TY - GEN
T1 - The comet method in 3-D hexagonal geometry
AU - Connolly, Kevin John
AU - Rahnema, Farzad
PY - 2012
Y1 - 2012
N2 - The hybrid stochastic-deterministic coarse mesh radiation transport (COMET) method developed at Georgia Tech now solves reactor core problems in 3-D hexagonal geometry. In this paper, the method is used to solve three preliminary test problems designed to challenge the method with steep flux gradients, high leakage, and strong asymmetry and heterogeneity in the core. The test problems are composed of blocks taken from a high temperature test reactor benchmark problem. As the method is still in development, these problems and their results are strictly preliminary. Results are compared to whole core Monte Carlo reference solutions in order to verify the method. Relative errors are on the order of 50 pcm in core eigenvalue, and mean relative error in pin fission density calculations is less than 1% in these difficult test cores. The method requires the one-time pre-computation of a response expansion coefficient library, which may be compiled in a comparable amount of time to a single whole core Monte Carlo calculation. After the library has been computed, COMET may solve any number of core configurations on the order of an hour, representing a significant gain in efficiency over other methods for whole core transport calculations.
AB - The hybrid stochastic-deterministic coarse mesh radiation transport (COMET) method developed at Georgia Tech now solves reactor core problems in 3-D hexagonal geometry. In this paper, the method is used to solve three preliminary test problems designed to challenge the method with steep flux gradients, high leakage, and strong asymmetry and heterogeneity in the core. The test problems are composed of blocks taken from a high temperature test reactor benchmark problem. As the method is still in development, these problems and their results are strictly preliminary. Results are compared to whole core Monte Carlo reference solutions in order to verify the method. Relative errors are on the order of 50 pcm in core eigenvalue, and mean relative error in pin fission density calculations is less than 1% in these difficult test cores. The method requires the one-time pre-computation of a response expansion coefficient library, which may be compiled in a comparable amount of time to a single whole core Monte Carlo calculation. After the library has been computed, COMET may solve any number of core configurations on the order of an hour, representing a significant gain in efficiency over other methods for whole core transport calculations.
KW - Coarse mesh method
KW - HTTR
KW - Hybrid method
KW - Prismatic block reactor
KW - Transport method
KW - Whole core transport
UR - http://www.scopus.com/inward/record.url?scp=84870324716&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:84870324716
SN - 9781622763894
T3 - International Conference on the Physics of Reactors 2012, PHYSOR 2012: Advances in Reactor Physics
SP - 3744
EP - 3755
BT - International Conference on the Physics of Reactors 2012, PHYSOR 2012
T2 - International Conference on the Physics of Reactors 2012: Advances in Reactor Physics, PHYSOR 2012
Y2 - 15 April 2012 through 20 April 2012
ER -