486
J. Apostolakis
Simulation serves many purposes at each point in the lifecycle of an experiment
or a facility. Different types of simulation are used, typically with an increasing
level of detail during a lifecycle. At first, fast simulation follows only the most
energetic particles, typically the particles arising from the primary interaction,
through simplified geometrical descriptions to obtain average values for energy
deposition in the key volumes in which a signal is generated. Later, in more detailed
simulation, the interactions create secondary particles, which in turn are tracked
and create further descendant particles; interactions are treated with more detail
and energy deposition includes fluctuations. This enables the estimation of many
quantities including the measurement resolution of detectors, and correlations.
To prepare a proposal for an experiment, different versions of the setup are
simulated. For each design a simplified version is constructed and simulated,
typically using a tool such as SLIC [1] or DDG4 [2] which provide templates
for detector components and hit generation. Putting this together with tailored
digitization and reconstruction, many key characteristics of a design necessary for
technical design report [3] can be evaluated.
The energy dispersion or resolution of calorimeters, the longitudinal and lateral
leakage, corrections to momentum measurements, the backscattering of particles
into trackers or other ‘upstream’ components can be estimated for different designs.
Accurate simulation is a powerful quantitative tool for optimizing the performance,
the size and cost of each sub-detector. In addition, its use extends to quantifying
the tradeoffs between the performance of combinations of detectors, and finally for
optimizing the global performance of a complex modern detector.
During the prototyping and calibration phase of a detector, simulation is utilized
for test beam setups of single detectors or combinations to ensure that their
performance is understood and can be accurately modeled. The accuracy expected
in today’s high precision experiments requires agreement between simulation and
test beam measurements at the level of 1% or better. This type of high-quality
quantitative comparison is the basis for evaluating through simulation significant
corrections to measurements in the experiment that cannot be obtained in a test
beam or directly from in-situ data once the experiment starts operating.
The possibility for detailed modeling of the conversion of energy deposition
into signal within a detector is a key strength of detector simulation. This requires
detailed knowledge of many detector-specific effects. Simulation is an important
tool also for the data analysis phase of an experiment. After its accuracy has
been validated against single-particle benchmarks and test beams, an experiment’s
simulation can be utilized to model tracks of different types, and even full signal
and background events. Detailed quantitative aspects of simulated tracks are used in
preparing methods for particle identification and measurement before the start of an
experiment, and continue to be crucial in many methods that utilize data to calibrate
complex quantities such as the jet energy scale. It can also be used in modeling the
separation of signal from background contributions to measured data.
The impact of a performant detector simulation, whose predictions matched data,
was demonstrated in the first years of the operation of the LHC experiments, during
J. Apostolakis
Simulation serves many purposes at each point in the lifecycle of an experiment
or a facility. Different types of simulation are used, typically with an increasing
level of detail during a lifecycle. At first, fast simulation follows only the most
energetic particles, typically the particles arising from the primary interaction,
through simplified geometrical descriptions to obtain average values for energy
deposition in the key volumes in which a signal is generated. Later, in more detailed
simulation, the interactions create secondary particles, which in turn are tracked
and create further descendant particles; interactions are treated with more detail
and energy deposition includes fluctuations. This enables the estimation of many
quantities including the measurement resolution of detectors, and correlations.
To prepare a proposal for an experiment, different versions of the setup are
simulated. For each design a simplified version is constructed and simulated,
typically using a tool such as SLIC [1] or DDG4 [2] which provide templates
for detector components and hit generation. Putting this together with tailored
digitization and reconstruction, many key characteristics of a design necessary for
technical design report [3] can be evaluated.
The energy dispersion or resolution of calorimeters, the longitudinal and lateral
leakage, corrections to momentum measurements, the backscattering of particles
into trackers or other ‘upstream’ components can be estimated for different designs.
Accurate simulation is a powerful quantitative tool for optimizing the performance,
the size and cost of each sub-detector. In addition, its use extends to quantifying
the tradeoffs between the performance of combinations of detectors, and finally for
optimizing the global performance of a complex modern detector.
During the prototyping and calibration phase of a detector, simulation is utilized
for test beam setups of single detectors or combinations to ensure that their
performance is understood and can be accurately modeled. The accuracy expected
in today’s high precision experiments requires agreement between simulation and
test beam measurements at the level of 1% or better. This type of high-quality
quantitative comparison is the basis for evaluating through simulation significant
corrections to measurements in the experiment that cannot be obtained in a test
beam or directly from in-situ data once the experiment starts operating.
The possibility for detailed modeling of the conversion of energy deposition
into signal within a detector is a key strength of detector simulation. This requires
detailed knowledge of many detector-specific effects. Simulation is an important
tool also for the data analysis phase of an experiment. After its accuracy has
been validated against single-particle benchmarks and test beams, an experiment’s
simulation can be utilized to model tracks of different types, and even full signal
and background events. Detailed quantitative aspects of simulated tracks are used in
preparing methods for particle identification and measurement before the start of an
experiment, and continue to be crucial in many methods that utilize data to calibrate
complex quantities such as the jet energy scale. It can also be used in modeling the
separation of signal from background contributions to measured data.
The impact of a performant detector simulation, whose predictions matched data,
was demonstrated in the first years of the operation of the LHC experiments, during