Sensitivity Analysis of a Bottom Reflood Simulation using Morris Screening Method

In the work presented in this paper, a sensitivity analysis using the Morris screening method was carried out for a model of the Flooding Experiment with Blocked Arrays (FEBA) separate-effect test facility. The sensitivity analysis was conducted in the context of preliminary checkup of the model, developed for the thermal-hydraulics system code TRACE. To that aim, 26 model parameters ranging from material properties to specific physical model parameters were selected to determine their importance to the modelling of bottom reflood. The predicted time-integrated cladding temperatures, at three different axial locations of the heated test section, were chosen as the quantity of interest. An implementation of the Morris method was developed in Python programming language. Two sampling strategies resulting in different convergence of the importance ranking. This, however, comes at a rather large overhead due to the additional trajectory optimization calculations. The resulting screening procedure was able to classify the set of parameters either as important or non-important with respect to the model output. Most of the highly important parameters were related to the parameters of the dispersed flow film boiling heat-transfer- and hydrodynamic-related models implemented in TRACE. The model output were also found to be highly sensitive to the spacer grid model. Thus, proper specification to this model is required.