A flexible methodology for generating 3D graphite dosimetry data in Advanced Gas-Cooled Reactors

Abstract

The safety cases for the UKs Advanced Gas-Cooled Reactors require a robust understanding of the structural integrity of their constituent graphite bricks. Both neutron damage and graphite weight loss have significant effects upon material properties. Graphite weight loss occurs via radiolytic oxidation, initiated by neutron and gamma-ray dose. This includes contributions from fast neutrons, gamma-rays from the fuel, and secondary gamma-rays. An accurate knowledge of the dosimetry history of trepanned graphite samples is important, as well as the ability to make forward predictions for any brick within the core. It is therefore necessary to calculate both these dosimetry quantities with a high degree of spatial accuracy, taking into account the time-varying nature of local fuel channel powers as well as the way in which weight loss itself affects the dosimetry data. The Monte Carlo method is the most accurate available for the calculation of radiation dosimetry data. However, it is impractical to undertake full scale Monte Carlo calculations for a whole reactor core and to represent every required period of time and graphite weight loss, which will also vary in time and space. An alternative approach has been developed by Jacobs which exploits the high accuracy afforded by the Monte Carlo method, but is also flexible and practical enough to provide 3D dosimetry data for any graphite brick within the active cores, at any point in time and at any required graphite weight loss. The Power Partitioning Method uses MCBEND-generated dosimetry “Contribution Matrices”, coupled with PANTHER “core follow” fuel powers, to calculate dosimetry data anywhere within selected fuel channels or interstitial bricks.