Assessment of interfacial area concentration models in RELAP5 and TRACE in application to bubbly and cap-bubbly flows in a large square channel

A thermal-hydraulic system analysis code is essential for evaluating the performance and safety of nuclear reactors. RELAP5 and TRACE are two widely used system codes that employ the two-fluid model to simulate gas-liquid two-phase flows. Interfacial area concentration (IAC) is one of the key flow parameters required to close the two-fluid model, and its modeling performance directly affects the simulation accuracy. Given its importance, several IAC models have been developed using existing experimental databases for various flow channels, including pipes, rod bundles, and rectangular channels. However, no IAC model development has been conducted on large square channels, despite their significance in the design of advanced light-water nuclear reactors, such as the ESBWR. To address the need for reliably analyzing two-phase flows in large square channels, the two-group (2G) IAC models in RELAP5 and TRACE were evaluated using a large square channel experiment that included 2G flow measurements for bubbly and cap-bubbly flows. Here, the 2G approach refers to a method that classifies bubbles into two bubble groups based on the drag coefficient acting on bubbles. Significant prediction errors were identified in both sub-models that comprise the RELAP5 and TRACE IAC models: the 2G void fraction (VF) model and the VF-to-IAC model (calculating IAC from VF). A recently developed 2G drift-flux correlation was recommended to improve the prediction accuracy of the 2G VF. The VF-to-IAC models in RELAP5 and TRACE were modified based on the experimental data. The mean relative deviations of IAC prediction for large square channels were −4 % and −1 % using the modified RELAP5 and TRACE IAC models, respectively, and improved from 73 % and −26 % using their respective default models. © 2026 Elsevier Ltd.