Choose timezone
Your profile timezone:
10th Quench Behavior Team meeting
→
Europe/Zurich
927/1-B23 (CERN)
Date: 2016-06-21, 09.00-11.00
Place: 927
Presents: J. C. Perez, D. Tommasini, J. P. Tock, M. Pojer, F. Savary, O. Bruning, S. Le Naour, R. Schmidt, A. Verweij, G. De Rijk, M. Bajko, E. Todesco
Possible interventions in LS2 [S. Le Naour]
Sandrine presents a preliminary list of interventions for the LS2 to have an idea about what part of the LHC, if any, could be kept cold during LS2. This is based on today knowledge and, obviously, could increase in the next 2-3 years.
- 11 T dipoles to be installed in 67 and 78 (one each). This is for HL LHC.
- Connection cryostat to be installed in 12 and 23 (one each). This is for HL LHC.
- 3 cases of quench heater failure (one high field circuit lost in each dipole, low field is being used). Sectors 23, 45 and 81. One can survive with any 3 circuits in the two apertures, so redundancy is still available. But as Marta points out, heater failure were sometimes common to the same coil, so if a high field fails the low field could also have the same weakness. Cause: magnet nonconformity.
- Two quadrupoles in sector 34 do not have the associated Landau octupole correctors (not available during repair). Cause: SSS not available after the 2008 incident.
- One quadrupole with very large internal resistance (sector 34). Cause: magnet nonconformity.
- 15 dipoles (all from 2000 series) with internal resistance between 10 and 15 nW in sector 12, 23 and 67. Large intervention, with no evidence from operation of a necessity of replacement. A long term test in SM18 could be envisaged. Cause: magnet nonconformity.
- ULO in 81 – plan (if any) is to be discussed
- Short in some MCS and MCO circuits of sector 78 (two in total): Arjan points out that is quite difficult to localize, but there are worries that the same problem could come in other sectors. Today the circuit is condemned.
- A quadrupole in 78 had two flattop quenches (in the few past weeks) – to be monitored
- A weak MQTL in 56
- Will we recover one or two magnets from the 3150-3200 batch to test them in SM18?
Oliver suggests that having at least one sector cold during LS2 would provide relevant information to the management on the possibility of keeping the LHC cold. A good candidate would be sector 45, taking the small risk of not curing the heater failure of one dipole, and weighting it against the numerous quenches required in this sector to reach 6.5 TeV. Other possibilities would be sector 34, 56 or 81.
Accelerated training [R.Schmidt]
Rudiger recalls the idea, proposed in 2008 and never realized, to increase the delay of the power converter switch off and energy extraction up to few seconds (not more than 5) to be able to quench more magnets at the same time. Rudiger presents a first estimate of the possible gain, based on the present knowledge of training times and recovery times from cryogenics. One could increase the training speed of a factor 2 to 3.
The risk associated to this operation should be carefully estimated. Massive quenches occurred in the LHC a few times (up to 50 quenches in 2008 before the incident), and in case of no failures the accelerator can take these events without risk. The problem is to estimate what can go wrong – in particular the presence of two simultaneous shorts could endanger the machine. It is not the mandate of QBT to explore these topics, but this possibility could be discussed in MP3 with all implications for hardware and commissioning.
Analysis of production data after thermal cycle [E. Todesco]
The second part of the analysis is presented. Data are postprocessed to remove i) the second, third and higher quenches ii) the magnets that were dismantled and reassembled (not done in previous talk). This last step can be controversial, since one could think that dismantled magnet belong to a special batch of non-conform magnets that have been all intercepted (and therefore should be removed from the statistics) or that they are anyway representative of the production (and therefore should be kept).
With this postprocessing, the data at 6.5 TeV are at the edge of being compatible with HC data, as concluded in the previous talk. At 7 TeV the data are compatible for the 3000 and 2000 extrapolation based on the Gaussian distribution (with the 3000 series always at the limit of the statistical error, as for 6.5 TeV). Unfortunately, as noted in previous talk, the sampling is far from being uniform, and relevant parts of the 3000 series have very low sampling. The 2000 series looks more uniform and also the sampling is more significant. A conclusion of this analysis is that a uniform sampling for the behavior after thermal cycle would have guaranteed a better control of the production and given more information on the magnet behavior.
For the 1000 series, due to the lack of HC data, an a priori estimate of 5% quench probability to reach 7 TeV was given at LMC. The ATC data suggest that this could be larger, of the order of 20% (if the sample is not biased). So new estimates for 7 TeV are presented using this number. There is not a significant change, with ~30 quenches needed to reach 7 TeV + 100 A in sectors 12, 34 and 45, plus the second quenches for which a rough estimate would be 1/3 to ¼ more. The same conclusions hold, ie. 45 and 12 would be the ideal candidates, and also 34 could be an interesting sector to push towards 7 TeV.
There are minutes attached to this event.
Show them.