Validation Studies of RELAP5-3D for High-Temperature Gas-Cooled Reactor Analysis Using a Refined RELAP5-3D Model of the High Temperature Test Facility

RELAP5-3D has been used extensively for high-temperature gas-cooled reactor (HTGR) analysis, but such analysis lies outside the validation basis for the code. Oregon State University, the U.S. Nuclear Regulatory Commission, and the U.S. Department of Energy’s Advanced Reactor Technologies gas-cooled reactor campaign worked together to design and operate the High Temperature Test Facility (HTTF) at Oregon State University to provide validation data for prismatic HTGR thermal-hydraulic analysis.Work using a legacy RELAP5-3D model of the HTTF demonstrated the ability to reproduce trends in HTTF data, but an inability to reproduce the values measured in multiple experiments within the uncertainty. Previous work hypothesized that a new model with a finer radial nodalization would perform better for HTTF analysis.In this paper, we present the development of that more refined model and analyze HTTF experiments PG-27 and PG-29, pressurized and depressurized conduction cooldown transients, respectively. Comparing the PG-27 results to the experimental data showed a better prediction of steady-state and transient temperatures; however, a deeper dive revealed that the better transient temperature predictions are a result of better initial conditions. The transient temperature rise in the core is comparable between the two models, but the legacy model predicted temperatures in the reflectors better.The new model demonstrated the ability to predict unique temperatures in each sector of the core during PG-29, capturing the azimuthal asymmetry; however, the temperature predictions still differed from the measured temperatures. These results suggest that gaps and uncertainties in the measured boundary conditions for the HTTF tests do not allow for fully validating RELAP5-3D for prismatic HTGR analyses. We demonstrated improved validation for steady-state analysis, but significantly improved transient performance has yet to be recognized.