Burnup-dependent effect of lattice-level homogenization and group condensation on calculated kinetics parameters for CANDU-type lattices

作者：Nichita E^*; Serghiuta D; Podobed S

来源：Nuclear Engineering and Design, 2015, 286: 253-260.

DOI：10.1016/j.nucengdes.2015.03.001

Modern analysis of nuclear reactor transients uses space-time reactor kinetics methods. In the Canadian nuclear industry, safety analysis calculations use almost exclusively the Improved Quasistatic (IQS) flux factorization method. The IQS method, like all methods based on flux factorization, relies on calculating effective point kinetics parameters, which dominate the time behavior of the flux, using adjoint-weighted integrals. The accuracy of the adjoint representation influences the accuracy of the effective kinetics parameters. Routine full core calculations are not performed using detailed models and transport theory, but rather using a cell-homogenized model and two-group diffusion theory. This work evaluates the effect of homogenization and group condensation at different burnups, for three fuel types: natural-uranium (NU) fuel, low-void reactivity (LVR) fuel and Advanced CANDU Reactor (ACR) fuel. Results show that the use of a two-group lattice-homogenized adjoint consistently overestimates the effective delayed neutron fraction by approximately 5% for all three fuel types and over a wide burnup range. The use of a two-group lattice-homogenized adjoint also introduces errors in the effective neutron generation time, but these are at most 1.3% (and their sign changes with burnup). Errors tend to vary with burnup by approximately 1% (of the individual parameter value). If a 69-group lattice-homogenized adjoint is used, the errors drop to approximately 2% for the effective delayed neutron fractions and 0.5% for the effective neutron generation time, for all three fuel types and over the entire studied burnup range. The variation of errors with bumup is also reduced.

• 出版日期2015-5