Evaluating single-sided natural ventilation models against full-scale idealised measurements: Impact of wind direction and turbulence

Commonly single-sided natural ventilation is used in temperate climates to provide comfortable and healthy indoor environments. However, within built-up areas it is difficult to predict natural ventilation rates for buildings as they depend on many flow factors and opening type. Here, existing models are evaluated using the nine-month Refresh Cube Campaign (RCC). Pressure-based ventilation rates were determined for a small opening (1% porosity) in a cubical test building (side = 6 m). The building was isolated and then sheltered in a limited staggered building array to simulate turbulent flows in dense urban areas. Internal and external flow, temperature and pressure measurements captured a wide range of scales of variability. Although the Warren and Parkins (1985, WP85) model performed best for 30-min mean ventilation rates, all four models tested underestimated ventilation rates by a factor of 10. As wind dominated the stack effect, new coefficients were derived for the WP85 wind-driven model as a function of wind angle. Predictions were mostly improved, except for directions with complex flow patterns during the sheltered case. For the first time, the relation between ventilation rate and turbulence intensity (TI) around a full-scale building was tested. Results indicate that the wind-driven model for single-sided ventilation in highly turbulent flows (0.5 < TI < 4) can be improved by including TI as a multiplicative factor. Although small window openings with highly turbulent flows are common for sheltered buildings in urban areas, future model development should include a variety of configurations to assess the generality of these results.