LEVERAGING SENSITIVITY ANALYSIS TO INFORM SEPARATE EFFECTS TESTS FOR NUCLEAR REACTOR SAFETY APPLICATIONS

Future commercial nuclear power is seeking enhanced performance through technologies that operate outside the current regulatory space. Accident tolerant fuels (ATFs) are a proposed method for enhancing safety in nuclear systems without sacrificing cost or performance. However, prior to licensing a new fuel form for commercial use, fuel qualification must be performed. Traditionally this process is time-consuming and costly, taking upwards of 20 years to perform and relying primarily on integral effects tests. A novel solution to accelerating the fuel qualification process is supplementing integral effects testing by leveraging an iterative process between separate effects tests and modeling and simulation to establish a database of fuel performance. Separate effects tests are cost-effective experiments which isolate specific fuel performance behaviors to provide fundamental data for developing theory and predictive models. Predictive models can then be used to inform the design of experiments to address regions of high uncertainty through the use of sensitivity analysis. The goal of my work is to identify high-impact, high-uncertainty parameters within predictive nuclear models to inform separate effects experimental design. This approach will be applied to two individual case studies.

The first study performs a knowledge gap analysis of transient critical heat flux and transient boiling heat transfer modeling within the RELAP5-3D code. I inspect discrepancies between experiments and code predictions due to rapid transient heating effects by performing experimental comparison and highlight areas of highest modeling uncertainty using sensitivity analysis. To address this region of high-uncertainty, I proposed an experimental test matrix that can provide insight for future models of transient boiling heat transfer and transient critical heat flux within this operational regime.

The second study inspects modeling nuclear fuel irradiation tests in collaboration with the Oak Ridge National Laboratory (ORNL). I perform an in-depth analysis of the irradiation subcapsule design to reduce model variance in fuel temperature predictions. This analysis revealed three important design parameters that can be controlled: initial gas pressure, contact pressure, and surface roughness. Using the results of the predictive model, I have developed a set of design recommendations which will reduce fuel temperature variance in future separate effects experiments.