EVALUATING MODELS FOR SUPERPOSITION OF WIND AND STACK EFFECT IN AIR INFILTRATION

Models designed for routine calculations of air infiltration rates into buildings often use a semi-empirical function to combine the separately calculated wind-effect and stack-effect flow rates. The actual superposition of wind and buoyancy-generated pressure fields is a complicated non-linear process that is strongly influenced by the distribution of leakage sites on the building envelope. In the present study large sets of hourly-averaged air infiltration measurements using a constant concentration tracer gas injection system were sorted to separate stack-driven, wind-driven and wind-direction shelter effects. These data sets were then used to test superposition errors for linear, quadrature and flow coefficient methods for superposition of wind and stack effects. By using measured values of wind and stack dominated extremes, tests of the superposition methods were made independent of theoretical models for the wind-effect and stack-effect flows. The best superposition model was one using simple pressure addition. Results show that simple non-linear superposition models are an acceptable approximation to estimate average infiltration rates for combined wind and stack effect.