Accession Number:

AD1104979

Title:

Comparison of Modeling Techniques for Radiation Dose in a Realistic Geometry

Descriptive Note:

Technical Report

Corporate Author:

AEROSPACE CORP EL SEGUNDO CA EL SEGUNDO United States

Personal Author(s):

Report Date:

2018-01-14

Pagination or Media Count:

38.0

Abstract:

The Geant4 open-source radiation transport code package provides a variety of ways to model the response to the space radiation environment of sensitive parts or detectors embedded in inert shielding or structural material. The simulated geometry can be built in the C user code entirely from primitive solid shapes boxes, cones, etc., or portions can be imported from CAD computer-aided design files. The nominal radiation transport technique provided is a forward Monte Carlo simulation, which provides the best accuracy for increased simulation speed with some reduction in realism of modeled physical processes, recent versions of Geant4 have also enabled adjoint Monte Carlo simulations, and we have also used Geant4 to implement a variant of the sector shielding technique with some improvements over the standard technique. In this report we compare the results of dose calculations employing various combinations of these capabilities with vs. without CAD import of some of the geometry, and forward vs. adjoint Monte Carlo techniques vs. sector shielding for protons and electrons irradiating a realistic geometry representing our REACH micro dosimeter payload and a portion of its host spacecraft. We find that using CAD import to avoid constructing parts of the geometry by hand-coded C gives results that agree with the fully hand-built geometry, but that runtimes are significantly longer. The adjoint Monte Carlo technique provides a great increase in speed, and results agree well with those from forward simulations for electrons, but the present state of the proton adjoint simulations in Geant4 gives results that are too different from the forward results for us to use them. The improved sector shielding technique gives an even greater boost in speed for protons, and results agree well with those from forward Monte Carlo simulations.

Subject Categories:

  • Numerical Mathematics
  • Computer Programming and Software

Distribution Statement:

APPROVED FOR PUBLIC RELEASE