Predicting natural ventilation in residential buildings in the context of urban environments

The objective of this dissertation was to develop, through systematic research and experimentation, a mathematical model for predicting exterior surface pressures and indoor air velocities for small-scale buildings in urban settings. The resulting model is a step-by-step series of functions that produce these results while accounting for various possible geometric relationships between the building and the urban surroundings.; This study was conducted in two phases. The first phase developed an empirical Pressure Prediction Model (PPM) for shielded surfaces using a sequence of wind tunnel tests. The model produces a non-dimensional Pressure Modification Coefficient (Cpm) using a set of geometric variables that describe urban surroundings in terms of obstruction blocks and the gaps between them. A number of empirical corrections account for horizontal displacement of obstructions and for wind direction effects. (Cpm) is then used to calculate the average pressure coefficient on shielded surfaces. The wind tunnel tests show that the shielding effect of an obstruction block is significant within a +/-70° arc around the wind direction, and that it is possible to predict the shielding effect of multiple obstruction blocks within this arc by averaging the shielding effects of individual obstruction blocks and summing the effects of all the gaps.; The second phase concentrated on the development of an Indoor Velocity Prediction Model (IVM). The IVM uses the PPM-predicted surface pressures on shielded walls as input to a model developed by Ernest (1991) to determine the Indoor Velocity Coefficients (IVC). The IVM model also adopts a procedure developed by Arens et al (1986) to convert remote weather station data into site-specific wind speeds. Arens' procedure corrects for the differences in height between the weather station and the site, the differences in terrain roughness characteristics between the two locations, and wind acceleration due to site topography.; The PPM was verified against Wiren's (1984) tests of an instrumented model in different arrays of similarly configured obstruction blocks, and against an instrumented model in a more complex layout. The predicted and the measured pressure values showed a reasonably good fit in both cases. The successes and limitation of the model are discussed.; The IVM predictions of interior airflow were not validated here. Ernest has validated his model in both unobstructed and simply-obstructed conditions, and the PPM is not expected to change the nature of the interior flows predicted by Ernest's model.