A benchmark test for the isothermal backward-facing step flow

CFD is one of the most important approaches for predicting and evaluating the airflow in the indoor environment. Although there has been a continuous improvement of computer performance and CFD technology in the past few decades, the uncertainty in CFD prediction remains high due to the user dependency on building meshes, selecting turbulence models, defining boundary conditions, and choosing numerical schemes. Benchmark tests are therefore an important approach to validate the CFD models and guide the selection of proper turbulence models under different flow regimes. This study aims to perform a benchmark test for the isothermal backward-facing step flow under different flow regimes. The benchmark test is conducted in a small-scale model and LDA is employed to identify airflow patterns. The benchmark results indicate that the SST k-ω model could be a good option to predict low turbulent flow while the Realizable k-ε and RSM models can be used for fully developed turbulent flow.