physics validation of the simulation packages in a lhc-wide effort
DESCRIPTION
CHEP’04 , Interlaken, 27 th September 2004. Physics Validation of the Simulation Packages in a LHC-wide effort. Alberto Ribon CERN PH/SFT on behalf of the LCG Simulation Physics Validation group. Track 2 “Event Processing” #493. Geant4 Project. FLUKA Project. - PowerPoint PPT PresentationTRANSCRIPT
Physics Validation of the Physics Validation of the Simulation Packages in a Simulation Packages in a LHC-wide effortLHC-wide effort
Alberto RibonAlberto Ribon
CERN PH/SFT CERN PH/SFT on behalf of the LCG Simulation Physics Validation group
CHEP’04, Interlaken, 27th September 2004
Track 2 “Event Processing” #493
Alberto Ribon, CERN/PH/SFT
2
Simulation Project LeaderG.Cosmo
FrameworkA. Dell’Acqua
WPWP
WP
Geant4J.Apostolakis
WPWP
WP
FLUKAIntegration
A.Ferrari
WPWP
PhysicsValidationF.Gianotti
WPWP
WP
ShowerParam
WPWP
GeneratorServicesP.Bartalini
WPWP
Subprojects
Work packages
Geant4Project
FLUKAProject
ExperimentValidation MC4LHC
Alberto Ribon, CERN/PH/SFT
3
LCG physics validation project goals: • compare Geant4, Fluka with the LHC test-beam data
• test coherence of results across experiments and sub-detector technologies
• study simple benchmarks relevant to LHC
• “certify” that simulation packages and framework are ok for LHC physics
• weaknesses and strengths of the packages
More details:
http://lcgapp.cern.ch/project/simu/validation/
Alberto Ribon, CERN/PH/SFT
4
•
• First cycle of electromagnetic physics validation completed at the percent level. We will focus here only on the (most difficult!) hadronic physics validation.
• As for the choice of the Geant4 Physics List, the validation should be targeted to each considered application domain: e.g. for high-energy physics one should consider different observables than, for instance, medical physics, or space science.
• The criteria to consider a simulation “good” or “bad” should be based on the particular application: e.g., for LHC experiments, the main requirement is that the dominant systematic uncertainties for all physics analyses should not be due to the imperfect simulation.
Physics Physics ValidationValidation
Alberto Ribon, CERN/PH/SFT
5
Validation project
Suppose that e.g. for e/ : (G4-test-beam data)~10%
LHC physics simulation
Does this meet LHC physics requirements (e.g. for compositeness) ?
Check with (fast ?) simulations thatthis is good enough
If not :
Needs input/help from the experiment physics groups
Alberto Ribon, CERN/PH/SFT
6
Two main types of test-beam setups:
1. Calorimeters: the typical test-beams (made mainly for detector purposes).
The observables are the convolution of many effects and interactions. In other words, one gets a macroscopic test.
2. Simple benchmarks: typical thin-target setups with simple geometry (made, very often, for validation purposes).
It is possible to test at microscopic level a single interaction or effect.
These two kinds of setup provide complementary information!
Type
Validation Validation setupssetups
Alberto Ribon, CERN/PH/SFT
7
Double-differential neutron production Double-differential neutron production (p,xn)(p,xn)
Proton beam energies: 113, 256, 597, 800 MeV
Neutron detectors (TOF, scintillators) at 5 angles.
Study of the neutron production spectrum (kinetic energy) at fixed angles.
Alberto Ribon, CERN/PH/SFT
9
benchmark studies
• ratio simulated / experimental data for data shown on previous slide• error bars include errors from experimental data (stat+syst) and from simulation (stat) - dominated by experimental syst. errors
• typical agreement at level of 1 σ to 2σ
FLUKA
G4: QGSP_BERT
G4: QGSP_BIC
Alberto Ribon, CERN/PH/SFT
10
K. Nakai at al., PRL 44, 1446 (1980)K. Nakai at al., PRL 44, 1446 (1980) D. Ashery et al, PR C23, 2173 (1991)D. Ashery et al, PR C23, 2173 (1991)
pi +/-
thin target (Al, Cu, Au)
detectors
beam monitoring counters
Pion absorption – experimentsPion absorption – experiments
•Nakai – look for gammas emitted after pion absorption
• Ashery – look for transmitted (not absorbed) pions
pi+/- beams ofenergies between
23 – 315 MeV
Alberto Ribon, CERN/PH/SFT
12
Absorption Xsection for Absorption Xsection for pi+pi+
both G4 and Fluka show both G4 and Fluka show reasonable agreementreasonable agreement
in some cases Fluka seems in some cases Fluka seems to be a bit betterto be a bit better
difficult to make more difficult to make more conclusions because of big conclusions because of big uncertainties in the uncertainties in the experimental dataexperimental data
Alberto Ribon, CERN/PH/SFT
13
Absorption Xsection for pi-Absorption Xsection for pi-
same remarks as for pi+same remarks as for pi+ for heavy material (Au) the for heavy material (Au) the
shape of the QGSP_BERT shape of the QGSP_BERT quite differentquite different
G4: best agreement for G4: best agreement for ‘medium-weight’ materials‘medium-weight’ materials
Alberto Ribon, CERN/PH/SFT
14
Hadronic interactions in ATLAS pixel Hadronic interactions in ATLAS pixel test-beamtest-beam
180 GeV/c nominal + beam
Geant4 Geometry. Use the same Geometry also with Fluka,using FLUGG (interface between the Transportation and Physics of Fluka and Geant4 Navigation of the Geometry).
Alberto Ribon, CERN/PH/SFT
17
Ratio max Eloss / total ElossRatio max Eloss / total Eloss
QGSP is in excellent agreementwith data.
Alberto Ribon, CERN/PH/SFT
18
Cluster sizeCluster size QGSP producestoo narrow clusters.FLUKA, LHEPand QGSC are in good agreement withdata.
In conclusion,FLUKA, Geant4are in reasonablegood agreementwith the data, but some observables canbe improved.
Alberto Ribon, CERN/PH/SFT
19
:
ATLAS:
HEC : copper + LAr HEC1 + HEC2, 4 longitudinal compartments 6-150 GeV for electrons; 10-200 GeV for charged pions; 120, 150, 180 GeV for muons.
Tilecal : iron + scintillator tile 2 extended barrel + 1 barrel + barrel 0 modules 20-180 GeV electrons and charged pions; 1, 2, 3, 5, 9 GeV charged pions.
CMS:
combined ECAL + HCAL : ECAL : prototype of 7 x 7 PbWO4 crystals HCAL : copper + scintillator tile each tile is read out independently Max magnetic field of 3 T 10-300 GeV muons, electrons, and hadrons.
LHC hadronic calorimeter test-beams
Alberto Ribon, CERN/PH/SFT
21
0.25 0.65
extended barrelmodule
extended barrelmodule
η=0.25η=0.65
ATLAS Tilecal test beam setup
Alberto Ribon, CERN/PH/SFT
26
ATLAS HEC: energy deposition
There are 4 longitudinal segments: 2 in HEC1 and 2 in HEC2.
F is the fraction of the total energy deposition in each layer.
Alberto Ribon, CERN/PH/SFT
27
ATLAS Tile: π shower profile
barrel / Etot
EB + M0 / Etot
M0
barrel
EB
EB
Alberto Ribon, CERN/PH/SFT
29
Conclusions Geant4 electromagnetic physics has been already
validated at percent level. In the next future, we will try to push it at the permil level.
First round of hadronic physics validation has been completed, with good results.
For the observables we have checked in the case of the simple benchmarks (pixels, neutron double differential, pion absorption) there is a reasonable agreement between data and both Geant4 and Fluka, more or less at the same level.
For the calorimeter test-beams, Geant4 describes well the pion energy resolution, σ/E, and the ratio e/.
The shape of hadronic showers still needs further improvements.
Geant4 studies of radiation background in the LHC caverns are in progress, and they will be soon compared with Fluka.