Improving Reliability of Novel TRISO Fuel Forms for Advanced Reactors via Multiscale, High-Throughput Characterization and Modeling

The anisotropic nature of graphite flakes and the unique fabrication routes for TRISO containing fuel forms leads to possible local and long-range texture. This texture could produce anisotropic bulk properties, as the graphite flake alignment surrounding particles or in a pebble's fuel free-region could be direction dependent. The inability to model local and long-range anisotropic thermal properties based on our current understanding of properties may result in either an over- or underprediction of fuel (and particle) temperatures. This has implications regarding prediction of fuel performance, as many relevant phenomena are related to temperature including, but not limited to, fission product release. Ultimately, accurate models of integral fuel form thermal properties supports a more accurate determination of fuel temperatures. We propose the use of a parallelized thermal conductivity (k) measurement device coupled with multiscale models to accurately predict the thermal conductivity of TRISO fuel composites. This proposal overcomes the issue plaguing many 'localized' microscale measurements, namely the inability to scale local measurements up to engineering scale properties. We will do this by using Bayesian inference techniques on the 105 k measurements we can collect in 1 week to predict an effective thermal conductivity of each sample and validate this value compared to a bulk k measurement computed from thermal diffusivity, specific heat, and density measurements. Quality assurance of the measured microscopic properties will be ensured by calibration of the P-SDTR device compared to standard reference materials and validation of local k measurements at several key locations with the IR microscope, albeit with a lower throughput.