Abstract/Summary

With 50 % of the worlds population dwelling in urban environments and over 70 % of people’s time is spent indoors (at home, work or in vehicles). It is important to understand how the urban area effects the internal-external air exchange for buildings and how this may impact on the occupants, though this will differ depending on location. The urban area is complex, requiring multidisciplinary expertise in order to understand the driving features. Urban areas may be simplified down for study to reduce some of the complexity. The study undertaken at Silsoe, UK, used a full-scale staggered array of nine 6 m3 cubes to gain an understanding of the effects of meteorological variables on the natural ventilation rate and pressure coefficient. After 6 months 8 cubes were removed, leaving the instrumented cube isolated for 2 months. All equipment logged constantly, creating a dataset which covers a wide range of wind directions, wind speeds, temperature differences and atmospheric stabilities making the dataset unique from previous work. Changes in wind direction cause changes in the pressure coefficient for both isolated and array cases. However defining wind direction is difficult for the array due to the complex interaction of obstacle wakes. The relation between reference and local wind directions is non-linear. The flow within the array was dominated by mechanical turbulence generated by the wakes of the array elements, with the local turbulence intensity being 7 to 10 times greater than for an isolated cube. The presence of an opening had no effect on the pressure coefficient when acting as an inlet. Stability was found to have no effect due to the building being low-rise and the effects of turbulence could not be discerned from 30 minute averages for both pressure coefficient and ventilation measurements. The full-scale data were compared to a wind tunnel model of the site. This allowed for increased array sizes to be used. It was found that the length and size of the rows have a non-linear effect on the pressure coefficient of a cube within the array, with a limited array reducing the pressure coefficient by 10 to 50 % ± 5 % depending on measurement location. Pre-existing models predict the pressure coefficient for an isolated cube well, but do not accurately predict the pressure coefficient for a limited array due to the lack of wind direction and shielding terms. This is also true for the full-scale data. The three methods used to predict ventilation rate (tracer gas decay, pressure difference and the volumetric method) were all affected by different variables such as the presence of thermally driven ventilation, wind direction, location of the wind speed measurement and amount of turbulence within the flow. The difference in the volumetric flow methods depended on the wind speed measurement used, highlighting the difficulty in gaining an accurate representation of ventilation rate using wind speed alone, especially in an urban area. All three methods show more agreement for the array cases than for the isolated cube cases. Pre-existing empirical models of urban wind speed (CIBSE), pressure coefficient (ASHRAE and AIVC) and ventilation rate do not capture the dual behaviour of the ratio of local and reference wind speeds found for the array. This dual behaviour is demonstrated for 1, 5, 10, 30 and 60 minute averaging periods. This behaviour is not correlated to changes in wind direction, the turbulence or speed of the oncoming flow or internal-external temperature differences. A combination of frontal area density, sheltering factor, wind speed, wind direction, opening location and temperature differences within a ventilation model is required to accurately predict ventilation rate.