2019 article

TREAT M2 experiment modeling for transient benchmark analysis

Sorrell, N. C., & Hawari, A. I. (2019, June). ANNALS OF NUCLEAR ENERGY, Vol. 128, pp. 398–405.

By: N. Sorrell n & A. Hawari n

co-author countries: United States of America 🇺🇸
author keywords: TREAT; Transient; Monte Carlo; Graphite
Source: Web Of Science
Added: May 6, 2019

The Transient Reactor Test Facility (TREAT) is a graphite moderated, air-cooled reactor historically operated for experimental fuel transient testing. This facility has returned to operation, and as part of that effort, the ability to model and predict the conditions for upcoming experiments is paramount to the development of the next generation of transient experiments and fuel testing. In order to design future experiments, methods are being developed to predictively reproduce data from experiments performed at TREAT prior to its shutdown in 1994. In this work, the M2 and M3 experiments, representing a set of transient tests, were selected to explore as TREAT transient benchmarks. In TREAT, transients are controlled largely by control/transient rod movement and temperature feedback that is attributed to the core’s graphite-fuel matrix. To capture these effects, a methodology for modeling these mechanisms using multi-physics coupled Monte Carlo simulations is developed. This includes performing full-core transient simulations using the Serpent Monte Carlo code with feedback based on temperature estimates derived from the OpenFOAM computational fluid dynamics code. Using this methodology, the developed model for the TREAT reactor was able to reproduce the pre-transient power maneuver from low power up to 27 MW and to initiate the power transient for the M2 2580 experiment. The results demonstrate the utility of this approach for dynamic modeling of the reactor. Additional investigation of the impact of key phenomena such as neutron thermalization and temperature response in TREAT’s graphitic core and the surrounding reflector was also performed under steady state conditions using the developed models.