Delay for the LHC

By John Conway | October 16, 2008 12:02 pm

As of my last post on the LHC, following the quench incident on Sept. 19, little was known about the cause of the incident nor the extent of the damage. In fact, so much liquid helium was released into the tunnel in the affected sector of the machine that it was too cold for people to enter until last week. A colleague of mine quoted the team who entered the sector as saying that when they got to the magnets that were affected, “it wasn’t a pretty sight”.

CERN has released a report today giving an initial summary of what happened and the extent of the damage, and indeed it turns out to have been quite serious: 24 of the long dipole magnets and 5 of the quadrupole magnets (which focus the beam) suffered serious mechanical damage when the liquid helium enclosure between two of the magnets ruptured, allowing helium into the vacuum jacket surrounding the enclosure. This led to a chain of events which resulted in an extreme overpressure in the vacuum jacket on a long chain of magnets, seriously damaging them. The forces were great enough to physically tear the magnet stands out of where they were bolted to the concrete floor.

Normally CERN shuts down accelerator operations every winter due to the high cost of electricity in the winter (Europeans mainly use electricity for home heating) and this was foreseen for December. Repairs to the damaged sector will proceed in parallel with the previously scheduled work during the shutdown. But clearly any hope of high energy colliding beams in 2008 was lost following this incident, and it looks likely to be months before the machine will turn on again. At that point, there is still a many-week period of commissioning before the machine can collide protons at high energies, probably 10 TeV initially.

But now the tough questions: what was the ultimate cause of this incident, and what can and must be done to prevent a similar occurrence in the future? The report concludes optimistically that improved quench detection systems and increased pressure relief devices will ensure safe powering of the machine. I take that to mean that they will do a retrofit on the entire 27 km accelerator. I guess we’ll see how long it really takes. The Director General’s statement accompanying the report only says the machine could be restarted “in 2009″…

CATEGORIZED UNDER: News, Science, Technology
  • Pingback: Relatório sobre falha do LHC « Ars Physica()

  • Ellipsis

    I hope the DG will quickly reach out to the world community, including the members of the experiments and technical staff at other laboratories, to assist the accelerator staff at CERN to fix this. There are plenty of people on ATLAS, CMS, LHC-B, and ALICE that would be happy to assist the shift crews that implement the retrofit (after taking any needed training), and any other help that is needed.

  • Sili

    This being Europe I sincerely doubt that even if Health & Safety would allow academics to help out the plummers and engineers, then the unions would likely veto it. The last thing we want now is a strike.

    24 dips and 5 quads – out of how many in total? And is this more or less than what they have of off-the-shelf replacements?

    Europeans mainly use electricity for home heating

    I get the point, but that’s like saying it never snows in Minnesota because there’re droughts and fires in California.

    That may well be true in France where they have plenty of nuclear power, but it’s pretty much a last resort up here in the Denmark. I can’t speak for Norway with all the hydroelectrics, though.

    (For those interested most non-urban settings use oil or gas if they’re connected to the net. Larger cities mostly have ‘distant heating’ from coalfired powerplants – hot water is piped through most of the urban area. As I recall it it does lower the efficacy of power production a bit, but the overall effiency of the coal is increased through the double usage.)

  • Ellipsis

    Hi Sili,

    Yes, that’s true — hopefully the unions will agree that getting needed assistance at a critical time will be more likely to save their jobs than refusing it.

  • James G

    Hey, who would have thought it, the first discovery of the LHC – reasons why pressure relief devices need to be really really great when you’re dealing with near -273 kelvin .

    On a more optimistic note, this is good news, since it’s gonna trake me another year or so to develop my ground-breaking science that saves the world theory.

  • James G

    ^^ obviously I meant -273 centigrade :)

  • not_entirely_surprised

    During my days at CERN I had a chance to interact with the magnet engineering division. We are talking almost 15 years ago, early days of the LHC project. Back then these folks thought that we (the physicists) had gone nuts, their reaction to the technical specs they were being asked to deliver. This is a quote from the division head at the time, dropped during a candid conversation. I am hoping things have changed in the interim, making this the first and last of such episodes.

  • Adam A.

    It’s not like we’ve never used super conducting magnets before. At Fermilab, the magnets also must be cooled by liquid helium, so the difficulties weren’t unexpected. However, it is true that every experiment poses its own challenges and that the LHC magnets are more complicated then normal due to having two beams of protons. In addition, the first time everything is hooked up, it’s always possible that someone made a mistake – you do everything you can to check multiple times, but with thousands of magnets, it’s always possible for something to slip by – the latest emails assert that there was a problem with the connection between two magnets which lead to this much larger problem.

    Near the beginning of Run II at the Tevatron, there was a much more rapid abort of the beam than had been anticipated. The control systems that steer the beam safely out of the detector couldn’t react fast enough, and the beam cut through a tungsten beam collimator. After that, the accelerator division figured out what to do to make sure that never happened again. Live and learn.

    So damaging that many magnets is very bad, but remember that we are dealing with machines that very desperately want to destroy themselves. The energy contained in the magnetic field of the magnets is immense – an improperly handled quench can melt the coils of the magnet. Rotating beams aren’t stable – they are always trying to expand their radius and tiny imperfections in the magnets (that the beam passes many, many times per second) make the deviations of the beam impossible to calculate a priori. Just remember to keep all of this in perspective – while every one hoped it would be easy, I don’t think anyone expected it to be.

  • Pingback: LHC status report « Peculiar Velocity()

  • ObsessiveMathsFreak

    Is there no other alternative to supercooled magnets? If every hiccup takes six weeks to repair, it could be well worth looking into more robust methods of guiding those protons.

  • Lab Lemming

    Well of course they need a lot of electric heating if they flood the entire joint with liquid helium. That stuff is cold!

  • Seth Zenz

    ObsessiveMathsFreak, sure there’s an alternative: using weaker conventional magnets, paying more for electricity, and tripling or quadrupling the circumference of the LHC!

  • wb

    The report notes that there are sufficient magnets to replace the damaged magnets. From the installation schedule just completed the change-out time can be accurately estimated. So why isn’t that time given in the report? The report notes that the beam tubes have been contaminated to a limited but not known distance. This certainly adds an unknown delay. Presumably there is a procedure and plan already to clean the beam tubes. It would be good to make these known beyond the CERN inner circle.
    It seems clear that the pressure relief system did not work as well as planned; hence the requirement for an upgrade. Hopefully these can be upgraded without warming the undamaged sectors. It seems that quench heaters behaved well, although the conclusion of upgrading the protection system, may suggest that some heaters were triggered unnecessarily; still better safe than very sorry.
    The fact that the other sectors had been run at high power without incident is good news suggesting that the bad connection is singular or such connections are limited to sector 3-4. Thermally cycling the entire ring is undesirable and adds downtime.
    Overall the report is useful to the general HEP community, but many questions are left unanswered. To be sure CERN management does not want to specify a repair duration, yet it could have been more forthcoming about a lower bound for remedies.
    When one considers that the storied energy in the magnet system is of order 10 GJ (not including the cryogenics), it does not take much to cause considerable damage.

  • ObsessiveMathsFreak

    ObsessiveMathsFreak, sure there’s an alternative: using weaker conventional magnets, paying more for electricity, and tripling or quadrupling the circumference of the LHC!

    Expensive I know, but if managing superconducting magnets at this scale proves to be beyond our feasible capability, perhaps these measures should ultimately be considered when building the next big particle accelerator. Lost time costs money as well.

  • none of the above

    One shouldn’t be too pessimistic. They have all the spares necessary to do the replacements, and 5 months to do them. They’ve already shut down the North Area fixed target program so that they could do the projected maintenance on the injection chain starting two months early, with a view to moving the starting date for LHC running from June to April, next year.

  • Count Iblis

    Did the physicists neglect volume 1 of the “Book of Physics”, see end of page one and top of page 2 of this article? :)

  • jfb
    is the LHC design report. There are 1232 main dipoles.

    Each of the eight 2K refrigerators at LHC is slightly larger than the previous record holder at Jefferson Lab. If there were more connections from 2K to room temperature the refrigeration costs would go up substantially. This is part of the risk-benefit tradeoff that was done during design. There are eight ~3km cryo loops.

    Burst disks for pressure relief are graphite machined so they will burst at a precise pressure differential. When I was in the MRI magnet business two decades ago we specified the things to burst at 19-21 psi differential if memory serves. Bigger burst disks mean bigger stainless pipes and flanges; these cost a lot more than the graphite disk. Retrofitting bigger disks to existing magnets, if that is what’s planned, will be touchy because once can’t afford to get chips into the super-insulation. One will have to warm up the magnets to increase the size of the disks because one has to bring the insulation vacuum space to atmosphere.

    Again, one models the anticipated failure modes, determines the gaseous helium flow rates needed to keep over-pressure down, and sizes the pipes and burst disks with appropriate engineering safety factor. Too big a disk and the cost is too high. CERN had to borrow from their employees pension fund to get the funds needed to build LHC – the member states wouldn’t increase their contributions enough to fund it otherwise. Cpst containment was VERY important.

    The 5 TeV energy for the initial run was chosen because all of the sectors exceeded the field needed for this energy in sector by sector tests. All the magnets trained to fields above those needed for 7 TeV at acceptance, but some relaxed at room temperature or during installation.

    I am sure CERN will work hard to determine what caused the interconnect to fail this time rather than during previous excursions to similar currents.

    Using steel magnets one would need a ring ~150 km in circumference for this energy. The LHC tunnel was dug for LEP so only the enlarged detector caverns were needed WRT civil construction. Superconducting magnets are the only solution for hadron machines at these energies.

    The next machine, if built, will be a linear collider. The larger international effort plans to use superconducting RF cavities to accelerate electrons and positrons.
    The smaller effort, CLIC at CERN, works with room temperature RF.

  • jfb

    Corrections to comment 17

    I hadn’t read the four page report before writing, just the press release.

    Pressure relief is via spring loaded disks in the short straight sections.

    This sector was the only one which hadn’t been taken to 5.5 TeV equivalent currents prior to the incident.

  • Blake

    W00t! SUSY Higgs at the Tevatron in late ’09 early ’10 or before 15 fb-1 acquired! am I right kids?

  • Harbles

    The fundamental cause of the incident is the failure ( going open circuit ) of an electrical connection between two superconducting magnets. I imagine the only thing left of said connection is that fine black soot that has contaminated the beam tubes in a large area either side of the failure. I have been trying to find out the details of how does one make connections between superconducting magnets or buss bars. Is the technique used in the LHC an old tried and true method with a great deal of engineering experience behind it or a new innovative method that has less time in use. If it’s an old method one can only hope that this particular incident was the result of improper installation and not a fundamental flaw with the connection method that is used some 6000 times in the entire collider.

  • Harbles

    Eureka! I have found the holy grail of electrical connections.

    Over 10000 13KA connections alone!

  • Harbles



  • Pingback: LHC - Take 2 | Cosmic Variance | Discover Magazine()


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Cosmic Variance

Random samplings from a universe of ideas.

See More

Collapse bottom bar