A Multiscale Coupled Framework for Prismatic HTGR Analysis: Integrating Homogenized CFD with Analytical TRISO Temperature Feedback

Generation IV prismatic high-temperature gas-cooled reactors (PHTGRs) require coupled thermal-hydraulic and neutronic analysis to characterize safety margins, yet explicit resolution of TRistructural ISOtropic (TRISO) particles in computational fluid dynamics (CFD) remains computationally prohibitive. Standard homogenization approaches are efficient, but they do not directly provide the intraparticle temperature information needed for accurate Doppler feedback evaluation. To address this, this study develops a multiscale coupling framework that combines Serpent Monte Carlo neutronics, OpenFOAM CFD, and a dedicated analytical heat conduction solver. In the proposed approach, the fuel compact is treated as homogeneous in the CFD solution, while TRISO coating layer and kernel temperature profiles are analytically reconstructed for each CFD fuel cell from the homogenized compact temperature field and supplied to the neutronics for cross-section evaluation. Compared with recently reported Serpent-OpenFOAM applications to research reactors, small modular reactors, molten salt reactors, liquid metal–cooled reactors, and light water reactors, the present work focuses on a PHTGR with TRISO fuel and distinguishes itself through full integration with Serpent’s unstructured mesh coupling interface and a per cell analytical reconstruction of TRISO temperatures within a block-scale Monte Carlo/CFD framework. The coupling approach is applied to a high-temperature engineering test reactor–based fuel block model, where the coupled simulation resolves intrablock power peaking, absorber proximity effects, and subpin radial power and temperature variations, predicting a maximum fuel rod temperature of 1249 K and a coolant outlet temperature of 950 K. The results show that the framework provides high-resolution thermal feedback information for PHTGR analysis without the prohibitive cost of explicit TRISO meshing and can serve as a high-fidelity reference tool for validation of lower-order system models.