Transcript clasPoster5 - University of Richmond

Introduction
Jefferson Laboratory (JLab) is a US national laboratory built to probe the
atomic nucleus and to unravel the quark-gluon nature of matter. Our best
theory of the strong interaction that binds nuclei together is Quantum
Chromodynamics (QCD) which is not well understood in the nuclear
environment. In this project we develop and optimize simulations of the
proposed CLAS12 detector as part of the JLab 12-GeV Upgrade.
CLAS12 Simulation Analysis
and Optimization
K.Dergachev and G.P.Gilfoyle
University Of Richmond - Department of Physics
CEBAF
The Continuous Electron Beam Accelerating Facility (CEBAF) is the heart of
the JLab complex and produces an extremely precise electron beam. The
accelerator is 7/8-mile around and sits in a racetrack-shaped tunnel 25 feet
underground. It is a superconducting linear accelerator capable of producing
electrons with energies up to 6 GeV. The electron beam makes up to five
laps around the track before being steered into one of the three existing
experimental halls; Halls A, B, and C in the figure below.
Figure 1: Jefferson Lab
Plugins
Plug-ins for event generators have been developed to separate the event
generation from the simulation. Developers can test different event
generation schemes without recompiling Sim12. The existing types are:
TXT – Text-based format supporting unlimited particles per event. Each
particle is characterized by Particle Data Group identification number,
vertex position (in cm), and 3-momentum (in GeV/c). File extension is .txt.
LUND – Text-based format that follows the LUND scheme for describing
the contents of an event (T.Sjostrand, Comp. Phys. Comm. 82 (1994) 74).
For particle decays only the final particles (children) are kept.
Upgrade Rationale
JLab was built to study the transition from the proton-neutron-meson picture of atomic nuclei to one based on quarks, gluons, and
QCD. The original accelerator was designed and built with an eye to the future to eventually increase the beam energy. In the time
since, a compelling scientific case has been made to upgrade CEBAF and the end stations to more fully explore this transition
region. The JLab 12-GeV Upgrade is now the highest priority project in nuclear physics in the United States.
CLAS12
To take full advantage of the new physics opportunities with the
upgraded CEBAF, a new detector called CLAS12 will be built in Hall B. It
will retain many design features of the original CLAS, but will have a
new superconducting magnet coils and additional improvements.
• Higher luminosity for hydrogen targets (10-fold increase over CLAS).
• Improve small-angle acceptance for electrons and large-angle
acceptance for recoil baryons.
Figure 4: CLAS 12 Event Simulation
• Increase momentum range of electron-pion separation.
• Better scintillator segmentation.
Simulation Results
• Add another Cerenkov counter.
Preliminary results from Sim12 are shown in Figures 4-6. Figure 4 shows
the event display for a single electron. The light blue components are the
drift chambers, the yellow ones are the torus magnet coils, the dark blues
elements are the electromagnetic calorimeters in the forward and central
detectors. The electron path is the straight, red line (the magnetic field is
off here) which shows secondary electrons knocked lose by the passage of
the scattered electron beam.
• Add central detector with dual purpose solenoid magnet design.
• Replace current mini-torus for Moller electron shielding.
• Provide magnetic field for large angle momentum analysis.
A
B
Figure 3: CLAS 12
C
CLAS12 Simulation
CLAS – 6 GeV
The CEBAF Large Acceptance Spectrometer (CLAS) in Hall B is used to
detect electrons, protons, neutrons, pions, photons, and other subatomic
particles. Drift chambers determine the path of charged particles after
bending in a toroidal magnetic field produced by superconducting electric
coils. Cerenkov counters separate electrons from negative pions.
Scintillators determine time of flight, and calorimeters measure energy. The
information from these components is used to determine particle momentum
and energy and to identify each particle.
Figure 2: CLAS
A detailed simulation of the CLAS12 (named Sim12) is
necessary to validate the design decisions of the upgraded
detector, understand the detector response, and eventually
for acceptance calculations (acceptance is the effectiveness
of the detector at measuring different events). A group at
JLab has already begun developing a GEANT4-based,
object-oriented simulation package.
GEANT4 - A software toolkit from CERN for the simulation
of the passage of particles through matter.
XML-based geometry - The geometry of CLAS12 will be
described with the Extensible Markup Language (XML). XML
is simple, flexible format used to define the structure of
information and to exchange data between different
processes, sites, etc.
Plug-ins for event generation - There are already a variety of
event generators available to members of the Collaboration.
Plug-ins will be used to pass simulated event data to Sim12.
This technique creates a clean separation between event
generation and simulation. Developers do not have to rebuild
Sim12 to test a new event generator.
Hit generation schemes – A library of generic detector types
have been written to give greater control to the user. The
choice of hit generation is made in the geometry files so the
method can be altered without modifying the Sim12 code.
Output data – Output from Sim12 is written in a form called
EVIO that is compatible with the CEBAF Online Data
Acquisition (CODA). Utilities exist to produce ROOT
histograms from EVIO files produced by Sim12.
Installation and
Run-Time Optimization
Figure 5
For Sim12 to be a useful, it must be quick to implement for new
users. We have written procedures, scripts, and documentation
to speed installation and optimize running.
Installation – Sim12 requires a large number of software libraries
shown in the table below.
CLHEP
MOTIF
HDDS
GEANT4
DAWN
JGEANT4
XERXES-C
EVIO
SIM12
Table 1. List of software packages for Sim12.
A detailed installation guide was written to guide new users
(http://clasweb.jlab.org/wiki/index.php/Getting_Started_Sim12_Offsite). Compile time is still an issue (up to 2 hours), so binary
packages were created allowing users to simply extract the files
from an archive for the Suse 10.2 and Fedora Core 6/7 Linux
distributions.
Run-time – Event Simulations are typically slow (a few Hz) and
the graphical user interface (GUI) for CLAS12 is computationally
intensive. The GUI was operated in either immediate mode or
stored mode. The former has no limitation on data size, and the
latter is fast for visualizing large data repetitively, and so is
suitable for animation. In stored mode, the OpenGL function
calls that have been executed are kept in a display list where
they can be executed after creation without the overhead of any
processing that was performed to obtain those function calls.
This enables much faster rendering of OpenGL at a cost of
increased memory use.
The histogram to the
left shows a simulation
of the energy lost by a
sample of 13,500
scattered electrons that
passed through region 1
of the CLAS12 drift
chambers. The region 1
drift chambers consists
of six layers (layer one
is the closest to the
target). As the electrons
move through the drift
chamber gas, they lose
increasing amounts of
energy
Figure 6
The histogram to the
right shows a simulation
of the electron vertex
position as a function of
the position along the
beam axis (z) and
perpendicular to it (y).
The target is treated as a
point giving rise to the
large peak at the origin.
The events downstream
of the target are beam
interactions with air or
secondary interactions by
small-angle electrons.