Andrew Puckett SBS Weekly Meeting June 3 2015 Outline Recent geometry additions to g4sbs Recent code improvements Highlights of recent g4sbs simulation resultsanalysis Ongoing projectsstatus ID: 799306
Download The PPT/PDF document "SBS Monte Carlo Simulation Update" is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
SBS Monte Carlo Simulation Update
Andrew Puckett
SBS Weekly Meeting
June 3, 2015
Slide2Outline
Recent geometry additions to g4sbs
Recent code improvements
Highlights of recent g4sbs simulation results/analysisOngoing projects/status:Missing/incomplete geometriesSimulation capabilities wish-listSimulation studies wish-listStatus of user documentationStatus of codeOutlook
6/3/2015
SBS Weekly Meeting June 3, 2015
2
Slide3Recent geometry additions:
HCAL (Freddy Obrecht, code from Vahe Mamyan)
BigBite shower/preshower calorimeters (Freddy Obrecht, geometry from S. Riordan thesis)
ECAL (Freddy Obrecht, transparency model from S. Abrahamyan & Bogdan W.)RICH geometry finalized (“full” details from CAD drawings, A. Puckett)6/3/2015
SBS Weekly Meeting June 3, 2015
3
Slide4GRINCH
Implementation
6/3/2015
SBS Weekly Meeting June 3, 20154
GEANT4 model developed by H. Yao (W&M) imported to g4sbs framework
To-do:
update PMT quantum efficiency (currently based on RICH PMTs)
Detailed description of containment volume
Understand and mitigate sources of background
Slide5RICH
update
6/3/2015
SBS Weekly Meeting June 3, 20155
Virtually complete description of RICH aluminum box and steel shielding for PMTs from CAD drawings
New acceptance, PID performance and background studies under way
Slide6HCAL Implementation
HCAL GEANT4 code from V. Mamyan imported to g4sbs by R. F. Obrecht (code version up-to-date as of ~Oct. 2014, doesn’t include more recent developments by CMU)
6/3/2015
SBS Weekly Meeting June 3, 2015
6
Slide7BigBite Calorimeter implementation in g4sbs
BigBite preshower and shower calorimeters implemented by R. F. Obrecht
Note—optical properties taken from GEP ECAL simulation; need to update for different type of lead-glass; should be reasonable approximation already
6/3/2015
SBS Weekly Meeting June 3, 2015
7
Slide8Highlights of g4sbs simulation results/analysis
6/3/2015
SBS Weekly Meeting June 3, 2015
8
Slide9Checkout of HCAL in g4sbs—coordinate resolution
Preliminary HCAL coordinate resolution in g4sbs ~2.8 cm (based on cluster center of gravity calculation with linear energy weighting)
consistent with CDR
6/3/2015
SBS Weekly Meeting June 3, 2015
9
Slide10Checkout of HCAL in g4sbs—Energy resolution
Preliminary
σ
E
/E ~ 23% @7 GeV
~ 60%/sqrt(E) (consistent with CDR number)
6/3/2015
SBS Weekly Meeting June 3, 2015
10
Slide11Notes on preliminary HCAL performance study
Code is the same as that used by CMU group; studies done to validate implementation in g4sbs program—consistency with previous results suggests implementation is correct.
Studies made extensive use of new code features in uconn_dev branch:
“detector maps” which provide cell coordinate info needed for clustering of hitsFT/FPP1/FPP2 track information utilized6/3/2015
SBS Weekly Meeting June 3, 2015
11
Slide12FPP simulated angular distributions (unpolarized)
Scattering angle distribution for all tracks in FPP1 relative to elastically scattered proton track in FT
Qualitatively similar to real data from e.g., Hall C FPP
6/3/2015
SBS Weekly Meeting June 3, 2015
12
Slide13FPP simulated distance of closest approach
FPP1
FPP2
6/3/2015
SBS Weekly Meeting June 3, 2015
13
Slide14FPP simulated “zclose” vs. theta
Left: theta vs. zclose for FPP1 relative to FT
Right: theta vs. zclose for FPP2 relative to FPP1
6/3/2015
SBS Weekly Meeting June 3, 2015
14
Slide15Notes on preliminary FPP studies
Unpolarized angular distributions look qualitatively similar to real data from e.g., Hall C
So far, track reconstruction “cheats” using MC “truth” information to group hits into tracks
Cuts applied on track momentum, chi2, etc. No polarization info included yet.Knowledge of MC “truth” info enables detailed study of FPP efficiency and particle energy distributions, particle IDs and multiplicities as a function of scattering angle—could inform polarization analysis
6/3/2015
SBS Weekly Meeting June 3, 2015
15
Slide16ECAL MC studies (R. F. Obrecht)
Presented at APS April 2015 Meeting in Baltimore
6/3/2015
SBS Weekly Meeting June 3, 2015
16
Slide176/3/2015
SBS Weekly Meeting June 3, 2015
17
Slide186/3/2015
SBS Weekly Meeting June 3, 2015
18
Slide196/3/2015
SBS Weekly Meeting June 3, 2015
19
Slide206/3/2015
SBS Weekly Meeting June 3, 2015
20
Slide216/3/2015
SBS Weekly Meeting June 3, 2015
21
Slide22Recent code improvements
6/3/2015
SBS Weekly Meeting June 3, 2015
22
Slide23Segregation of g4sbs commands into pre/post-initialization
Added optional segregation of configuration commands into “pre-init” and “post-init” execution:
Motivation: g4sbs user needs the option to turn off computationally expensive optical photon production without recompiling the code.
Need to work around (apparently long-standing) GEANT4 bug that crashes g4sbs when attempting to disable scintillation using “/process/inactivate Scintillation”Solution: added commands “/g4sbs/useckov” and “/g4sbs/usescint” to toggle Cerenkov/Scintillation processes on/off in G4SBSPhysicsList
Commands only have an effect if invoked in the pre-initialization phaseHow it works: program interprets
“./g4sbs file1.mac file2.mac” to mean:Execute the commands in file1.mac before
RunManager::Initialize()
Execute the commands in file2.mac
after
RunManager::Initialize()
If only one macro file is given on the command line, will execute
after
initialization (this was default behavior previously)
For most existing g4sbs commands, GEANT4 doesn’t care whether they are invoked pre-init or post-init. Known exceptions:
/g4sbs/useckov and /g4sbs/usescint must be pre-init
/g4sbs/run must be invoked post-initialization.
Elimination of existing GEANT4 bug
could
eliminate the need for this feature, but:
Potentially useful for other purposes such as user configuration of geometry/materials and other parameters defined pre-initialization.
6/3/2015
SBS Weekly Meeting June 3, 2015
23
Slide24Encapsulation of calorimeters for optional “fast” simulation
Enclose all heavy showering detectors (ECAL/HCAL/BBCAL) in mother volumes that can be optionally made “total absorbers” with “calorimeter sensitivity” using “/g4sbs/steplimit” command
When enabled, all particles entering the mother volume enclosing a calorimeter are killed and recorded without simulating any details of the shower development.
Speeds up simulation when calorimeter info not wanted for e.g., rate estimates and acceptance calculations6/3/2015
SBS Weekly Meeting June 3, 2015
24
Slide25Ongoing Projects/Status
HCAL performance checkout (UConn):
nearly done, update code to latest version from CMU
ECAL (UConn):Done, but need final geometryCDet (SMU)GRINCH backgrounds (JLab, Dasuni A.)FPP efficiency and polarimetry (UConn) RICH performance (UConn)Documentation and user support (UConn)
6/3/2015
SBS Weekly Meeting June 3, 2015
25
Slide26Simulation capabilities wish-list:
Particle spin-tracking (relevant mainly to GEP and new GEN recoil proposal)
“Smart” retention of more limited subset of “particle history” info for heavy showering detectors—current approach is “all” (complete history) or “nothing” (hit info only; e.g. nphe, total edep, etc)
Fast, parametrized detector response for optical photon/shower detectorsFast, parametrized “full digitization” to generate ADC/TDC pseudo-dataMore detailed GEM response simulation including strips, clustering, etc. (need to parametrize response from real test data, not planning to simulate ionization/amplification process)
Interface to “minimum-bias” event generators such as PYTHIA and others
Investigating use of ROOT-PYTHIA6 interface (“TPythia6”)
6/3/2015
SBS Weekly Meeting June 3, 2015
26
Slide27Simulation studies wish-list
SBS spin precession with GEANT4 spin-tracking and realistic
magnetic field map—ray-tracing AND fitting
SBS and BigBite optics and resolution (already done, but needs to be re-done for all experiments with realistic field maps)Optimize layout of the SIDIS experimentUpdate background rates for all detectors in all experimentsDevelop reconstruction for all detectors with realistic background conditions
6/3/2015
SBS Weekly Meeting June 3, 2015
27
Slide28Status of User Documentation
Draft documentation of new tree structure in “uconn_dev” written by R. F. Obrecht
End-user and API documentation coming soon
Encourage use of “uconn_dev” branch; contact Andrew Puckett/Freddy Obrecht with specific questions“Standard” scripts for all experiment configurations coming soon. Documentation of g4sbs commands coming soon6/3/2015
SBS Weekly Meeting June 3, 2015
28
Slide29Status of Code
Major changes represented by “uconn_dev” branch largely tested, debugged.
Need to update default/standard simulation scripts for all experiment configurations
Different tree structure depending on detectors in useneed standard “TTree::MakeClass()” files (skeleton ROOT analysis macros) for each experiment configuration. Near future—merge stable version of “uconn_dev” into “master” branch
6/3/2015
SBS Weekly Meeting June 3, 2015
29
Slide30Outlook
Major improvements to g4sbs largely tested/debugged
New capabilities already bearing fruit in analysis of simulation results
Current development priorities:Clean-up of simulation scripts and ROOT macros to make compatible with new tree structure and code: need to work “out-of-the-box”“Official release” of new g4sbs version, merge of “uconn_dev” into master branch of g4sbs, integration of parallel effortsUser and developer documentation:
How to run simulationsHow to analyze data—documentation of tree structure
How to add new detectors/geometriesOpportunities for collaboration—many geometries still missing/incomplete
6/3/2015
SBS Weekly Meeting June 3, 2015
30
Slide31SIDIS Cherenkov Background Rates
6/3/2015
SBS Weekly Meeting June 3, 2015
31
PRELIMINARY
RICH (left) and GRINCH (right) counting rates for SIDIS configuration
Caveats: results obtained with “bare-bones” He-3 target in air—target is only ~1/3 of material along beamline
More effective shielding and collimation of target can reduce rates
Investigation of background sources for Cherenkov counters ongoing using (new) detailed particle histories
RICH: 700 kHz/PMT average (0.7% occupancy for 10 ns window)
GRINCH: 3.8 MHz/PMT average (3.8% occupancy for 10 ns window)
Slide32SIDIS GEM Backgrounds
6/3/2015
SBS Weekly Meeting June 3, 2015
32
SBS GEM tracker for SIDIS average hit rate at first plane ~ 35 kHz/cm
2
(10X below GEP case)
Slide33ECAL position resolution from GEANT4
6/3/2015
SBS Weekly Meeting June 3, 2015
33