EPJ Web of Conferences (Jan 2023)

Comparison between GEANT4 and MCNP for well logging applications

  • Varignier Geoffrey,
  • Fondement Valentin,
  • Carasco Cédric,
  • Collot Johann,
  • Pérot Bertrand,
  • Marchais Thomas,
  • Chuilon Pierre,
  • Caroli Emmanuel,
  • Doan Mai-Linh

DOI
https://doi.org/10.1051/epjconf/202328801002
Journal volume & issue
Vol. 288
p. 01002

Abstract

Read online

MCNP and GEANT4 are two reference Monte Carlo nuclear simulators, MCNP being the standard in the Oil & Gas nuclear logging industry. While performing a simulation benchmark of these two software for the purpose of “Cased Hole” wellbore evaluation, discrepancies between MCNP and GEANT4 were observed: computational experiments were performed first in a theoretical and simplified environment using spherical models, then in a more realistic “Open Hole” wellbore context with simplified logging tools. Results of this comparison show an excellent overall agreement for gamma-gamma physics and an acceptable agreement for neutron-neutron physics. However, the agreement for neutron-gamma physics is satisfactory only for certain lithologies and energy windows, but not acceptable for other operating conditions. These results need to be put in perspective with the current use of nuclear simulation in the logging industry. Indeed, wellbore evaluations rely on charts simulated with Monte Carlo codes in various contexts. In the case of radially heterogeneous environments such as “Cased Hole” wellbores, nuclear simulations are mandatory to precisely determine the radial sensitivity of logging tools via the so-called sensitivity functions. The feasibility of wellbore inversion relies on the physical validity of such sensitivity functions obtained from nuclear simulations. This MCNP vs. GEANT4 benchmark was conducted with the perspective to secure the physical fundamentals used for building the sensitivity functions of logging tools.

Keywords