Assessment of computational method for highly subcooled flow boiling in a horizontal channel with one-sided heating and improvement of bubble dispersion

This study employs computational and experimental methods to explore local hydrodynamic and heat transfer characteristics of subcooled horizontal flow boiling of FC-72 in a 2.5-mm × 5.0-mm rectangular channel heated along its 2.5-mm bottom wall. Limited studies have been conducted investigating the limit of the multiphase model and the way to improve the dynamic interfacial behavior between phases. The 3D CFD solver is constructed in ANSYS Fluent in which the Volume of Fluid (VOF) model is combined with the shear stress transport (SST) k-ω turbulent model, a surface tension model, and interfacial phase change model, along with a model for shear-lift force and bubble dispersion. Discussed are simulation results for void fraction distribution, pressure drop, and local fluid and wall temperatures. The simulation results are compared to experimental data to validate the solver's ability to predict the asymmetrically heated horizontal flow configuration. Body force is shown to have an appreciable impact on both predicted and measured results, the outcome of bubbles along the heated wall migrating toward the top adiabatic wall and causing asymmetry in fluid velocity and temperature within the cross-sectional area and along the channel. These asymmetrical effects are quite pronounced at low mass velocities, where the upstream fully developed liquid boundary layer is disrupted by bubble formation along the heated bottom wall. Predictions are presented for simulations, including shear-lift force with and without bubble collision dispersion force. The shear-lift force is shown to govern mostly the dynamic behavior of small bubbles stuck on the heated bottom wall and therefore has a greater impact on both heat transfer and heated wall temperature. Absent bubble collision dispersion force, the simulations show excessive bubble coalescence creating inordinate vapor pockets along the adiabatic wall. But by including this force, coalescence of densely packed bubbles in the bulk region is significantly inhibited, with larger bubbles even incurring additional breakup into smaller bubbles and culminating in far less vapor accumulation along the top wall. Including the bubble collision dispersion force is shown to yield better agreement with experimental interfacial behavior along the channel. The mean deviation and error of wall temperature across cases are 2.2 °C and 3.34%.