Choose timezone
Your profile timezone:
9th Quench Behavior Team meeting
→
Europe/Zurich
Date: 2016-05-31, 09.00-10.30
Place: 927
Presents: P. Hagen, E. Todesco, R. van Weelderen, J. C. Perez, P. Ferracin, D. Tommasini, G. Willering, J. P. Tock, M. Pojer, F. Savary
Update on flattop quenches [G. Willering]
Gerard presents the update on the flattop quenches in Run II operation. In the period March-November 2015, we had five flattop quenches (quenches at operational current corresponding to 6.5 TeV, during stable beams) of five different magnets, all belonging to 2000 series. In November 2015, a powering to 6.5 TeV plus 100 A has been done for sectors 34 and 67: the nominal current was reached with two quenches (one per sector), one of a 3000 series and one of a 2000 series magnet.
After the winter shutdown, in March 2016 all sectors we brought to 6.5 TeV plus 100 A, with only one quench of a 3000 series magnet. Since then LHC operated at 6.5 TeV for three months without flattop quenches. A plot of the number of flattop quenches versus time spent at 6.5 TeV during the 2015-2016 run shows a saturation (slide 3). The conclusion is that the phenomena of flattop quenches exist, but it is rare, it is becoming more and more rare, and is not affecting operation.
A flattop quench in the main quadrupoles was observed in May 2016, in a particularly long run lasting 36 hours at 6.5 TeV. The magnet quenched at flattop operational current of 10361 A.
Test of 3409 [G. Willering]
Gerard recalls the situation for the test of 3409. This magnet is the only 3000 series in surface with series number larger than 3100 (the first 100 magnets of 3000 series show a good quench performance both in SM18 and in the tunnel). The magnet was tested several times, but last three quenches were all above nominal. One the other hand, quenches below nominal, in the 11500 A range, were observed in previous runs. It is suggested to have two more runs with thermal cycle to verify if a quench below nominal happens. In this case, 3409 would exhibit the behavior observed in the tunnel, and therefore should be equipped with quench antenna to be intensively studied. If no quench below nominal is observed, this magnet will be considered not to be representative of the behavior of the slow trainers of 3000 series, and other strategies will be defined (for instance removing some magnets from the tunnel belonging to the worse batch).
Analysis of production data after thermal cycle [E. Todesco]
During the production, magnets not reaching ultimate went for a second test After Thermal Cycle (ATC). The sample is of 136 magnets, 42 of 1000 series, 58 of 2000 series, and 36 from 3000 series. These data were analysed in the past, but in this analysis we inject the information relative to the commissioning of the whole LHC to 6.5 TeV in 2015.
ATC data show that over the sample of 136 magnets, 8 magnets quenched to reach 6.5 TeV plus 100 A: none of 1000 series, 5% of 2000 series and 14% of the 3000 series. So the differences between the three series observed in the tunnel are clearly visible in the production data ATC. The statistical error associated to the 3000 series estimate of 14% probability to quench to reach 6.5 teV plus 100 A is 12% (with two sigma). For the 3000 series, 34% of the magnets quenched in the LHC tunnel, so the trend is there but data are not compatible.
Indeed, the sample of the ATC is far from being homogeneous. Spitting in 8 batches of 52 magnets each, we see that one third of the 3000 series magnets tested ATC belong to best part of the production 3000-3104. Removing this first biased part, we end up with a quench probability of 22%±17%, compatible with the tunnel behavior of all magnets of 34%.
It is interesting to note that batch 3105-3156, belonging to the production part that underwent two crisis (curing and coil protection sheet, see presentation of P. Fessia on 10th December) was intensively tested ATC (10 magnets) showing only one quench. On the other hand, batch 3157-3208, showing the worse performance in the tunnel, and installed in 45, had only two magnets tested ATC, and one of them quenching before 6.5 TeV plus 100 A. So unfortunately we have very little data in the most critical part of the production.
The prediction at 7 TeV for the 3000 series is also compatible with ATC data, once the initial part is removed.
At 7 TeV according to the ATC data, the difference between firms becomes less important. This is not in agreement with the present extrapolation to 7 TeV. Further study will be focused on checking the pattern of the sampling used for the 1000 and 2000 series, and provide estimates where the bias due to a non uniform sampling is removed.
There are minutes attached to this event.
Show them.