Minutes of the 5th LHC Insertions Upgrade Working Group held on 15th November 2007
Present: V. Baglin, F. Bordry, O. Brüning, F. Cerutti, S. Fartoukh, A. Ferrari, M. Giovannozzi, J. Kerby, J.-P. Koutchouk, K.H. Meß, R. Ostojic, L. Tavian, D. Tommasini, R. Van Weelderen, E. Wildner
Excused:
Invited:
1. News and approval of the minutes of the last meeting
The minutes of the last meeting are approved without any comments.
2. Limits of the present cryogenics in IP1/5 and possible "Phase I" upgrades (L. Tavian, ppt slides)
Laurent started his presentation by reminding the main cryogenic limitations of the 1.9 K loop. At the level of the inner triplet, the heat-exchangers and other hardware equipments have been designed to be able to extract a power of the order of 500 W. On the other hand, only 400 W at 1.9 K, corresponding to a ~2KW cryogenic load on the 4.5K refrigerator, are reserved for the inner triplet under ultimate conditions. This is just compatible with the heat load expected in the inner triplet for a luminosity of 2 1034, assuming that the power generated by the debris coming from the IP and by the primary beam (synchrotron radiation, electron cloud) is dissipated into the 1.9 K helium bath. This leaves even some comfortable margin if this power is extracted at 4.5 K (e.g. by a thicker beam-screen). However, the specified and measured capacity of each 4.5 K cryoplants amounts to 23KW while the required ultimate capacity jeopardized by the electron cloud induced heat load has been lately estimated to 28KW for sector 45 (4kW reserved for the RF cavities of IP4) and 24 KW or less in the other sectors. This limitation is therefore rather worrying for phase I but may be at the end a non-issue if the cryogenic budget reserved for the electron multipacting effect (~10KW @ 4.5 K per sector) is found to have been overestimated. RO concluded that while a new refrigerator will not be installed in IP4 for the RF, a net dissymmetry (4kW) will subsist between the left and right side of IP5, possibly impacting onto the performance of the triplet 5L. The first LHC runs will help towards a clarification of the situation.
Laurent then analyzed different possible operating temperatures for the beam-screen (5-20 K as currently, or 40-60 K as proposed by Vincent at the LIUWG#3 to reduce, if needed, the equilibrium gas density by a factor of 2). Also different beam-screen configurations can be envisaged (with two, four or six capillaries or a so-called annular configuration). The 40-60K temperature range is not existing in the LHC but may be eventually produced by sub-cooling the helium coming from the thermal shield cooling circuits. Also, this temperature range does not help in reducing substantially the requirement imposed on the minimum inner dimensions of the beam screen capillaries: 5.5 mm for a configuration with two capillaries @ 5-20 K to extract a power of 10 W per meter of beam-screens (i.e. 400 W for a 40 m long triplet), compared to 5 mm @ 40-60K and 3.7 mm for the nominal LHC triplet beam-screens. In other words, this option will be retained only if there is a strong argument related to the vacuum stability. SF insisted that the configuration with only two capillaries is the most convenient to optimize the beam clearance in the LHC triplets. In this configuration, he then suggested that the aperture loss of 5.5-3.7=1.8 mm in the smallest beam-screen dimension could be significantly minimized by changing the cross-section of the capillaries: e.g. half-moon or elliptical shape (instead of circular) of much bigger area, which should now be allowed by the net increase of the beam-screen absolute dimensions. Laurent a priori agreed.
Finally, Laurent summarized the main modifications of the cryogenic system to be brought in the region of the triplet, namely: the change of the flow characteristics of some valves, the implementation of cryogenic extensions to be compatible with the new longitudinal positions of the triplet magnets and DFBX and, depending on whether the cold option will be retained for D1 and if this new D1 will be connected or not to the cryogenic string of the triplet quadrupoles, the implementation of new service modules and individual valve boxes for the D1 magnets. SF commented that, within a marginal impact on the D1 aperture requirements, installing D1 as close as possible to the inner triplet will always be beneficial for the optics (e.g. D1/D2 integrated strength, total number of parasitic encounters).
RO pointed out that in order to intervene in the region of the triplet quadrupoles and in general in the long straight sections, the two sectors on either sides of the IP must be warmed up, that is four sectors in total for two low-beta insertions. This induces additional costs and delays, and non-negligible technical risks. RO then asked Laurent if the cryogenists planned to install a system able to decouple the arcs from the experimental insertions. Laurent replied that, presently, such a system was not under study.
3. Possibilities of improving the heat transfer from cable to HeII (D. Tommasini, ppt slides)
Davide reminded all the advantages of the Nb3Sn cables with respect to NbTi coils, in terms of temperature margin (a factor of 7 higher under similar conditions, i.e. same field and current density) and power evacuation (about 3 times more efficient assuming the same heat deposition, and similar operational conditions in relative w.r.t. the critical surface B(J,T)). KHM commented however that for a given energy flux impacting the coils, the heat deposition will be higher in the Nb3Sn cables (higher Z compared to NbTi), which may at the end counterbalance the Nb3Sn performance in terms of heat evacuation. AF confirmed that the heat deposition varies indeed linearly with respect to the atomic number of the material.
For the NbTi coils, the first barrier to the removal of the heat generated in each individual cable is the electrical insulation. The latter consists in a winding of two polyimide insulation layers around the cable with an overlap level of about 50% per turn and per layer. This makes the insulation semi-permeable to HeII while the heat evacuation via the convection of super-fluid helium is by far the most efficient way to thermalize the coils. In order to favor the creation of HeII micro-channels of finite cross-section, it is then proposed to change the way the insulation layers are winded around the NbTi cables, with no direct overlap for each individual layers and a third layer playing the role of spacers between the internal and external layers (see Davide's slides for more details). Simple porosity checks using a moderately pressurized air flow shows very promising results, even in the presence of a transverse compression of the cables going up to 50MPa. Then, first tests of heat transfer in cryogenic environment seems to indicate the huge potential of this new approach: a gain of a factor of 5 with respect to the standard LHC cables in terms of capability of power evacuation (e.g. 1 W per meter of cable length) for a given temperature increase inside the cable (e.g. 0.2K).
The audience was impressed by the results obtained. Ranko strongly recommended to study this option under all possible aspects (e.g. electrical behavior) in order to conclude rather quickly on its possible implementation for the NbTi cables which will equip the phase I inner triplets.
4. A.O.B.
Next meeting scheduled for 29 November 2007: Report on IR'07 (F. Zimmermann), Highlights from THERMOMAG (R. Van Weelderen), Reusing the MBXW's for D1 (M. Karppinen).
S. Fartoukh and R. Ostojic