11 Detector Simulation
503
• the production of π 0 particles and the fluctuation of the energy fraction in
this channel which leads to prompt EM energy deposition plays a determining
role in the resolution of calorimeters [94, 95]; these π 0 ’s can be the result
of charge exchange or other hadron–nucleon collisions, or formed by the soft
fragmentation, e.g. in the QCD string view of high energy reactions;
• the simulation of neutron generation, transport and interaction, which contribute
to prompt and delayed signal, activation, and escaping energy;
• a component of missing energy (that influences the resolution) is the energy lost
to nuclear breakup. The accuracy of the modeling of all hadronic reactions, but
in particular the de-excitation stage, determines the quality of its simulation;
• the simulation of leading particles which determines the shower profile—the
profile of energy deposition—and in particular the sharing between longitudinal
compartments and the amount exiting in the direction of the projectile.
The accuracy of the modeling of the time dependence of different interactions,
both in a tool and in an experiment simulation, are also essential.
The fluctuation in the fraction of energy going into each type of secondary
particle (gammas, charged hadrons, neutrons, neutrinos) in all reactions is an
essential feature of a simulation tool. It is important for predicting and modeling
the resolution and other aspects of detector performance.
11.3.11 From Full to Parameterized (‘Fast’) Simulation
In order to obtain high statistics, it is necessary for some applications to utilize
simulation which is much faster (typically by two orders of magnitude) than the
detailed full simulation. There is a spectrum of such simulations with different
approximations and compromises.
The coarseness of simulation, from detailed to fast, is determined by a number
of variables: whether the geometry is described approximately or in great detail;
whether secondary particles are generated from the interaction of primary particles;
the degree to which particles are eliminated during tracking—for example, the
relation between the energy of primary particles and an energy threshold; and the
type of physics models utilized. Choosing the least level of detail and following
only the primary tracks distinguishes the fastest simulation. In other variants, some
aspects are simulated in more detail in order to obtain more precision. Different
ways can be used: simulating more particles, adding physics or geometry volumes.
In the first two LHC runs, it has been possible to produce billions of simulated
events using detailed simulation. In some experiments the forward sub-detectors
were simulated faster, e.g. use by ATLAS of frozen showers for forward calorimeters. Projections for the High Luminosity LHC era foresee an order of magnitude
gap between the statistics possible using GEANT4-based detailed simulation on
projected 2025 hardware and the needs of most analyses. This gap is driving the
continued development of fast simulation methods, and the research into methods
503
• the production of π 0 particles and the fluctuation of the energy fraction in
this channel which leads to prompt EM energy deposition plays a determining
role in the resolution of calorimeters [94, 95]; these π 0 ’s can be the result
of charge exchange or other hadron–nucleon collisions, or formed by the soft
fragmentation, e.g. in the QCD string view of high energy reactions;
• the simulation of neutron generation, transport and interaction, which contribute
to prompt and delayed signal, activation, and escaping energy;
• a component of missing energy (that influences the resolution) is the energy lost
to nuclear breakup. The accuracy of the modeling of all hadronic reactions, but
in particular the de-excitation stage, determines the quality of its simulation;
• the simulation of leading particles which determines the shower profile—the
profile of energy deposition—and in particular the sharing between longitudinal
compartments and the amount exiting in the direction of the projectile.
The accuracy of the modeling of the time dependence of different interactions,
both in a tool and in an experiment simulation, are also essential.
The fluctuation in the fraction of energy going into each type of secondary
particle (gammas, charged hadrons, neutrons, neutrinos) in all reactions is an
essential feature of a simulation tool. It is important for predicting and modeling
the resolution and other aspects of detector performance.
11.3.11 From Full to Parameterized (‘Fast’) Simulation
In order to obtain high statistics, it is necessary for some applications to utilize
simulation which is much faster (typically by two orders of magnitude) than the
detailed full simulation. There is a spectrum of such simulations with different
approximations and compromises.
The coarseness of simulation, from detailed to fast, is determined by a number
of variables: whether the geometry is described approximately or in great detail;
whether secondary particles are generated from the interaction of primary particles;
the degree to which particles are eliminated during tracking—for example, the
relation between the energy of primary particles and an energy threshold; and the
type of physics models utilized. Choosing the least level of detail and following
only the primary tracks distinguishes the fastest simulation. In other variants, some
aspects are simulated in more detail in order to obtain more precision. Different
ways can be used: simulating more particles, adding physics or geometry volumes.
In the first two LHC runs, it has been possible to produce billions of simulated
events using detailed simulation. In some experiments the forward sub-detectors
were simulated faster, e.g. use by ATLAS of frozen showers for forward calorimeters. Projections for the High Luminosity LHC era foresee an order of magnitude
gap between the statistics possible using GEANT4-based detailed simulation on
projected 2025 hardware and the needs of most analyses. This gap is driving the
continued development of fast simulation methods, and the research into methods