Simulation of mixed convection flow in a room with a two-layer turbulence model.

Most indoor airflows are mixed convection. In order to simulate mixed convection accurately and efficiently, this paper uses a two-layer turbulence model. The two-layer model combines a one-equation model for near wall flow together with the standard k-epsilon model for outer-wall flow. The model has been used to predict the mixed convection by displacement ventilation in an office. The computed results agree well with the corresponding airflow pattern and the distributions of air temperature, air velocity, air velocity fluctuation, and tracer-gas concentration. The model can predict correctly heat transfer from a wall where the standard k-epsilon model and re-normalization group (RNG) k-epsilon model with wall functions often fails. The computing cost required by the two-layer model is comparable to that of the standard k-epsilon model and RNG k-epsilon model and is significantly less than that by a low-Reynolds number model.