Distributions of expected air infiltration and related energy use in buildings based on statistical methods with independent or correlated parameters.

The equivalent leakage area algorithm is used to illustrate the use of statistical simulations to predict distributions of infiltration and energy loss for buildings. The important parameters in the model are: leakage at 50 Pa pressurisation, indoor and outdoor temperature, leakage in the ceiling and the floor, wind speed, building height and shielding class. Most of these parameters are not known accurately. In the statistical method we assumed for each a distribution based on measurement or good guess. To find the resultant distribution of infiltration we make 1000 simulations with random values from the distributions of the parameters. This has been done with normal, uniform and/or Weibull distributions. We find that the infiltration energy loss for a house in Oslo in average is 7025 kWh/year (case 1) and 10% is above 10500 kWh/year and 10% is below 4200 kWh/year. The results are also given as mean air change for infiltration. Results from five other simulations are described. We normally assume that the parameters are uncorrelated, but it is easy to calculate what happen, if we have a positive or negative correlation. We found, that in this case made correlation between some parameters practically no difference in the distributions. The simulation method can be used on more complicated and correct models to find out how uncertain knowledge of parameters will influence the final results.