BESIII Simulation Software

1
BESIII Simulation Software

• Deng Ziyan, IHEP
• ACAT 2007, Amsterdam
• April 23-27

2
Outline

• BESIII experiment
• BESIII offline software
• Simulation Algorithm
• Performance
• Summary

3
BEPCII and BESIII

• BEPC (Beijing Electron-Positron Collider)
• BESI/BESII detector worked on it from 1988 to
  2005
• Beam energy 1.5 - 2.8GeV
• BEPC ? BEPCII
• Luminosity
• 1.01031cm-2s-1 ? 1.01033cm-2s-1
• Number of beam bunches
• 1? 93
• BESIII a new spectrometer to be working on
  BEPCII
• Very good energy and angle resolution for photon
  measurement
• Accurate 4-momenta measurement of charged
  particles with low momentum
• Good hadron identification capabilities

?
?
?
4
BESIII Detector

• MDC
• Small cell, He gas based main drift chamber
• TOF
• Plastic scintillator based PID system
• EMC
• CsI(Tl) crystal electromagnetic calorimeter
• SSM
• Superconducting solenoid magnet of 1T
• MUC
• RPC based muon counter

Detector moving into the collision region by Jan.
2008
5
BESIII Offline Software System BOSS
Calibration Database
FACS
Reconstruction
Simulation
Generator
Transient Data Store
Simulation
Digitization
6
Simulation Algorithm

• Components
• Geometry description
• Event Generator
• Physics processes
• Magnetic field
• Digitization
• MC truth
• Data I/O
• Trigger simulation
• Background mixing

• Developed in BOSS framework
• Based on Geant4 toolkit
• External packages
• Geant4.8.1p01
• GCC 3.2.3
• CLHEP 1.9.2.3
• GDML 2.8.0
• Xerces_C 2.7.0
• ROOT 4.03.02

Project Leader Liu Huaimin Cao Guofu, Deng
Ziyan, He Miao, Liang Yutie, Pan Minghua, You
Zhengyun, Yuan Ye
7
Detector Description
xy (z0)
yz (x0)
8
Detector Description
stereo cell G4TwistedTub
MDC
TOF
G4IrregBox
EMC
MUC
9
Class for Special Geometry

• G4TwistedTub
• Lately imported to Geant4 source code
• Error in step calculation at special points
  inside MDC stereo cells, step out of world volume
• already reported to the author
• G4IrregBox
• Not in current Geant4 release
• Useful in many experiments, such kind of geometry
  is hoped to be included in later Geant4 and GDML
  release

10
Detector Description with GDML

• GDML (Geometry Description Markup Language),
  originally developed by Geant4 group
• Peking University group expanded the GEANT4
  Schema and developed a new ROOT Schema for BESIII
  applications.
• GDML Detector Description can be used for
  simulation, track extrapolation, event display
  and reconstruction.

11
Event Generator

• BES has 30 event generators written in Fortran
• Use C HepEvt_Wrapper to access the kinematics
  infor generated by the generators
• Other generators
• EvtGen, Bhlumi, KKMC
• G4HepMCInterface being used
• Why removed in latest Geant4 release?

12
Physics Processes

• General decay, transportation
• Electromagnetic
• standard (used), low energy (tried), optical
  photon (not used)
• Hadronic
• hadron in flight elastic, inelastic
• hadron at rest absorption, capture
• Different hadronic models used and compared
• LHEP LEP/HEP parameterized model
• QGSP Quark-Gluon String Precompound, theory
  driven model

13
Physics Processes

• Bugs in hadronic processes
• Classes list (Geant4.8.1p01)
• G4LEAntiProtonInelastic
• G4HadronElastic
• G4UHadronElasticProcess
• G4MuonMinusCaptureAtRest
• G4NucleiProperties
• G4InelasticInteraction
• Cause dead-loop or crash, hard to detect
• Modification or protection made in our Geant4
  package
• Reported to Geant4 team, also confirmed by other
  groups and users

14
Hadronic Models (QGSP)
Old Geant4.7.1
New Geant4.8.1p01
Energy(GeV)
Energy(GeV)
no obvious peak fake photons increase10
a peak around 70MeV
Fake photon energy in EMC p-
(momentum100500MeV)
15
Magnetic Field
From SMG

• Uniform magnetic field being used
• Simulation and reconstruction can use the same 3D
  magnetic field.
• Linear interpolated algorithm has been checked
  carefully and works well

Y0 plane
From BOSS
16
Digitization

• MDC
• Drift distance-gt drift time
• dE/dx
• Wire resolution added
• noise added
• TOF
• Light emission light propagation
• PMT response threshold discrimination

• EMC
• Waveform added
• Simulate main amplifier output
• Provide risetime to eliminate noise
• MUC
• Strip ID
• Noise added

17
MC truth

• Purpose
• To provide truth information for reconstruction
  and physics analysis
• What to save
• True state of the particles and their association
  with detector response
• Particle
• particles from generator
• uninterrupted decayed daughters
• Detector response
• hits in sub-detectors associated with particles

18
MC truth

• Track and vertex
• (associated with particle)
• Track
• PDGcode
• charge
• original vertex
• terminal vertex
• four-momentum
• decayed daughters
• Vertex
• parent track
• position
• time

• truth hits
• (associated with track)
• Considering demand of each reconstruction
  developer
• MDC
• Hit history
• TOF
• First-hit
• EMC
• Total energy loss
• MUC
• Hit history

19
Data I/O

• Output MCtruth and raw data in two formats
• ASCII format (simple, convenient, but big size)
• Binary format (just as online, small size, for
  large sample)
• Easy for user to choose data format
• Easy to turn on/off

EVHEAD TRACKTRUTH, VERTEXTRUTH MDCTRUTH,
MDCDIGI TOFTRUTH, TOFDIGI EMCTRUTH,
EMCDIGI MUCTRUTH, MUCDIGI
20
Trigger simulation
Trigger Table
jobOptions
Event Filter
Output Files
Trigger Algorithm
Trigger Service
Mdc Trigger Conditions (Total 9 conditions)
Mdc Digi
Mdc Track Segment Finder
Track Finder
Global trigger
Isolated Cluster
Emc Trigger Conditions (Total 16 conditions)
Emc Digi
Emc Trigger Cell
Energy Sum and Balance
Tof Trigger Conditions (Total 6 conditions)
Tof Digi
Tof Hit Count

• Trigger simulation with three sub-detectors
  almost completed
• Trigger with Muc is ongoing
• Cross check has been made with trigger group

21
Background Mixing

• Event mixing (developed by Dubna group)
• Signal Background ? Real event
• Algorithm implemented
• Only signals and noise with the same identifier
  are added up
• If no corresponding MC digit found, new digit is
  created with background information
• New charge is sum of MC charge and background
  charge
• New time measurement is the earliest of two
• These rules can be changed individually for each
  sub-detector, if requested

22
Background Mixing

• Simulated background
• Trigger group did some study on this
• MC background events are used for mixing
• Background data taking
• There were some discussions
• Single beam, or separated beams without collision
• Data taking with random trigger

23
Performance speed

• Average 0.9 second per physics event
• About 1.5 times slower than BESII simulation (G3)

Responsible J. Apostolakis / G. Cosmo
Description "Compared to G3 simulation, under
similar circumstances G4 is reported by the LHC
experiments, to be a factor 1.5-2 slower. A study
group started last year to address this issue,
and should continue with more priority. This is
expected to be a collaboration between G4 and the
users."
Geant4 Technical Forum on November 20th, 2006
24
Performance MDC
µ- at pt 1GeV/c
spatial resolution 110µm
momentum resolution 0.4
25
Performance TOF
time resolution at z0 85ps (1GeV electron)
beta vs momentum (e, µ,p, K, p)
26
Performance EMC
energy resolution of gamma 2.1_at_1GeV
spatial resolution of gamma
5.8mm_at_1GeV
27
Performance
Sub-detector Sub-detector Designed MC
MDC spatial 130um 130?m(smeared as designed)
MDC spatial 130um 110?m(smeared based on beam test)
MDC momentum 0.46_at_1GeV 0.46(?xy 130?m)
MDC momentum 0.46_at_1GeV 0.41(?xy 110?m)
MDC dE/dx 6-7 6
TOF time 80-90ps 85ps
EMC energy 2.5_at_1GeV 2.1
EMC spatial 6mm_at_1GeV 5.8mm
MUC µ ID 95 96
MUC p miss_ID 6 6.2
28
Summary

• Simulation software running stably
• Key quantities consistent with design target
• Mass production starts for physics study
• More studies on physics processes and
  digitization needed