Status of BESIII Event Reconstruction System BESIII offline
Status of BESIII Event Reconstruction System BESIII offline software group 2006. 10. 31
BESIII Experiment Main Drift Chamber (MDC): xy = 130 m P/P = 0. 5 %@1 Ge. V d. E/dx = 6 -7 % TOF System: T = 90 ps barrel 110 ps endcap Super-conducting Magnet: 1. 0 Tesla Muon Chamber (MUC): RPC based EM Calorimeter (EMC): E/E = 2. 5 % @ 1 Ge. V z, = 0. 6 cm @ 1 Ge. V
Offline software BOSS – BES Offline Software System based on Gaudi.
Calibration Recon. and simulation algorithms retrieve calibration data from calibration data store v Calibration constants are stored in ROOT files v GUI Client Reconstruction algorithm Calib. Func Svc Calibration Data Service Calib. Root Cnv. Svc My. SQL database Simulation algorithm Calibration constants (ROOT) Calibration Algorithm
The BESIII Event Reconstruction Modules and General Flow Event. Start. Time (Fast-tracking, PID, EST ) MDC Tracking d. E/dx Rec. (Particle ID) Track Fitting (Kalman Filter) Track Ext. TOF Rec. Veefind Event. Start. Time Track Matching EMC Rec. Mu. C Rec. Evtvert
MDC Fast Tracking Module Track parameters: d 0 , Ф 0, Κ, dz , tgλ
TEST (Event Start Time) Module The TEST is important for the momentum Res. , spatial Res. of the charged particles, but it is uncertain in the online system due to: 1) The BESIII trigger system can’t separate every event up to beam bunch. 2) The start time of charged event is determined by TOF_T, but the TOF reach time is different for different particle and different momentum. TEST should be calculated by offline system. REF (2 ns) Bunch (8 ns) Trigger Clock (24 ns) TOF Test tev TDCM(TOF) MDC TDCM(MDC)
TEST (Event Start Time) Flow MDC Fast Reconstruction TOF 1. Calculate Test by 2. straight line segments fit Yes No. TOF No Have TOF Information ? No 2. Calculate Test by TOF Inf. And MDC Fast tracking Inf. PID MDC Fast Rec. Information? Yes 4. Calculate Test 3. Calculate T by MDC Inf. est Only by MDC Inf. And Fast tracking Inf.
Preliminary Result of TEST TOF: 96. 78% MDC: 3. 22% Segment Fit Test Eff. : 100% Muon , Test=0 ns, 8 ns, 16 ns, 5000 events TOF: 97. 2% MDC: 2. 6% Segment Fit: 0. 2% Test Eff. : 96. 4% error: 0. 63% e+e- hadrons , Test=0 ns, 8 ns, 16 ns, 5000 events
Mdc Tracking Algorithm Two algorithms for MDC track reconstruction: Trk. Reco & Mdc. Pat. Rec Trk. Reco Mdc. Pat. Rec Personpower Zang S. L. (IHEP) Zhang X. Y. (SDU) Zhang Y. (SDU) Based on Belle reco. algorithm Ba. Bar reco. algorithm Segment finding Conformal transform Pattern matching Track fitting Least square method
MDC Tracking Module(1) (based on Belle Lib. ) ØInit: Geometry-Survice. , Cal. -Const. , Adjust-constant Get Hit Information… ØR- Tracking: segment finding by conformal transformation and histogram method Conformal transformation ØZ Finding Ø 3 D Helix Fit S-Z calculation in Z finding
MDC Tracking Module(2) (based on Babar Lib. ) ØInit: Geometry-Survice. , Cal. -Const. , Adjust-constant Get Hit Information… ØSegment finding: Search segments in each super-layer using a pattern look-up table ØTracking: Link segments to 2 D tracks, add stereo segments Ø 3 D Helix Fit Sequence of segment Finder For every superlayer, every wire in 2 nd Layer form a group of 8 wires For every group (1) Try 4 hit pattern for group (2) Try 3 hit pattern for group Set massage of segment list 7 6 4 5 3 2 1 0 0 1 0 1 7 6 5 4 3 2 1 0 Wires No. 0 -7 Set one word for a group of 8 wire, each bit for a wire. Set “ 1” for a hit wire, others “ 0” This octal value used for its group No.
Preliminary Result of MDC tracking Momentum resolution μ- at pt = 1 Ge. V Momentum resolution σp = 0. 40% Spatial resolution μ- at Pt = 1 Ge. V spatial resolution σ ~= 110μm
Preliminary Result of MDC tracking Efficiency vs Pt (e-, μ-, π, p) 0. 1 mm d 0: signed distance from the pivot to track in x-y plane • Efficiency > 95% for single track Pt > 200 Me. V z 0: signed distance from the pivot to track in z direction 0. 8 mm
Kalman filter tracking fitting Module The MDC reconstruction results are not very good, have to do Kal. Fit: Ømultiple scattering Øenergy losses Ønon-uniform magnetic fields …… Parameters of a track before Kal. Fit & after Kal. Fit(proton 300 Mev)
d. E/dx Module v Principle: MDC tracking d. E/dx~f(v ) Particle type info P =βγ · m Ø d. E/dx codes developed successfully, released for physics study. Ø Particle ID is tested with MC samples, d. E/dx resolution, distributions, PID efficiency is reasonable.
Pre. Performance of d. E/dx seperation for 5 particles(MC) seperation power with d. E/dx üGood particle seperation in a wide range for different particles üThe π/K seperation(3 σ ) reach about 800 Me. V/c üParticle identification efficiency is more than 90% with MC samples
Track Extrapolation Module Function: Extrapolate MDC tracks to Outer detectors: TOF, EMC and MUC Algorithm: based on GEANT 4, Energy loss and Magnetic field. Multiple scattering effect put into error matrix.
TOF Reconstruction Module Initialize() Tof. Geom. Svc Tof. Calib. Svc Geometry data Calibration constants Data base Get Data Tof Digits MDC Tracks TDS Get Tracks execute() Make Hits Match Hits and Tracks Reconstructed Tof Tracks finalize() Tof Tracks TDS
Primary TOF Recon. Result v Reconstruction Efficiency : 99. 82% (1 Ge. V electron data) beta vs momentum time resolution: 113 ps
EMC Reconstruction Module EMC is used to measure energy and position of electrons and photons, and to provide neutral energy trigger digit hit cluster shower
Pre. Performances of EMC single gamma with 0. 5 Me. V noise σE/E≈2. 2%@1 Ge. V theta phi Recon Effi. of π0
Muc Reconstruction Module Use Ext Track from Mdc as seed; Window Searching hits gap by gap; Searching on Barrel first, then End. Cap; Also Searching on neighboring segments; Ext track Line Fit with hits on track; Compare with Mc truth; Barrel End Cap Fired strips
Pre. Performances of Muc Ratio % μ/π Identifcation With different θ μ efficiency 90 70 50 40 30 π(fake μ ) p (Ge. V/c)
Summary v v Based on Gaudi, the BESIII offline software system (Boss) has developed Simulation Reconstruction Calibration system Physics analysis All sub-system work well on the Boss 6. 1. 0, got the Pre. Performance by the simple data sample, and passed the checking by typical physics channel. More works need to do to make each code work in the best status by closely the data to real data case step by step, such as: “noise study”, NUMF and “low momentum tracking” and S-T relation in MDC etc. Data challenge and software performance optimization are the major task in the next step.
- Slides: 26