LHC Beam screen Heat Loads Benjamin Bradu BEICS
LHC Beam screen Heat Loads Benjamin Bradu, BE-ICS On behalf of BE-ICS and TE-CRG E-cloud meeting 12 th December 2016
Content ■ Main interrogations in the current data we observe 1. 2. Variability of heat loads between half-cells in the same sector Variability between sector heat load averages ■ Heat load details in instrumented half-cells in sector 45 ■ Heat transfer test Note: all heat load plots and data in this presentation have been re-computed offline using the same algorithm and the same parameters for a fair comparison. B. Bradu. Beam Screen heat loads
1. Variability between half-cells ■ Valve characteristics are not well know over the machine Ø Ø ■ Temperature sensors can provide wrong information Ø Ø Ø ■ Falling sensor: the half-cell is automatically discarded in the average. Wrong calibration/installation. Few cells per ARC seems concerned (to be investigated). Normally, precision < 0. 1 K < 1% of error on the beam screen heat load. “Calibration test” setup by CRG-OA in 2016 without beam: Ø Ø Ø ■ Valve rangeability and pre-constraint uncertain. Induce error in the valve massflow calculation error in the heat load calculation. Valve is positioned at the same aperture than a reference fill (ex: 50%) Electrical heater is powered such that BS temperature is the same than reference fill (ex: 20 K) Electrical heater power is compared to the Qdbs calculated (ex: 100 W Vs 109 W ɛ = 9% ) Test done in ARC 12, 23, 78 in 2016 Ø Ø Ø Error statistically distributed over the sector. Error is always the same between 2 identical fills. Error evolves with the valve aperture. Max error per half-cell in ARC 12 for ~2000 bunches fill: +/- 30 %. Average heat load relative error over ARC 12 : + 3%. B. Bradu. Beam Screen heat loads
1. Variability between half-cells ■ 30 Results already presented by E. Rogez in S 12 error over BS loop s 12 20 10 0 -10 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 -20 -30 -40 B. Bradu. Beam Screen heat loads
Error behaviour ■ Error of valve characteristic evolves with valve opening (so with cooling power) Ø Example for an error evolution with +/- 30 % on the valve rangeability and on the valve pre-constraint S 45 Average In S 45 = 35% S 23 Average In S 23 = 53% S 45 Average In S 23=105 W Average In S 45 = 60 W B. Bradu. Beam Screen heat loads S 23
2. Variability between sectors ■ Possible reasons: E-cloud difference across sectors ? Ø Vacuum experts say “improbable”… ü ABP experts say that the power profile injected in RF per bunch in 1 turn is the same profile as e-cloud signature… ü Unknown phenomena across sectors ? Ø Image current and Sync. Radiation checked and are OK. ü Dark Heat Load in the LHC ? (to be discussed with dark matter experts ) ü Error in the heat load calculation between sectors ? Ø Ø Next slide. . B. Bradu. Beam Screen heat loads
2. Variability between sectors Why it CANNOT be a calculation error ? ■ When there is no e-cloud (50 ns) The calculation gives similar results in ALL sectors. The result fits theoretical Synch. Radiation and image current losses. Ø Ø ■ The total LHC BS heat load calculated is compatible with the power re-injected in the RF the total sum is correct ! Ø ■ We obtain different results in 4 sectors before and after LS 1 with the same algorithm. ■ Let’s imagine that the cold mass cools the beam screens and can explain the differences Should be statistically distributed over the machine, not per sector Conduction heat transfer is proportional to BS temperature and we don’t see it when no beam at 20 K… Should see a significant 1. 8 K additional heat load. Ø Ø Ø ü ■ Not at all the case: for instance, fill 5045 (38 hours in June 2016), in ARC 34 (lowest BS heat load after LS 1) the coldcompressor flow was very stable during the 38 hours of the fill (65 g/s) no additional 1. 8 K load. Conclusion: Ø Ø It cannot be due to a wrong average heat load calculation ! There is a REAL heat load difference between sectors, which is about a factor 2 today ! B. Bradu. Beam Screen heat loads
Check without e-cloud (50 ns) ■ Fill 3134 (6 th October 2012) Ø 50 ns_1374_1368_0_1262_144 bpi 12 inj @ 3 Te. V Theory 11. 3 W B. Bradu. Beam Screen heat loads
Before & After LS 1 ■ Fill 3429 (13 Dec 2012): Ø Ø Ø ■ 25 ns_804 b_72 bpi_12 inj_2012_MD @ 4 Te. V Intensity/beam = 0. 857. 1014 p+ Average ARC e-cloud (=Qdbs-Qic-Qsr) = 44 W/hc Standard deviation between sector average: 5 W/hc Max difference between sectors: 37 % Fill 4363 (14 Sept 2015) Ø Ø Ø 25 ns_745 b_733_565_580_144 bpi 8 inj @ 6. 5 Te. V Intensity/beam = 0. 772. 1014 p+ Average ARC e-cloud (=Qdbs-Qic-Qsr) = 55 W/hc Standard deviation between sector average: 11 W/hc Max difference between sectors: 92 % +115% +48% +8% +9% B. Bradu. Beam Screen heat loads +15% +35% +71%
Before & After LS 1 ■ LHC half-cell total heat load details (485 loops) S 12 S 23 S 34 S 45 S 56 S 67 S 78 S 81 B. Bradu. Beam Screen heat loads
During 2016 ■ Fill 5043 (25 th June 2016) Ø Ø Ø ■ 25 ns_2076 b_2064_1717_1767_96 bpi_23 inj @ 6. 5 Te. V Intensity/beam = 2. 45. 1014 p+ Average ARC e-cloud (=Qdbs-Qic-Qsr) = 71 W/hc Standard deviation between sector average: 29 W/hc Max difference between sectors: 125 % Fill 5451 (26 th October 2016) Ø Ø Ø 25 ns_2220 b_2208_1940_2036_96 bpi_24 inj @ 6. 5 Te. V Intensity/beam = 2. 35. 1014 p+ Average ARC e-cloud (=Qdbs-Qic-Qsr) = 48 W/hc Standard deviation between sector average: 24 W/hc Max difference between sectors: 132 % -29% Average decreased by 31% in 4 months -16% -27% -26% -30% -38% B. Bradu. Beam Screen heat loads -48%
During 2016 ■ LHC half-cell total heat load details (485 loops) S 12 S 23 S 34 S 45 S 56 S 67 S 78 S 81 B. Bradu. Beam Screen heat loads
Extra-instrumented cells in S 45 ■ 3 half-cells in S 45 with intermediate temperatures Ø Ø Ø 1 sensor seems dead in 34 R 4: need correction to take only 1 sensor between the 2 dipoles (EYETS). 1 half-cell (14 L 5) is in opposite direction and online calculation was wrong: need correction (EYETS) Strange sensor values in 14 L 5: wrong calibration/installation ? B. Bradu. Beam Screen heat loads
12 R 4 half-cell ■ Offline re-computation of each magnet heat-load Ø Fill 5030 (20 th June 2016): 2076 bunches at 25 ns and 6. 5 Te. V - Similar dipole behaviours Quadrupole very small heat load Total almost respected Qdbs TOTAL DIPOLE QUADRUPOLE B. Bradu. Beam Screen heat loads
32 R 4 half-cell ■ Offline re-computation of each magnet heat-load Ø Fill 5030 (20 th June 2016): 2076 bunches at 25 ns and 6. 5 Te. V - Quadrupole very small heat load - Dipole LBALA_34 L 5 is VERY strange but: - Calculation seems OK - Temperature sensors before/after seems OK - Total almost respected - Reproduced in all fills Qdbs TOTAL DIPOLE QUADRUPOLE DIPOLE B. Bradu. Beam Screen heat loads
13 L 5 half-cell ■ Offline re-computation of each magnet heat-load Ø Fill 5030 (20 th June 2016): 2076 bunches at 25 ns and 6. 5 Te. V - Similar dipole behaviours Quadrupole very small heat load Problems of temperature sensors calculation error on quadrupole Qdbs TOTAL DIPOLE QUADRUPOLE B. Bradu. Beam Screen heat loads
Heat transfer test ■ 1 test was made in Sector 2 -3 in open loop Ø Ø Half cell 31 R 2_947 on 24 th November 2016 Needs 7 hours for 2 points at 26 K and 45 K Normal operation with beam = 1 W Theory = conduction + radiation losses as described in the LHC Project Note 330 (L. Tavian, Beam screen regenerative heating: cryogenic impact and feasibility, 2003) B. Bradu. Beam Screen heat loads
Conclusion ■ Variation in a sector Ø Ø ■ Variation between sectors Ø Ø ■ Not due to a calculation error There is a real heat load difference Sector 45 instrumented cells: Ø Ø ■ Partly because of not well known valve Can come from bad sensors (failure or wrong calibration/installation) Instrumentation to be consolidated in the current instrumented half-cells Extension to other half-cells would be interesting Heat transfer test Ø Ø One half-cell done Need more slots to perform other tests B. Bradu. Beam Screen heat loads
- Slides: 18