MODEL BIASES IN HIGH-BURNUP FAST REACTOR SIMULATIONS

A new code system called the Advanced Reactor Modeling Interface (ARMI) has been developed that loosely couples multiscale, multiphysics nuclear reactor simulations to provide rapid, user-friendly, high-fidelity full systems analysis. Incorporating neutronic, thermal-hydraulic, safety/transient, fuel performance, core mechanical, and economic analyses, ARMI provides 'one-click' assessments of many multi-disciplined performance metrics and constraints that historically require iterations between many diverse experts. The capabilities of ARMI are implemented in this study to quantify neutronic biases of various modeling approximations typically made in fast reactor analysis at an equilibrium condition, after many repetitive shuffles. Sensitivities at equilibrium that result in very high discharge burnup are considered ( and >20% FIMA), as motivated by the development of the Traveling Wave Reactor. Model approximations discussed include homogenization, neutronic and depletion mesh resolution, thermal-hydraulic coupling, explicit control rod insertion, burnup-dependent cross sections, fission product model, burn chain truncation, and dynamic fuel performance. The sensitivities of these approximations on equilibrium discharge burnup, k{sub eff}, power density, delayed neutron fraction, and coolant temperature coefficient are discussed. (authors)