Application of GRS Methodology to Coolant Void Reactivity Uncertainty Analysis in the ACR-1000

Abstract

The primary focus of this paper is to describe the methods used to calculate the uncertainty in the full-core, instantaneous, coolant void reactivity (CVR): this is defined as the change in reactivity from a fully cooled core at nominal conditions to a totally voided core, where the only change between the two states is the coolant density. This uncertainty analysis considers only equilibrium conditions. Therefore specific operational states (such as initial core, transition to reference core, and start-up after long shutdown) are outside the scope of the work. The ACR-1000 is designed to have a slightly-negative full-core CVR under nominal design-centre conditions so that if all the coolant is lost, core reactivity would still decrease.The coolant void reactivity uncertainty analysis for ACR-1000 relies on an integrated approach based on the Best Estimate and Uncertainty (BEAU) analysis and Gesellschaft fuer Anglagen und Reaktorsicherheit (GRS) method of uncertainty calculation, using the MCNP5[1] code. This analysis method identifies the sources of uncertainty that have an impact on the value of the safety margin parameter, determines their associated uncertainty range and distribution, and ranks them through a Phenomenon Identification and Ranking Table (PIRT) process. The sources of uncertainty considered are, for example, those associated with the boundary conditions in the analysis and core representation (modeling of the core). The identified uncertainties are then propagated through the analysis to provide an overall uncertainty assessment for the safety margin parameter, and the one sided 95%/95% tolerance limit for the Figure of Merit (FOM) CVR is evaluated by the Ordered Statistics approach used in the GRS method.