Numerical Benchmark of Strongly to Loosely Coupled Assemblies Using the Transient Fission Matrix Method

Abstract

Advances in computational methods have given rise to the study and simulation of different aspects of reactor behavior. As such, topics associated with high computational costs become feasible candidates for further investigation and one of them is reactor space-time kinetics (STK). Until recently, STK simulation and point kinetics approximation were limited to deterministic codes, with Monte Carlo codes being too costly to start with. However, recent developments in this area have allowed the use of certain methods in stochastic codes. One such technique is based on the Transient Fission Matrix (TFM) model, a hybrid method that uses a system response obtained through Monte Carlo and stored in fission and time matrices as input for deterministic calculations. The result enables a view of the STK in terms of neutron propagation probability and propagation time across the system. The TFM method was applied to a simple coupled core configuration to generate a numerical benchmark. The Serpent 2 Monte Carlo code was used for the stochastic part of the calculation. The configuration consists of two fuel assemblies placed in a light water tank, with a water blade of varying width between them. TFM, flux and fission results were obtained for varying water blade widths, ranging between 0 cm and 20 cm. The data is then used to analyze the behavior of the system, as well as the effects of the coupling between the two assemblies. As the assemblies move further apart, the system slowly transitions from two tightly coupled assemblies that essentially form a single core, to two almost independent cores. This study enables to produce a benchmark for future calculations and predefine an innovative way of designing high dominant ratio configurations, required for tackling Monte Carlo residual problems. An actual experimental program could be led in ad hoc zero power reactor (ZPR), such as the KUCA reactor of Kyoto University.