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.

It's Sabotage!

[Anonymous Coward]

Coincidentally, Fermilab stands to gain most from delays at Cern. Its researchers also operate a rival but less powerful particle accelerator, the Tevatron. Fermilab staff are pushing the Tevatron to ever-higher energies hoping that they might find the Higgs boson before the LHC switches on. An LHC researcher said: "Ironically, this delay could be all they need."

Gasp! Sabotage! A divide by zero here, a drop of the carry there...who would notice!? :)

Re:Super Conducting Super Collider

[Anonymous Coward]

Calling the democrats neoluddites while the republicans are still around?

Sorry, you are in the wrong place, parallel universe is the next thread down.

Helium is not friendly!

[Sockatume]

When you're working with liquid nitrogen and liquid helium (as coolants for superconducting magnets) it's easy to assume they're harmless because they're chemically inert. However a small volume of liquid boils into a huge volume of gas, which will exclude the air - and precious oxygen - from the vicinity. A big helium leak is no laughing matter because of the asphyxiation risk.

Something smells fishy

[Hoi Polloi]

So Fermilab, CERN's competition, designed the magnets that happened to have a basic design flaw? Hmmmm, cue The Beastie Boys tune "Sabotage"!

Re:Super Conducting Super Collider

[644bd346996]

Mr. Tubes is a Republican. You can't get much worse than him.

Re:Super Conducting Super Collider

[khallow]

So how did democrats balloon the cost of the project from $4.4 billion to $12 billion (in 1993 dollars)? I notice that the Large Hadron Collider has only cost around $3 billion in current dollars including cost overruns. The SSC seems to have a similar track record to the International Space Station (which was formed about the same time, started also with an initial cost of $4 billion and eventually grew to over $40 billion). So IMHO it's likely that it wouldn't have stopped at just $12 billion. The LHC cost overruns appear to be less than 50% of the initial project cost. At some point, you have to recognize that the SSC was poorly run with heavy pork and out of control costs associated with it. If the SSC has maintained some sort of financial discipline, we'd probably be using it today.

Translation

[iamlucky13]

The linked article, which has more useful information in each paragraph than the entire original article from the story submission, is a little technical. Lemme try and simplify the important parts:

preliminary indications are that structures supporting the inner "cold mass" of one of the three magnets within its enclosing cryostat broke at a pressure of 20 atmospheres, in response to asymmetric forces applied during the test.

The magnets are chilled with liquid helium to keep the temperature near absolute 0. Some of the support framework which holds one of the magnets inside the coolant pipes ("cryostat") failed at 300 psi because the loads did not line up in a way the framework was designed to hold.

Such forces are expected on occasion during normal operation of the LHC. The failure does not concern the magnets or the cold masses themselves, but rather their assembly in the cryostat.

These conditions do not happen often, but they were known and were apparently overlooked by the engineers. Fortunately, the functional design of the system appears sound, it's just the design of physical supports that needs to be modified.

While the full cause of the problem is not yet known, failure to account for the asymmetric loads in the engineering design of the magnet appears to be a likely cause.

Contrary to what the submission and article imply, the math was probably fine, but they engineered the design for a less stressful load than it actually experiences in the worst anticipated case.

From 1998 to 2002, Fermilab conducted four engineering reviews of the magnets by experts from Fermilab, other US national laboratories and CERN. The reviews do not appear to have addressed these asymmetric loads. Tests at Fermilab were done on single magnets where such loads do not develop.

Per common practice in large projects, other professionals checked their work, including the customer (CERN). Nobody else thought to account for this case, either. Internal tests did not naturally replicate the failure conditiosn.

Large teams at Fermilab and at CERN are working on understanding and addressing the problem. They are reviewing the documentation, redoing calculations, performing mechanical tests and analyzing the damage to the magnet triplet at CERN.

Stay tuned for more news, while they confirm their failure theory and come up with a fix.