IMPLEMENTATION OF AUTOMATIC ANGULAR COLLIMATOR ALIGNMENT G AZZOPARDI

IMPLEMENTATION OF AUTOMATIC ANGULAR COLLIMATOR ALIGNMENT G. AZZOPARDI Supervised by: B. Salvachua Ferrando and G. Valentino LHC Collimation Working Group 12. 10. 2020

INTRODUCTION § Collimators aligned with BLM devices are currently aligned with zero tilt angle w. r. t. the beam Ignores any tank misalignments Risk of requiring more alignments for system operation with retractions below 1. 5 σ* § The 4 jaw corners of a collimator can be moved individually, so angles in the jaws are a possibility Angular alignments are occasionally used in cases of an isolated issue with individual jaws in operation § Automated procedures were introduced to align collimators at different angles A new FESA class implements 3 methods for automatic angular alignments MD 2504 Collimation (semi-automatic) (fully-automatic)” Technical report, ** * A. Mereghetti. Tested et al. , “ during β*-Reach–IR 7 Hierarchy and Limit. MD 3344 and Impedance. CERN, 2016. ** G. Azzopardi et al. , “Automatic Angular Alignment of LHC Collimators. ” Technical report, CERN, 2017. G. Azzopardi et al. , “ MD 3344 - Fully-Automatic Angular Alignment of LHC Collimators using Machine Learning. ” 2

IMPLEMENTATION - METHOD 1 Angular alignment using reference collimator § This method is based on the standard BBA procedure § The selected collimator is aligned with parallel jaws at different angles each time § The reference collimator (TCP) is aligned before and after each angular collimator alignment to create a reference halo § The angle where the jaws are parallel to the beam, is calculated by extracting the angle which generates the smallest measured beam σ Ø This method identifies when there is a real offset in the G. Azzopardi et al. , “ Automatic angular alignment of LHC collimators tank as the jaws are kept parallel 2017. . ” ICALEPCS 2017, pages 928– 933, 3

IMPLEMENTATION - METHOD 2 Angular alignment at maximum angles G. Azzopardi et al. , “ Automatic angular alignment of LHC collimators 2017. . ” ICALEPCS 2017, pages 928– 933, 4

IMPLEMENTATION - METHOD 3 Angular alignment using a jaw as reference § One jaw of the selected collimator is used as reference to define the beam halo - no reference collimator required § The other jaw is aligned at different angles each time § The reference jaw is aligned before and after each angular jaw alignment § The angles where the jaws are parallel to the beam, are calculated by extracting the angles which generate the smallest measured beam σ Ø This method determines the most optimal angles for both jaws independently thus can detect asymmetries in the G. Azzopardi et al. , “ Automatic angular alignment of LHC collimators collimator itself 2017. . ” ICALEPCS 2017, pages 928– 933, 5

MD RESULTS EXAMPLE § The best angle can indeed be obtained using each of the three methods § It is possible for them to converge to the same angle G. Azzopardi et al. , “ MD 3344 - Fully-Automatic Angular Alignment of LHC Collimators using Machine Learning. ” 6

? METHOD 4 § The selected collimator is aligned one jaw at a time, with each corner individually until spike § No reference collimator is required § Used to determine the TCSG. D 4 L 7. B 1 angle - D. Mirarchi Ø Quick method requiring only 4 alignments Ø Not thoroughly tested and compared to other methods 7

SOFTWARE STATUS New FESA class implements 3 methods for automatic angular alignments Achievements Limitations § Fully-automatic alignment (using ML) § § Provides 15 -20 alignments (at different angles) per collimator in ~10 minutes The FESA navigator must be used directly to control the angular alignments § A GUI has been temporarily designed to monitor the jaw movements and BLM losses as the collimators are aligned at different angles 70% faster than current procedure § No human intervention required and provides instant results 8

SOFTWARE STATUS 9

SOFTWARE STATUS 10

PROPOSED SOFTWARE UPDATES ? § User interface to align multiple collimators using selected angular alignment method § Incorporate crosstalk-model to handle parallel selection of collimators to align simultaneously Currently user must take care of this § Setup s heet to save centres and beam sizes and plot results § Automatic angular alignment tab for each collimator 11

DISCUSSION POINT ? MD 3344 on 27 -10 -2018 The 3 methods converged to ~200 μrad The BPMs aligned at an angle -350 μrad How to proceed with this discrepancy? G. Azzopardi et al. , “ MD 3344 - Fully-Automatic Angular Alignment of LHC Collimators using Machine Learning. ” 12

THANK YOU FOR YOUR ATTENTION!

ANGULAR METHODS CALCULATIONS 1. 2. §l If there is a misalignment of the collimator w. r. t. the beam, the measured beam size of the collimator will be larger than the real one § Calculate jaw gap to calculate measured beam size (using half gap) §l Calculate upstream and downstream centers separately § The difference between the two can be converted to an angle using the small angle approximation 3. §l Calculate the measured beam size at each of the aligned angles § Δx i j = average of up and down positions of jaw aligned at angle j (mm) Δx i 0 = centre of collimator jaws when aligned at 0 angle (mm) G. Azzopardi, “ Automation of the LHC Collimator Beam-Based Alignment Procedure for Nominal 14

ANGULAR ALIGNMENT USE-CASE § Loss maps to analyse beam loss distribution in Beam 1 § Hierarchy breakage caused by TCSG. D 4 L 7. B 1 (-350 μrad) : 02/12/2017 § Semi-automatic angular alignment applied one jaw corner at a time: ~40 minutes G. Azzopardi, “ Results of Automatic Collimator Alignments. ” CWG 07 -05 -2018. 15

TCSG. D 4 L 7. B 1 MD RESULTS § 02/12/2017 alignment: Mirarchi -350 μrad by D. § New methods: Right = 300 μrad, Left = – 500 This suggests an asymmetry within the collimator § Note: According to the collimation angular tilt convention, the right jaw angle is the negation of the left jaw angle when the jaws are parallel, but to plot them together to determine a single optimal angle, the inverse of the right jaw angles are plotted. For the TCSG the right jaw needs to be taken individually, therefore G. Azzopardi et al. , “ MD 3344 - Fully-Automatic Angular Alignment of LHC Collimators using Machine Learning. ” when using the convention the most 16

SOFTWARE ARCHITECTURE G. Azzopardi, “ Automation of the LHC Collimator Beam-Based Alignment Procedure for Nominal Operation. ” Ph. D thesis, 2019. 17

MD RESULTS EXAMPLE 1 collimator at 41 angles using 3 methods @Injection: ~25 minutes G. Azzopardi, et al. , “ Operational Results on the Fully-Automatic LHC Collimator Alignment. ” PRAB, 2019 18