Three-Field subchannel modeling of CANDU fuel thermal-hydraulics with MEFISTO-T code

Abstract

To support the development and validation of core thermal–hydraulic simulation codes, OECD/NEA organized a benchmark focused on thermal–hydraulic behavior of CANDU fuel assemblies. The benchmark was based on full-scale heat transfer experiments conducted in a 28-element horizontal test bundle, covering conditions representative of CANDU core operations. A non-blind phase was included, providing a limited set of axial pressure drop (single-phase and two-phase) and critical power data, followed by a blind phase for predictive assessment. The tested geometries simulated CANDU pressure tubes under both uncrept and crept (i.e., aging) conditions, as well as two different bearing pad heights. The crept geometry introduces a significant bypass on the upper side of the rod bundle, associated with complex crossflows that are notoriously challenging to capture by simulation codes. The Westinghouse MEFISTO-T subchannel analysis code, employing a mechanistic three-field model of annular two-phase flow and originally developed for BWR fuel applications, was applied to simulate all benchmark cases. The code calibration parameters related to form loss and drop deposition enhancement due to structural components were adjusted based on the provided dataset. For the non-blind phase, the code accurately predicts axial pressure drop and critical power, and correctly identifies the dryout rod and orientation, although with a downstream bias in axial location. Further results from the blind phase demonstrate generally good agreement, particularly in capturing the effect of pressure tube creep on axial pressure drop and the influence of system parameters (pressure, mass flow rate and inlet temperature) on critical power. However, the significant reduction in critical power observed in the tests due to pressure tube creep and lower bearing pad height was underpredicted, indicating areas for further model refinement. These results highlight the strong potential of applying the MEFISTO-T three-field model to CANDU fuel geometry, while also revealing opportunities for further improvements.