ILC Simulation using BDSIM Grahame Blair On behalf
ILC Simulation using BDSIM Grahame Blair On behalf of: John Carter and Ilya Agapov March 12 th 2006 LCWS 06
BDSIM and collimation/background simulations for ILC Ilya Agapov RHUL BDSIM status ● BDSIM geometry description concept ● RHUL Repository ● Other issues ●
BDSIM status BDSIM is a Geant 4 -based beamlne simulation toolkit ● Fast transportation and comprehensive set of physics processes ● Current version 0. 1, minor releases planned approx. every month ● Number of users growing ● http: //flc. pp. rhul. ac. uk/bdsim. html ● CVS repository ● Ilya Agapov
Functionality „almost every“ beamline geometry ● „almost all“ physics processes ● Tracking in „almost arbitrary“ field configuration (except probably 3 d field maps which are presently too slow) ● Major concern – validation ● Current performance philosophy – use Grid ● Ilya Agapov
BDSIM physics Transportation ● e+/e- physics ● Muon physics ● Synchrotron radiation ● Hadronic physics ● „Physics lists“ can be created as sets of processes of which BDSIM has plenty ● Several predefined physics lists available ● Ilya Agapov
BDSIM validation Tracking cross-check with DIMAD (with O. Dadoun et. al. , LAL ) minor disagreements in higher-order magnets. Under investigation. ● Tracking cross-check with STRUCT (with F. Jackson, A. Drozhdin) in progress. Preliminary results agree well. ● Planned hadronic physics cross-check, joint effort with Geant 4 collaboration. Possible connection to ILC dump studies. ● Ilya Agapov
BDSIM plans Some new physics processes – gas scattering, wakefield kicks (1 -2 months) ● Fast algorithms for tracking in fields described by 3 d maps under study (6 months) ● Activation and dosimetry (1 yr) ● Introduce more „tallies“ (currently only sampling and energy loss histograms available) (6 months) ● Introduce Open Cascade CAD with GDML interface (1 -2 yrs) – ESAled project ● Further validation (2 -3 yrs) ● Ilya Agapov
BDSIM geometry description concept Current BDSIM input format is GMAD ● Any other formats are supported as external „drivers“ ● Mokka and GDML drivers are available in BDSIM ● Ilya Agapov
GMAD Extension of a MAD subset to support geometry features ● Commands for run control, process cuts etc. ● Provides „drivers“ to other geometry formats, so arbitrarily complicated geometries are possible ● Can be used as a standalone parser library ● Ilya Agapov
GMAD example qd : quadrupole, l=0. 5 * m, k 1 = qdk 1; qf : quadrupole, l=0. 5 *m, k 1 = qfk 1; d : drift, l=2*m; dt : drift, l=3*m, tilt = pi/4; sex: sextupole, l=1 * m, k 2 = 10; sbvu : sbend, l=2*m, angle=pi/7, tilt=pi/2; sbvd : sbend, l=2*m, angle=pi/7, tilt=pi; sbvr : sbend, l=2*m, angle=pi/7, tilt=0; sb 1 : sbend, l=2*m, angle=pi/7; td: transform 3 d, z=0, phi=pi/4; td 1 : transform 3 d, theta=-pi/3; td 2 : transform 3 d, theta=pi/3; vrot : transform 3 d, psi=pi/2; ivrot : transform 3 d, psi=pi; ! method 1 - using coordinate transformations test 1: line=(d, vrot, sb 1, d, ivrot, sb 1, d); Ilya Agapov
Mokka is a My. SQL database of various ILC components ● No standardization – people tend to use random formats. ● BDSIM has a driver to certain formats of My. SQL databases dumped into files (supported by J. Carter) ● CAD-Mokka conversion can be performed, but requires heroes like Lawrence Deacon ● Ilya Agapov
A picture of a cryomodule coded in Mokka for laserwire diagnostics studies L. Deacon RHUL Ilya Agapov
And an example of a Mokka file CREATE TABLE TUBE_CONE ( PARENTNAME VARCHAR(32), # POSX DOUBLE(10, 3), # POSY DOUBLE(10, 3), # POSZ DOUBLE(10, 3), # RED DOUBLE(10, 3), # GREEN DOUBLE(10, 3), # BLUE DOUBLE(10, 3), # VISATT VARCHAR(32), # I = INVISIBLE, S = SOLID, W = WIREFRAME LENGTH DOUBLE(10, 3), # RINNERSTART DOUBLE(10, 3), # RINNEREND DOUBLE(10, 3), # ROUTERSTART DOUBLE(10, 3), # ROUTEREND DOUBLE(10, 3), # ROTPSI DOUBLE(10, 3), # ROTTHETA DOUBLE(10, 3), # ROTPHI DOUBLE(10, 3), # MATERIAL VARCHAR(32), # MATERIAL, CGA LITERAL NAME VARCHAR(32) # NAME OF SOLID, LOGICAL, AND PHYSICAL VOLUME ); Ilya Agapov
GDML driver is provided with Geant 4 and is available in BDSIM ● Existing GDML detector descriptions can be loaded, but this level of detail is not required for collimation/background studies ● Ilya Agapov
Future geometry concept Aim – XML-based language with CAD interface and easy from-MAD conversion (or beam optics code that support a subset of that language) provided ● Medium-term solution – gmad (+ Mokka + GDML) ● A substantional effort is required to achieve an agreement on the standard accelerator description language requirements which should complete about 95% of the task. Some sort of „steering committee“ is probably required. ● Ilya Agapov
RHUL Repository GMAD decks for the beam delivery system and some input bunch files currently under cvs control 1/23/2022 http: //cvs. pp. rhul. ac. uk ● Standardization and version control seem to work in the BDSIM circles but spreading it further obviously requires a higher level of coordination ● Ilya Agapov
Other issues We are producing a BDSIM/Geant/Root/etc. package for Grid. Anyone wanting to join is welcome ● More Communication with physics people desirable for a clearer understanding of collimation/background requirements (my point of view) ● Communication with the same people might help to establish a useful physics validation program ● Ilya Agapov
Extraction Line Power Losses for 2, 14 and 20 mrad designs John Carter • • 2 mrad Losses 14 mrad losses 20 mrad losses Future Plans March 2006 LCWS 06
Simulation Details § Using the following optics versions: § 2 mrad - 9 th Aug 2005 § 14 mrad - 28 th Sep 2005 § 20 mrad - 24 th May 2005 § High statistics Guinea-Pig files by combining: § Main Beam files from http: //www. hep. man. ac. uk/u/robert/files/ § Tail files from http: //www. slac. stanford. edu/~seryi/ILC_new_gp_files/ § (see individual tables for exact stats used) § Studied cases (each with 0 and a given vertical offset): § 0. 5 Te. V CM Nominal § 0. 5 Te. V CM High Lumi. § 1. 0 Te. V CM Nominal § 1. 0 Te. V CM High Lumi. § Tracking and losses simulated in BDSIM § Set to give ’hard losses’ with no shower development § Assume that 100% of energy is deposited John Carter March 2006 - LCWS 06
2 mrad Losses - Charged Beam & Tail Losses in W 0. 5 Te. V Nom. 0. 5 Te. V High Lumi. Total Ext. Power 11. 03 e 6 10. 89 e 6 10. 53 e 6 10. 18 e 6 Vert. Offset [nm] 0 200 0 120 QD 0 0 < 0. 24 106 ± 3. 27 356 ± 35. 7 SD 0 0 < 0. 19 0. 31 ± 0. 05 8. 33 ± 4. 80 SF 1 0 < 0. 47 4154 ± 105 2775 ± 319 ECOLA 0 < 0. 27 264 ± 5. 52 2294 ± 112 HCOL 0 < 0. 22 6. 67 ± 0. 71 0 < 0. 25 VCOL 1. 21 ± 0. 43 4423 ± 468 1157 ± 19. 8 1. 02 M ± 10. 7 k VCOL 2 735 ± 10. 0 614 ± 52. 4 63. 9 k ± 0. 4 k 103 k ± 2. 5 k HCOL 2 813 ± 57. 6 655 ± 100 845 k ± 8. 0 k 545 k ± 17. 7 k HCOL 3 230 ± 6. 35 367 ± 52. 5 10. 0 k ± 0. 2 k 18. 0 k ± 0. 9 k ECOL 0 10. 1 ± 1. 11 6711 ± 535 26. 8 k ± 0. 6 k 56. 2 k ± 2. 5 k ECOL 1 137 ± 15. 1 27. 3 k ± 0. 7 k 11. 7 k ± 0. 4 k 475 k ± 4. 6 k ECOL 2 1. 65 ± 0. 51 19. 4 k ± 0. 6 k 44. 4 ± 3. 25 413 ± 6. 0 k ECOL 3 0 < 41. 5 2563 ± 190 0 < 41. 5 209 k ± 3. 7 k Main Beam 640 k - Tail 2 x 10 k 2 x 25 k 2 x 500 k John Carter March 2006 - LCWS 06
2 mrad Losses - Charged Beam & Tail Losses in W 1 Te. V Nom. 1 Tev Nom. 1 Te. V High Lumi. Total Ext. Power 17. 14 e 6 16. 68 e 6 15. 27 e 6 14. 36 e 6 Vert. Offset [nm] 0 100 0 80 QD 0 0. 86 ± 0. 32 4. 48 ± 0. 94 9897 ± 153 17. 0 k ± 0. 2 k SD 0 0. 09 ± 0. 09 0. 77 ± 0. 39 132 ± 3. 83 1155 ± 30. 7 SF 1 0 < 0. 31 0. 17 ± 0. 17 282 ± 31. 1 1211 ± 40. 6 ECOLA 55. 7 ± 3. 02 265 ± 11. 2 21. 8 k ± 0. 5 k 47. 4 k ± 0. 8 k HCOL 0. 73 ± 0. 33 2. 16 ± 0. 71 1257 ± 17. 7 919 ± 41. 0 VCOL 66. 7 ± 3. 71 62. 8 k ± 2. 5 k 6655 ± 281 963 k ± 20. 6 k VCOL 2 45. 2 k ± 0. 4 k 110 k ± 2. 2 k 740 k ± 7. 5 k 407 k ± 6. 6 k HCOL 2 10. 3 k ± 114 32. 3 k ± 1. 7 k 3. 50 M ± 31. 9 k 5. 27 M ± 45. 2 k HCOL 3 3746 ± 51. 1 6078 ± 259 26. 2 k ± 0. 7 k 20. 9 k ± 0. 7 k ECOL 0 0 < 0. 99 8603 ± 365 62. 8 k ± 0. 1 k 359 k ± 9. 7 k ECOL 1 3683 ± 49. 1 101 k ± 3. 3 k 76. 4 k ± 0. 1 k 581 k ± 7. 9 k ECOL 2 0 < 58. 1 125 k ± 2. 5 k 0 < 58. 1 201 k ± 3. 0 k ECOL 3 0 < 64. 5 59. 6 k ± 1. 3 k Main Beam 630 k 620 k 635 k 625 k Tail 2 x 200 k 2 x 300 k 2 x 175 k 2 x 500 k John Carter March 2006 - LCWS 06
20 mrad Losses - Charged Beam & Tail 0. 5 Te. V Nom. 0. 5 Te. V High Lumi. 11. 03 e 6 10. 89 e 6 10. 53 e 6 10. 18 e 6 Vert. Offset [nm] 0 200 0 120 ECOL 1 1. 71 ± 0. 60 14. 3 ± 1. 44 47. 9 k ± 0. 8 k 384 k ± 4. 8 k ECOL 2 28. 7 ± 2. 26 71. 6 ± 3. 54 74. 5 k ± 1. 2 k 319 k ± 2. 7 k SC Quads 0 < 0. 14 14. 5 ± 0. 85 27. 5 ± 6. 66 Warm Quads 0 < 0. 11 51. 2 ± 1. 39 226 ± 17. 5 Bends 0 < 0. 28 0. 10 ± 30. 6 ± 1. 28 886 ± 48. 3 Max E-Loss Density [W/m] Power Loss [W] Total Ext. Power 0. 10 Main Beam 640 k 625 k - Tail 2 x 10 k 2 x 25 k 2 x 450 k 2 x 500 k John Carter March 2006 - LCWS 06
20 mrad Losses - Charged Beam & Tail 1. 0 Te. V Nom. 1. 0 Te. V High Lumi. 17. 14 e 6 16. 68 e 6 15. 27 e 6 14. 36 e 6 Vert. Offset [nm] 0 100 0 80 Power Loss [W] Total Ext. Power ECOL 1 606 ± 13. 3 4847 ± 57. 7 63. 8 k ± 1. 1 k 136 k ± 1. 0 k ECOL 2 126 ± 5. 34 551 ± 15. 2 41. 3 k ± 1. 3 k 54. 2 k ± 1. 2 k 0 < 0. 20 0 < 0. 32 1128 ± 14. 9 1248 ± 32. 7 Warm Quads 1. 14 ± 0. 25 6. 34 ± 0. 79 4046 ± 70. 5 5227 ± 59. 4 Bends 3. 58 ± 0. 57 207 ± 28. 4 2256 ± 86. 6 8812 ± 0. 2 k Max E-Loss Density [W/m] SC Quads Main Beam 630 k 620 k 635 k 625 k Tail 2 x 200 k 2 x 300 k 2 x 175 k 2 x 500 k John Carter March 2006 - LCWS 06
14 mrad Losses - Charged Beam & Tail 0. 5 Te. V Nom. 0. 5 Te. V High Lumi. 11. 03 e 6 10. 89 e 6 10. 53 e 6 10. 18 e 6 Vert. Offset [nm] 0 200 0 120 Max E-Loss Density [W/m] Power Loss [W] Total Ext. Power ECOL 1 1. 14 ± 0. 62 2946 ± 45. 5 38. 0 k ± 0. 6 k 1. 01 M ± 10. 6 k ECOL 2 1007 ± 77. 5 27. 1 k ± 0. 7 k 233 k ± 3. 5 k 856 k ± 8. 0 k ECOL 3 527 ± 80. 4 1863 ± 150 41. 6 k ± 1. 0 k 46. 0 k ± 0. 8 k SC Quads 0 < 0. 11 0. 92 ± 0. 20 1. 32 ± 1. 32 Warm Quads 0 < 0. 13 28. 5 ± 1. 19 156 ± 16. 0 Bends 0 < 0. 27 0. 05 ± 27. 8 ± 1. 18 589 ± 36. 7 0. 05 Main Beam 640 k 625 k - Tail 2 x 10 k 2 x 25 k 2 x 450 k 2 x 500 k John Carter March 2006 - LCWS 06
14 mrad Losses - Charged Beam & Tail 1 Te. V Nom. 1 Te. V High Lumi. 17. 14 e 6 16. 68 e 6 15. 27 e 6 14. 36 e 6 Vert. Offset [nm] 0 100 0 80 Power Loss [W] Total Ext. Power ECOL 1 223 ± 7. 28 1011 ± 22. 6 28. 1 k ± 0. 7 k 574 k ± 6. 6 k ECOL 2 231 ± 7. 27 15. 2 k ± 0. 2 k 369 k ± 5. 9 k 1. 27 M ± 11. 2 k ECOL 3 16. 3 ± 1. 96 1780 ± 32. 4 62. 2 k ± 1. 3 k 52. 6 k ± 1. 1 k 0 < 0. 17 0 < 0. 28 235 ± 4. 27 236 ± 10. 7 Warm Quads 0. 66 ± 0. 20 5. 36 ± 0. 81 2865 ± 33. 4 3773 ± 68. 6 Bends 2. 74 ± 0. 45 52. 4 ± 3. 40 2596 ± 75. 8 8084 ± 194 Max E-Loss Density [W/m] SC Quads Main Beam 630 k 620 k 635 k 625 k Tail 2 x 200 k 2 x 300 k 2 x 175 k 2 x 500 k John Carter March 2006 - LCWS 06
Conclusions & Future Plans § The 2, 14, and 20 mrad extraction lines have undergone a preliminary study to evaluate the hard power losses § The results show that more optimisation of the optics may be required for the extraction lines to be able to handle all the parameter sets. § Full shower simulations currently being run on the RHUL Grid Cluster with aims to make some tracking optimisations to speed up CPU time. § Only the charge beam and tail have been considered for this study - more losses on early elements should be expected from Radiative Bhabha’s (see work done for 2 mrad QD 0 in a separate 2 mrad talk given by Rob Appleby) § BDSIM has started to be used to give extremely detailed evaluations of localised energy deposition (m. W/g). John Carter March 2006 - LCWS 06
SUMMARY BDSIM has developed over several years ● Standardised input and output formats ● Matches to accelerator decks and ILC detector descriptions ● Producing “next generation” simulation results ● Grid implementation underway ● Further benchmarking planned ● Used at ATF and PETRA to simulation laser-wire ● Code available and a manual (EUROTe. V report). ● Grahame Blair March 2006 - LCWS 06
Extra Slides Power Loss into 2 mrad QDO (Also being shown by Rob Appleby in a separate talk) John Carter March 2006 - LCWS 06
2 mrad Losses - Scoring of QD 0 § Ring 1: 200 segments 3. 4 < ~R < 3. 5 cm Material: Aluminium (QD 0 Scored into 300000 volumes) § Ring 2: 200 segments 3. 5 < ~R < 4 cm Material: Nb. Ti dz = 1 cm X § Ring 3: 200 segments 4 < ~R < 5 cm Material: Nb. Ti § Ring 4: 200 segments 5 < ~R < 8 cm Material: Nb. Ti § Ring 5: 200 segments 8 < ~R < 13 cm Material: Nb. Ti § Ring 6: 200 segments 13 < ~R < 20 cm John Carter Material: Nb. Ti March 2006 - LCWS 06
2 mrad Extraction Line Losses in SC QD 0 from Radiative Bhabhas § Radiative Bhabha files from http: //www. slac. stanford. edu/~tvm/pairs/ § Assumed Coil is 100% Nb. Ti with density 5. 6 g/cm 3 @ 4 Kelvin from R=35 mm to 200 mm (i. e. no support structure or gaps between 4 coils, etc. accounted for). § Using Aluminium beampipe with 1 mm thickness, density = 2. 7 g/cm 3 § Using Tungsten liner with 3 mm thickness (~1 radiation length), density = 19. 3 g/cm 3 (as suggested by L. Keller & T. Maruyama 26/7 SLAC BDS Meeting) § Total energy deposits recorded per segment� with showers tracked down to 10 Ke. V (charged and neutral) Total Incident Total Power Peak Power Extracted Power [W] 0. 5 Te. V Nom 0. 5 Te. V High Lumi 1. 0 Te. V Nom 1. 0 Te. V High Lumi John Carter No Liner Tungsten Liner 70. 498 161. 37 179. 12 483. 03 Power [W] deposited in beampipe & Coils R<20 cm [W] Density in Beampipe [m. W/g] Density in Nb. Ti Coils [m. W/g] 0. 45 ± 0. 01 0. 27 ± 0. 01 0. 71 ± 0. 03 1. 90 ± 0. 07 0. 61 ± 0. 02 0. 13 ± 0. 004 0. 26 ± 0. 01 0. 11 ± 0. 01 1. 07 ± 0. 09 0. 60 ± 0. 05 1. 60 ± 0. 15 4. 27 ± 0. 41 1. 44 ± 0. 12 0. 32 ± 0. 03 0. 61 ± 0. 05 0. 27 ± 0. 02 1. 12 ± 0. 07 0. 65 ± 0. 04 1. 35 ± 0. 08 4. 11 ± 0. 23 1. 49 ± 0. 09 0. 34 ± 0. 02 0. 67 ± 0. 04 0. 30 ± 0. 02 3. 61 ± 0. 31 2. 08 ± 0. 18 4. 41 ± 0. 38 13. 7 ± 1. 2 4. 95 ± 0. 43 0. 69 ± 0. 07 1. 40 ± 0. 15 0. 62 ± 0. 07 March 2006 - LCWS 06
2 mrad Extraction Line QD 0 Power Density Maps § Power density maps for the first 5 rings (6 th ring had no deposits) § 0. 5 Te. V Nominal Case § No Tungsten Liner § All density units in W/g John Carter March 2006 - LCWS 06
- Slides: 31