Abstract
In this study, a non-intrusive and stochastic method is used to accomplish an Uncertainty and Sensitivity
(U&S) analysis in a control rod drop transient. This transient is included in the Anticipated Operational
Occurrences (AOOs). The U&S analysis and perturbation generation is done through the DAKOTA
statistical tool, developed at Sandia National Laboratories. As input parameters to the U&S analysis, 43
different thermal-hydraulic variables are chosen. Similarly, three different output parameters are chosen:
total reactor power, enthalpy and reactivity. The number of total perturbations (146) is obtained using
Wilks¿ formula considering double tolerance limits with 95% of uncertainty and 95% of statistical
confidence for the output parameters. The results include the tolerance bounds of output parameters and
sensitivity of input parameters as a function of time. Therefore, the most important thermal-hydraulic
variables, regarding this AOO, could be isolated. As a new feature in the thermal-hydraulic model, the core
is modeled using fully 3D components. A cartesian vessel is used to model the fuel assemblies (without
collapsing) and a cylindrical vessel is used to model the bypass and downcomer zones.
