Minutes of the 17th LHC Insertions Upgrade Working Group held on 17 July
2008
Present: B. Auchmann, C. Bertone, E. Bielert, F. Bordry, F. Butin, J. Coupard, R. Denz, J. Kerby, N. Kos, H. Mainaud Durand, K. Foraz, J.C. Guillaume, K. H. Meb, Y. Muttoni, D. Niesbet, R. Ostojic, H. Prin, P. Proudlock, S. Roesler, S. Russenschuck, N. Schwerg, T. Taylor, J.P. Tock, L. Walckiers, S. Weisz, E. Wildner.
1.
General information
Ranko reminded the meeting that the conceptual design review is scheduled for 31 July. In this context the meetings of today and of the 24 July are planned to review the remaining outstanding issues, one of which is the draft planning for the implementation of Phase-I. In a recent LHCC meeting, the experiments, the injectors (Linac4 and PSB) and the LHC have agreed to tentatively plan their upgrade activities for the period Nov 2012-June 2013, which will be followed by a shorter physics run in 2013. Any Phase-I upgrade activities should keep this guideline in mind.
2. IR upgrade implementation: activities and preliminary planning for the shutdown of 2013 (K. Foraz pptx file)
Katy presented a first draft of the schedule for replacing the LHC triplets in the ATLAS and CMS insertions. The schedule is based on the experience gained during the LHC installation and includes the activities starting with the sector warm-up up to and including the final pressure and leak test. The estimated time needed per IP side (triplet) is about 30 weeks without taking into account any modification on the QRL (estimate not known at present). The total duration for the four triplets in IP1 and IP5 is estimated between 34 to 42 weeks depending on the available tooling (1 or 2 sets) for disconnecting the present triplets. This assumes no limitation of manpower. It is also assumed that all four sectors can be warmed-up at the same time and that the full quantity of helium can be stored appropriately.
Katy listed the activities in the main ring and in the UJs, their estimated duration and responsible engineers/group. It is assumed that the work can be carried out in parallel in two triplets and is divided in three consecutive global activities: the removal of the present triplets, D1 and other equipment; the installation of the new magnets and other equipment; and finally the interconnections (including the tests).
François confirmed that the existing scenario to replace the TAS is independent from the activities in the tunnel as it will be performed from the experiment cavern side. Replacement of the TAS requires about two weeks and does not interfere with the opening of the experiments.
Jean-Philippe commented on the time necessary for disconnecting the present triplets: the four weeks per triplets take into account that the operations have to be performed with care to allow reconnection of the quadrupoles, which will be kept as spares for IR2 and 8. He pointed out the need for new tooling which needs to meet the ALARA requirements. Concerning the new interconnections, the time necessary is based on the present experience and is scaled up by the number of interconnections.
Ranko asked if any activity has been identified which could be preformed during the shut-down in 2012, preceding the Phase-I upgrade. Sylvain stressed that the passage for eventual transport of magnets has to be ensured and he warned that the shielding must be kept in place. Jean-Claude remarked that preparations for new cable trays could be done during the preceding shut-down. He also noted that the removal and re-installation of cables has to be taken into account in the planning. It was also remarked that the cool-down and commissioning of the sectors is not taken into account in the present schedule.
3. Protection of the Triplet, D1, Correctors and Superconducting Links/Leads (R. Denz ppt file)
Rainer presented the global electrical protection scheme for the phase I triplet circuits. He recalled that the protection systems should be installed in radiation free areas if at all possible. He gave an overview of the layout based on bridge type protection system for each individual magnet with dedicated instrumentation and electronics for the protection of the superconducting links and current leads. Digital systems are preferred if the risk from SEU is acceptable. Three options are considered for the by-passes depending on the powering scheme: no by-pass, if the quadrupoles are individually powered; using thyristors on the warm side of the circuit, close to the feed box; and finally using cold diodes connected at the 20 K part of the leads, within the feed box. The use of cold diodes in the magnets (as in the LHC dipole/quadrupole) was not recommended due lack of rad-hard devices. Instead of thyristors, warm diodes could be envisaged but with the inconvenience of high leakage current.
Rainer concluded that it would be advantageous to include energy extraction in all circuits. In that case, semi-conductor switches and extraction resistors similar to the main quadrupoles ones could be used. The question was raised whether quench heaters would be needed if the energy extraction is considered as sufficient. The opinion of the meeting is that quench heaters and energy extraction should be foreseen to increase the redundancy of protection. Rainer remarked that the new QPH power supplies should include pulse shaping, so as to limit shocks in the quench heater connections at the start of the discharge, while providing the same power. These units, as well as detection electronics could be placed far away from the magnets.
4. Alternatives for the triplet powering (D. Nisbet ppt file)
David
presented several options for powering of the triplet taking into account the
electrical complexity, the associated protection scheme, the volume
requirements and the total cost. He recalled the three proposals of powering
strategies that were presented in the 7th
LIUWG meeting : (i) a single circuit where the
four quads are powered in series, (ii) an individual powering of each quad, and
(iii) a mixed solution consisting in two circuits with Q1-Q3 and Q2a-Q2b in
series with an additional trim. The first proposal is not retained due to lack
of flexibility even if the cost and volume are the lowest. For the two
remaining proposals the possible options were listed with their advantages and
drawbacks. Seven options were commented and classified in two main schemes:
individual or split powering. The current decay time, the number of current
leads and the quench protection scheme vary among the options but the main
difficulty is to locate the 7.5 to 24 racks, depending on the option, in the
limited volume in radiation free areas. Another difficulty, common to all
options, is the control of the negative ramp rate at low current, and the need
to deal with inductive coupling. To simplify the inductive and resistive
coupling and reduce the volume for power converters, three split powering
options were commented. The number of racks in these cases is 14. The meeting
considered the split powering option with energy-extraction as the best
compromise and suggested that a tunnel layout is studied for it.
For the
cold D1 additional PC racks are needed. For the correctors, the intention is to
reuse the existing power converters as much as possible.
David
raised the question whether a cold link to the surface would be feasible for Phase-I.
Tom commented that an SC link based on MgB2 is a new development which needs to
be fully tested and validated on lengths of 10-20 m before extrapolating the
technology to long stretches. Vertical links, very attractive for the future,
require in particular extensive validation.
5. First studies of the triplet
quench protection (
Nikolaï reported on the first quench simulations made for the 120 mm aperture quadrupole. As a preliminary step, a magnet without any protection was simulated; as expected, the current decays very slowly and the temperature in the coils rises above 600 K before the current is half-way down. Quench heaters were then included in the model. The effective power and delay time of the heaters were chosen to reproduce the measurements on the LHC dipoles. The simulations show that in this case the quench distributes quickly across the coil and the temperature reaches only 160 K.
A third case was studied with a dump resistor and switch in the circuit. Varying the voltage across the resistor, it was found that bellow 200 V the dI/dt is too small and that there is no quench-back. In case of a 500 V resistor the quench-back is such that the quench distributes quickly in the two layers so that the temperature stays bellow 120 K. Cases with half of the nominal current and longer quench heater delay show that the magnet is equally well protected in these conditions. It was recommended in the discussion to increase the voltage well above the quench-back threshold (e.g. to 1 kV) to gain in safety margin and in extracted energy. Louis suggested a test campaign to validate the model and Stephan noted that the simulation parameters have to be updated to take into account the pre-stress in the quadrupole coil (different from the main dipole).
Nikolaï concluded that either quench heaters alone, or a dump resistor alone, as well as both systems together result in low peak temperature after quench for wide variety of parameters. He briefly presented the new version of the cross section and a different quench heater layout, which he did not have time to analyse. Stephan insisted that the input files received from the magnet designers need to have consistent winding and numbering schemes to allow efficient quench protection simulations for future variants.
R. Ostojic and H. Prin