Go to the Indico Home Page Oldest event Older event Up to category Newer event Newest event Printable version
  • Compact style
  • Indico style
  • Indico style - inline minutes
  • Indico style - numbered
  • Indico style - numbered + minutes
  • Indico Weeks View
    • Switch to the management area of this event
    Login

    13th Quench Behavior Team meeting

    → Europe/Zurich
    Hide

    Date: 2016-09-27, 09.00-10.00

    Place: 927

    Presents: G. Willering, A. Verweij, J. C. Perez, P. Ferracin, P. Hagen, R. Schmidt, M. Modena, E. Todesco, R. V. Weelderen

    Quench performance of the main dipoles in SM18: test between 2008 and 2016 [G. Willering]

    Gerard recalls the data about quench performance of the main LHC dipoles after the LHC start up (September 2008). A total of 78 magnets have been tested (including 21 2000 series spares acceptance tests). The goal of Gerard analysis is to draw more information about the behavior after a thermal caycle. For this reason, magnets that have been tested only once have been discarded. The sample consists of 41 magnets, 19 extracted from sector 34 and 7 from LS1 – among them, 7 from the 3000 series (see Table in slide 3).

    • The first observation is that the behavior during LHC commissioning campaign at 6.5 TeV, all manufacturers together, is in line with the biased sample of magnets tested after thermal cycle (plot in slide 5).

    • The first question concerns the impact of a short (few days) versus a long thermal cycle (several months). The analysis is not trivial since the set of short thermal cycle is biased (these were the magnets not reaching nominal performance), whereas the 41 magnets with a long thermal cycle had (in principle) no bias. Long thermal cycle are better behaving for 1000 and 2000 series, whereas for the 3000 series, within the very limited statistics, the training looks longer in the range <7 TeV (action: add the statistical error).

    • The second remark is a comparison of the behavior at the acceptance test and after a long thermal cycle for the 41 magnets. 1000 and 2000 series show a large improvement, whereas the 3000 series shows a relevant degradation. The 3000 series requires 1 quench per magnet to reach 7 TeV (compared to ~0.25 quench per magnet in 1000 and 2000 series). On the other hand, the 3000 series recovers for higher energies, and at ultimate (7.7 TeV) the three series require 1.5 to 2 quench per magnet.

    • The third remark is the correlation between first virgin quench to first quench after a thermal cycle: no correlation is found. Action: it would be interesting to find if there is come correlation between the last virgin quench (max current reached) and the first quench after thermal cycle.

    • Gerard also makes a correlation plot between the gain in current (after thermal cycle-virgin) and the time between cool downs, finding no correlation.

    Summarizing the results, also in these data (i) the 3000 series has a very slow start, but recovers to others at ultimate, and (ii) no evidence of role of long term phenomena in the magnet performance.

     

    Update on 3409 [G. Willering]

    3409 showed a loss of memory after the first and second thermal cycle. In the successive four thermal cycle, all first quenches were above nominal, and in the last thermal cycle ultimate was reached without quench. So it has been concluded that the memory loss is not visible in this magnet (or it is not visible any more) – the magnet will not undergo further thermal cycles.

     

    Update flattop quenches [G. Willering]

    In the past 2000 hours of operation, no flattop quenches have been observed. We had a total time of 3400 hours spent at nominal current, with 5 flattop quenches. A semilog fit is shown, even though it has no physical interpretation.

     

    There are minutes attached to this event. Show them.