Testing beam-induced quench levels of LHC superconducting magnets

In the years 2009–2013 the Large Hadron Collider (LHC) has been operated with the top beam energies of 3.5 and 4 TeV per proton (from 2012) instead of the nominal 7 TeV. The currents in the superconducting magnets were reduced accordingly. To date only seventeen beam-induced quenches have occurred; eight of them during specially designed quench tests, the others during injection. There has not been a single beam-induced quench during normal collider operation with stored beam. The conditions, however, are expected to become much more challenging after the long LHC shutdown. The magnets will be operating at near nominal currents, and in the presence of high energy and high intensity beams with a stored energy of up to 362 MJ per beam. In this paper we summarize our efforts to understand the quench levels of LHC superconducting magnets. We describe beam-loss events and dedicated experiments with beam, as well as the simulation methods used to reproduce the observable signals. The simulated energy deposition in the coils is compared to the quench levels predicted by electrothermal models, thus allowing one to validate and improve the models which are used to set beam-dump thresholds on beam-loss monitors for run 2.

Figure 1

3D image from neutron tomography of cable interstices in a stack of Rutherford-type cable [8].

Figure 2

MQED (up) and MQPD (down) as a function of beam-loss duration for heat pulses of constant power. The quench levels are computed with QP3 [9] on the horizontal plane of a main-dipole magnet, for the geometrical loss pattern of [10], with magnet currents corresponding to injection beam energy (450 GeV), 3.5 TeV, and 7 TeV.

Figure 3

Overview of the analysis methodology for quench tests.

Figure 4

Vertical trajectory of the injected beam. Black line: MAD-X simulation of injection with $80\mu \mathrm{rad}$ kick by MCBCV.9R2 vertical orbit corrector. Red line: BPM readings for of $80\mu \mathrm{rad}$ kick. Green line: BPM readings for $750\mu \mathrm{rad}$ kick.

Figure 5

Up: Comparison of BLM signals measured during the strong-kick event and simulated with FLUKA. Beam direction is from the left to the right. Down: FLUKA simulated peak energy deposition in the coils, integrated over the event. The gray box indicates magnet cold mass, black boxes indicate the location of BLMs.

Figure 6

Simulated transverse energy density distribution from FLUKA for the strong-kick event in MB.B10R2 coils at the position where the maximum energy deposition occurs. Results correspond to $2\times {10}^9\mathrm{protons}$ impacting on the magnet beam screen. Spatial coordinates are with respect to the center of the vacuum chamber.

Figure 7

FLUKA simulated peak energy density deposited in the coil during the short-duration collimation quench test. The gray box indicates the magnet.

Figure 8

Simulated transverse energy density distribution for the short-duration collimation quench test in the Q6.L8 coils (2500 A) at the position where the maximum energy deposition occurs.

Figure 9

BLM (black) and QPS (red) signals registered during the orbit-bump quench test with intermediate loss duration. A drop in QPS heater-voltage indicates heater firing (green), which was synchronized with the moment of beam dump. The QPS signal measures resistive voltage in a magnet. The 0.1-V intercept of the QPS signal was timed to precede the heater firing by 10 ms, i.e., the QPS evaluation time.

Figure 10

Up: Comparison of integrated BLM signal accumulated during the wire-scanner quench test and simulated with FLUKA. Down: FLUKA simulated peak energy density deposited in the coils over the entire event. Gray boxes indicate magnet cold masses, black boxes indicate the locations of BLMs.

Figure 11

Simulated transverse energy density distribution from FLUKA, integrated over the event, in D4.L4 coils at the location where the maximum energy deposition occurs during the intermediate-duration orbit-bump quench test.

Figure 12

BLM (black) and QPS (red) signals registered during the orbit-bump quench test with intermediate loss duration. A drop in QPS heater-voltage indicates heater firing (green), which was synchronized with the moment of beam dump. The QPS signal measures resistive voltage in a magnet. The 0.1-V intercept of the QPS signal was timed to precede the heater firing by 10 ms, i.e., the QPS evaluation time.

Figure 13

Comparison of MAD-X simulation and data collected by a beam-position monitor during the intermediate-duration orbit-bump quench test.

Figure 14

Up: Detail of the comparison between the BLM signal, accumulated during the intermediate-duration orbit bump quench test, and the simulated signal from FLUKA. Down: FLUKA simulated peak energy density deposited in the coils (black) and MAD-X histogram of protons lost on the beam-screen (green), normalized to the total number of simulated lost particles. The gray box indicates the magnet, black boxes indicate the locations of BLMs.

Figure 15

Simulated transverse energy density distribution from FLUKA in MQ.12L6 coils at the location where the maximum energy deposition occurs during the intermediate-duration orbit-bump quench test. Results correspond to $4\times {10}^8\mathrm{protons}$ impacting on the magnet beam screen.

Figure 16

Measured peak power losses by the beam in the collimation system versus time during the collimation quench tests in 2011 and 2013. In 2013 tests, beams were dumped after achieving the targeted loss rate.

Figure 17

Up: Local excerpt of the comparison of BLM signals and FLUKA simulation results for ramp 3 of the 2013 collimation quench test, both respectively normalized to the BLM of the skew primary collimator (TCP.B6R7) located at $+200\mathrm{m}$ [16, 60]. Down: Longitudinal pattern of the power density (averaged over the inner coil radial thickness) at the transition between cells 8 and 9.

Figure 18

Simulated transverse power density distribution from FLUKA during the steady-state collimation quench test at the MB.A9L7 maximum in Fig. 17 (down). The beam direction enters the figure. Recall from Sec. 3 that in FLUKA coils are extruded, not adequately representing the coil ends.

Figure 19

Comparison of the highest BLM signal time profiles for the 2010 dynamic orbit-bump quench test and the 2013 static orbit-bump quench tests. Note that in 2013 static losses were only achieved in the second attempt.

Figure 20

Up: Detail of the comparison between the BLM signal and the simulated signal from FLUKA for the dynamic orbit-bump quench test. Down: FLUKA simulated peak energy density (black) deposited in the coils and MAD-X normalized distribution of protons lost on the beam screen (green). The gray box indicates the magnet, black boxes indicate the locations of BLMs.

Figure 21

Simulated transverse power density distribution from FLUKA during the dynamic orbit-bump quench test in MQ.14R2 coils at the position where the maximum energy deposition occurs. Results correspond to $2.54\times {10}^9\mathrm{protons}$ impacting on the magnet beam screen. Spatial coordinates are with respect to the center of the vacuum chamber.

Figure 22

Results of particle-tracking and particle-shower simulations. Simulation 1 considers a smooth beam screen, whereas simulation 2 considers a 20-cm-long, 30-$\mu \mathrm{m}$-high aperture restriction close to the magnet center. Up: Detail of the comparison between the BLM signal and the simulated signal from FLUKA for the static orbit-bump quench test. Down: FLUKA simulated peak energy density (black) deposited in the coils and MAD-X normalized distribution of protons lost on the beam screen (green). The gray box indicates the magnet, black boxes indicate the locations of BLMs.

Figure 23

Simulated transverse power density distribution during the static orbit-bump quench test in MQ.12L6 coils at the location where the maximum energy deposition occurs. Results correspond to $3\times {10}^8\mathrm{protons}$ impacting on the magnet beam screen.

Figure 24

Electrothermal quench level estimates for the inner layer of the LHC main bending magnet. Shading in the MQED estimates indicates uncertainties following from the analysis of the intermediate-duration orbit-bump quench test. Shading in the MQPD estimates indicates uncertainties due to the unknown cooling efficiency of intralayer spacers in the main magnets of the LHC.