Conference Agenda

This study presents a computational fluid dynamics (CFD) methodology for simulating multi-regime two-phase flow in an in-line tube bundle. The approach is based on a mixture multiphase model, incorporating scale separation between large interfaces resolved according to the mesh size and the dispersed phase modeled using a three-dimensional slip velocity to account for kinematic disequilibrium between phases. The primary objective was to validate this methodology by comparing CFD results with experimental data. The quantities of interest are the Power Spectrum Density of the forces applied in a tube and the local distribution of the void fraction around an instrumented tube. Analysis shows good agreement between CFD results and measurements regarding the quantities of interest investigated. This work demonstrates the consistency and the reliability of the methodology for two different mixtures: Water/Air and fluids simulating water/steam. In addition, this approach reduces the computational complexity in comparison to two-fluid models while maintaining a good accuracy in predicting two phase flow topology in the tube bundles.

This work represents a significant advancement towards developing a one-fluid formulation methodology for simulating complex multi-regime two-phase flows in tube bundles, with particular emphasis on studying fluid-structure interaction (FSI) in the U-bend region of steam generators.

The OpenSTREAM computational environment is a new open-source platform designed to facilitate efficient and collaborative development and validation of one-dimensional, multi-field, two-phase flow simulation models across research institutions. It includes several simulation frameworks: a mixture model, a two-fluid model, a three-field model, and an advanced four-field model of annular two-phase flow. The current implementation supports single-component, incompressible, steady-state, and transient boiling two-phase flows in single straight channels under reasonable simplifying assumptions. The two-fluid model solves a six-equation system governing mass, momentum, and energy conservation for each phase, capturing hydrodynamic and thermal non-equilibrium effects. The three-field model follows a classical framework (vapor, drops and film) for annular two-phase flow, while the advanced four-field model explicitly represents both the base liquid film and dispersed disturbance waves as separate fields. In all solvers, field interactions and wall closure models have been implemented either from well validated models from the literature or from simple considerations, providing a foundation for future collaborative improvements. Simulations of a representative boiling water two-phase flow case using all simulation frameworks show consistent and reasonable predictions. A comparison with the TRACE system code demonstrates that the implemented two-fluid solver produces reliable and consistent results. Finaly, the validation exercises from the original four-field model development are reproduced. OpenSTREAM, along with its validation and application database, will soon be publicly available on dedicated GitHub repositories under the permissive MIT license.

Since innovative reactors' flow conditions and geometries may differ from those of conventional reactors, the applicability of the models and correlations used in the design works should be appropriately checked. In the design phase, there is a high possibility that the flow conditions and geometries will be changed. Therefore, applying detailed numerical simulation is expected to achieve efficient design works. In nuclear reactors, two-phase flow will appear in many situations. Confirming the applicability of models and correlations for two-phase flow conditions is important. The detailed two-phase flow simulation codes must be useful to confirm the applicability of two-phase flow models and correlations. However, there is no established methodology to properly validate detailed two-phase flow simulations. We have, therefore, started the research project to develop a methodology for validating detailed two-phase flow simulation codes. In this project, we have developed two-phase flow measurement technologies to obtain detailed information on the gas-liquid interface and two-phase flow database by using developed measurement technologies and performing detailed two-phase flow simulations for the developed two-phase flow database. We will compare detailed two-phase flow simulation results and the two-phase flow database and discuss the proper methodology with the reactor manufacturer. Finally, we will investigate the validation process of a detailed two-phase flow simulation code. In this presentation, we will show the outline of this project and future plans.

Two-phase liquid-gas flow has a wide variety of applications in the nuclear industry, including active thermal control systems, steam generators, and nuclear reactors. In order to model and predict the pressure drop and flow regimes in a reactor core, the void fraction must be accurately predicted. This paper presents a new mathematical model that can accurately predict the two-phase void fraction requiring only knowledge of the geometry of the channel, liquid and vapor mass flow rates, and properties of the working fluid. The predicted void fraction is validated by void fraction data collected using an in-house capacitance sensor and a unique vertical, air-water flow calibration loop. Compared to measured void fraction data, the new mathematical model has a better performance than commonly used models such as Lockhart Martinelli model, Wheeler model, Chen model and homogeneous model.

Two-phase flows are critical in various industrial settings, including nuclear reactors, heat transfer systems, chemical processes, and wastewater treatment. Air bubbles within these flows enhance mixing, enable oxygen transfer, and alter heat fluxes. In nuclear reactors, bubble dynamics and oxygen transfer play pivotal roles in containment cooling, pressure control, gas stripping, hydrogen/oxygen management, corrosion control, and thermal-hydraulic modeling, making a comprehensive understanding essential. Computational Fluid Dynamics (CFD) simulations offer powerful insights into these systems beyond what sensors alone can achieve.

This study examines the impact of two diffuser configurations on flow mixing and oxygen transfer in a 1.3-meter water-filled reactor with 16 air diffusers. Two configurations were tested: all diffusers active (configuration A) and only the central lines active (configuration B), both operating at a flow rate of 20 m³/h. Simulations using OpenFOAM's twoPhaseEulerFoam solver incorporated the two-film resistance model with Clift’s mass transfer coefficient. Experimental data collected on void fraction, bubble velocities, and liquid flow supported validation.

Findings showed strong alignment between simulated and experimental results for void fraction and velocity profiles, allowing for detailed analysis of flow patterns. Configuration B demonstrated a 15% reduction in oxygen transfer efficiency experimentally, while CFD predicted a 24% decrease, effectively capturing the trend. These CFD simulations offer pre-construction insights into diffuser performance, informing design decisions on hydrodynamic interactions and oxygen transfer efficiency. Future work will enhance model accuracy and explore additional flow rates and dynamic aeration configurations.

While many advanced reactor-design concepts do not rely on subcooled boiling, the Pressurized Water Reactor still dominates world-wide installed capacity and accurate prediction of transient and steady characteristics of the nucleate boiling heat transfer regime has a first-order impact on the reactor design efficiency and safety margins. After more than 70 years of study there remain gaps in knowledge and uncertainties in empirical models and correlations. With the continued increase in available computational power, interface resolving high-fidelity simulations have become an important tool in closing these gaps in knowledge. Numerical investigations at practical scales involving thousands of bubbles are now possible. However, resolving the micro scale surface topology and roughness necessary for in situ prediction of bubble inception and inertial growth remains computationally out of reach for the foreseeable future. In this work, we will present a closure model of vapor bubble nucleation waiting time and inertial phase growth aimed at reducing uncertainty in existing high-fidelity numerical investigations of nucleate boiling heat transfer. The model is based on a simplified energy and force balance on the extant vapor bubble retained in the micro-cavity. The inception of vapor bubble growth considers local thermodynamic effects and surface conditions and is formulated as a renewal time. Care has been taken to provide a numerically stable and computationally efficient closure to the higher-order thermal hydraulic simulations.