Physics:Simulation of showers

Although the basic physical processes occurring in electromagnetic showers are well known, this is not quite so for hadronic showers. The simulation of showers in calorimeters needs to follow all particles to rather small energies; for hadrons, phenomenological approximations for intra-nuclear cascades and intermediate-energy processes have to be made (see Ferrari93, Ferrari97), and also electromagnetic simulation results can be sensitive to multiple low-energy cutoff parameters. The number of particles in a shower is very large, particularly at high energies, so that even the computing resources of large laboratories can be challenged by full simulation programs. Simulation being a central tool in optimizing calorimeters, the validity of results and questions of efficiency on computers has led to multiple publications and comparisons (e.g. Fesefeldt90b). Much overall understanding can be derived from average shower parameters . (see Electromagnetic Shower, Hadronic Shower). Full shower simulation programs are EGS (see Nelson85) for electromagnetic showers, or GEISHA (see Fesefeldt85, Fesefeldt90a) or FLUKA (see Fass`o93) for both electromagnetic and hadronic showers. One should note that multiple versions for these programs exist, and that they are usually applied embedded in programs allowing one to introduce the geometry of a detector. Recently, many initiatives exist to allow the detector designers' form of describing detector components (in a CAD program's data base) to enter directly the simulation programs (which simulate not only showers, but also tracking detectors). Up-to-date manuals and specialist expertise are needed for more information.

The tuning of the simulation parameters, in particular of cutoffs and integration step sizes, is delicate and depends on the goal of the simulation; some codes (e.g. FLUKA) use mathematical methods (like importance sampling) to achieve robustness and speed up calculation, alleviating somewhat this tuning problem. In major simulation projects at high energies users have also resorted to relatively high cutoff parameters, using for further shower development randomly selected showers at lower energies, precomputed in full simulation and stored in ``libraries (see Graf90). This strategy can save substantial amounts of computer time.