Compilation of a Prediction-Based Validation Dataset for Heat Transfer Modeling of the Paks Spent Fuel Interim Storage Facility

This study presents and systematizes a high-reliability measurement and technological dataset suitable for prediction-based validation of the Spent Fuel Interim Storage Facility (SFISF) of the Paks Nuclear Power Plant. The primary objective of this dataset is not the validation of a general-purpose software tool, but to establish a reproducible experimental basis for the objective and quantitative validation of a three-dimensional, facility-scale heat transfer and buoyancy-driven flow model of the SFISF, developed using the finite difference method (FDM), in a passively cooled system where heat conduction, thermal radiation, and natural convection simultaneously occur. The applied measurement systems (SMAS, CTRS, and the in-house developed CFEPR), their spatial arrangement, accuracy characteristics, as well as data post-processing and the generation of model execution inputs are described in detail. Special emphasis is placed on the functional separation of the available data into initialization data, model execution data, and independent validation datasets, ensuring that model assessment does not rely on calibration or parameter fitting. Furthermore, the estimation of decay heat generated by the stored fuel assemblies is presented using both a standard correlation method (ANSI/ANS-5.1) and isotope inventory-based calculations, and the discrepancies between these approaches are treated as input uncertainties and sensitivity analysis factors. The spectral solar load is considered based on the ASTM G-173 reference spectrum, while during cloudy periods an effective irradiance estimation derived from on-site lux measurements is applied. The results indicate that the available measurement and technological information is sufficient for supporting reproducible, transparent, and quantitative validation studies of the three-dimensional numerical model of the SFISF, as well as for assessing the impact of dominant input uncertainties.