The Impact of Fuelling Operations on Full Core Uncertainty Analysis in CANDU Reactors

Abstract

Safety analysis methods have evolved considerably from the deterministic analyses and conservative code predictions towards best-estimate codes with quantified uncertainties. Within these emerging best-estimate-plus-uncertainty (BEPU) approaches code output uncertainties can be derived from comparison against suitable full-scale experiments or from the propagation of fundamental or microscopic uncertainties, or from some hybrid combinations. This paper examine the use of fundamental nuclear data uncertainties for predicting and understanding macroscopic reactor physics behaviours and also how such uncertainty analyses may be impacted by other multi-physics phenomena or by control system actions. Specifically, perturbations to microscopic nuclear data libraries are propagated through reactor physics calculations and provide a measure of the uncertainty in observable outputs from computer codes (E.g., channel power). In addition, daily fuelling operations performed in CANDU reactors also affect fuel element compositions as a function of time and hence also impact the observed channel powers. Hence when comparing propagated uncertainties to channel power measurements one must consider both the uncertainty in nuclear data and fuel composition, but also the role of fueling on the observable deviations in channel power. In this work the SCALE 6.2.2 code was used to generate 151 perturbed multi-group cross sections libraries each based on a set of perturbed microscopic nuclear data. Subsequently theses data were processed into few-group cross sections and used to generate 151 full-core diffusion models in PARCS. The impact of these nuclear data perturbations lead to changes in core reactivity for a fixed set-of fuel compositions of 4.5mK. The actions of a fuelling engineer were simulated using a series of fueling-rules and applied to each one of the 151 full-core models to generate a suitable fuelling history. Each of the cores was then simulated for an extended time period wherein burnup and fuelling operations were included. It was found that the fuelling engineer’s feedback reduced the standard deviation in core reactivity by 99% from 0.0045 to 2.7x10-5. The average relative standard deviation in channel power was found to decrease by 35%, from 3.89% to 2.51% after the simulated actions of the fuelling engineer. Uncertainties in the zone powers were reduced by between 80% and 99%, while the standard deviation in LZC % fill was found to decrease, indicating less intervention from the RRS was required due to the continuous reactivity and power balancing achieved in fuelling selection.