Quantification of the impact of RANS turbulence models on airflow distribution in horizontal planes of a generic building under cross-ventilation for prediction of indoor thermal comfort

Abstract

Computational fluid dynamics (CFD) simulations based on the steady Reynolds-Averaged Navier-Stokes (RANS) approach are widely employed to predict airflow field and thermal comfort in cross-ventilated buildings. RANS models are highly sensitive to the selected turbulence parametrization. Although the impact of turbulence models on airflow distribution in the vertical center plane of a building has been studied, limited studies have focused on detailed analysis of horizontal planes.

The current study investigated the impact of RANS turbulence models (Standard k-ε, Standard k-ω, Realizable k-ε, Renormalization group k-ε model) on indoor airflow distribution of horizontal planes in an isolated generic building. CFD simulations were validated with wind tunnel measurements. Airflow distribution in each turbulence model was quantitatively assessed in terms of the percentage area of a horizontal plane which holds a certain airflow speed (Au). The indoor thermal comfort in the horizontal planes was also evaluated with predicted mean vote (PMV). Moreover, the surface area, volume, angle, spreading width of incoming jets were studied.

Best overall performance in validation was demonstrated by standard k-ε model. Turbulence models exhibited distinct airflow patterns and significant differences in Au, even in horizontal planes which showed similar results for validation metrics. Similar differences were observed in PMV distribution. This study provides new insight on the prediction of airflow distribution and thermal comfort in horizontal planes, by showcasing the necessity of acquiring wind tunnel data in carefully selected points in the horizontal planes (i.e. not limited to the vertical center plane) for CFD model validation.