch?v=TBLr_XrooLs

Re:NPR Story missed this one

[Waffle Iron]

If you look up "force" on dictionary.com, you'll find 36 different meanings. From the context of the sentence, it was clear that the GP post was using meaning #3 (energy, power), not the formal physics definition #12 (f=ma). I guess that I lazily copied the term assuming the GP post's context, not remembering how literalistic the audience here tends to be. Sorry for any inconvenience.

Re:Forgetting...

[Enlightenment]

That worst case scenario was only stated because of a strict adherence to the principles of science: since nobody's done it yet, there's technically no way to know what happens. But common sense dictates that the world is probably _not_ going to end as a result of a high-energy collision between two particles. The energies involved aren't great enough to do that by, say, cracking the globe in two, so some weird effects from the collision would be the only way. You really have no cause for concern.

Re:NPR Story missed this one

[HoldenCaulfield]

The relevant quotation from the story on NPR:

"It's the energy of a bus moving at a normal velocity," De Rujula says. So imagine a bus rolling along -- which has something like 10,000 trillion, trillion particles -- but transfer all that energy into one single particle. There will actually be a beam of protons; a whole fleet of subatomic particles, each carrying the energy of a bus.

In other words, the grandparent just mis-remembered the story, or didn't realize how important the distinction could be when talking physics . . .

Not a Dupe, indeed

[Anonymous Coward]

I am a physicist working at CERN, but have nothing to do with the superconducting magnets. I work on the detectors. From my standpoint, if CERN admitted to the fact that there will be a delay and set some reasonable amount of time for said delay, then it would make the life of others who are trying to play catch-up constantly, a lot easier. There is a huge amount of pressure on those constructing the detectors, getting the computing infrastructure ready, etc. to be ready for data taking by summer 07. Such a pressure cooker is a good incubator for problems to come. Some breathing room would be most welcome!

Indeed, a delay of a few weeks is nothing compared to ~20 years some physicists have been working on the project. (I am considered a newbie, after 3 years on it!) The director general (DG) of CERN has "promised" to the funding agencies to deliver the beam by the end of 2007 and right now, with the current schedule, that would be Dec 07. A few weeks makes that 08, which would make CERN look like it did not live up its promises. Also remember that the LHC has mostly been built with loans. The sooner it starts up, the sooner CERN can pay the money. Most of the cost is construction. Although the electricity bill is so high that CERN will not be able to afford running in winter again after the first startup. (Electricity prices in France change on a daily basis -- often dictating what happens when.)

The former DGs of CERN had been physicists for the last 50 years. But this DG is actually an engineer. This decision was a concious and good decision as the biggest challenge for the *start* of the LHC is an engineering challenge -- not a physics challenge. The physicists have a lot of say in the design of the LHC, the construction of the detectors and the analysis of the data -- but *not* in the construction of the LHC, which is the biggest cost. So all those making fun of physicists, well, remember, the LHC is an *engineering* project, not a physics project. Any chance you are an engineer?? ;)

(For those who are wondering why I said that everything needs to be ready by summer 07. Before the LHC collisions with two beams colliding head-on every 25 nanoseconds, with ~20 interactions in every bunchcrossing and a resulting animal zoo of 200 or more particles spewing out of the interaction point, the LHC will have single beams going around the ring, probably summer/fall 07. Although having single beams in the collider sounds like "no fun" -- actually it is. Because the beampipe will not (can not) be at total vacuum so there will be interactions between the gas in the beam and the single beam going around the ring. We will not find ths Higgs in the data, but is crucial for calibrating the detectors... )