Use of statistics for predicting distribution of air infiltration.

Calculation of air infiltration in a large number of cases can give information of expected variations in yearly air change and energy consumptions. As model is used the equivalent leakage area model written in a spreadsheet computer program. For a typical Norwegian house an analysis of the influence of some parameters is made. The influence of climate is found to be small, if we compare the mean yearly air change for towns in Norway. Prediction of air infiltration is made from known variations in indoor temperature, 50 Pa pressurization air change and leakage and pressure characteristics. 500 simulations with random generated normally distributed numbers a r e done. The mean yearly air change has been found to be 0,24 in Oslo. But 10% will have values above 0,33. Regression shows a good correlation between 50 Pa pressurization air change and the mean yearly value. A good prediction of the mean yearly air change can therefore be based on the pressurization test without exact knowledge of other parameters. The simulation method can be used on more complicated and correct models to find out how uncertain knowledge of the parameter will influence the final results.