Verification of calculation models of air infiltration using three types of test houses.

In order to verify the calculation models of air infiltration using three wooden test houses which have the same type of construction but have different leakage distributions, airtightness of building components of these three houses were measured by means of the fan pressurization method, and then air infiltration was measured twenty-two times by C02 concentration decay technique. Some of the leaks were sealed so that total leakage of each of the three houses was equal, but the leakage distribution was different between House A and House B, and the amount of total leakage of House C was twice that of House A and of House B. Secondly, air infiltration was calculated by means of the LBL model and the BRE model. The values as calculated by the LBL model were unexpectedly two to three times the values of measurement. The values as calculated by the BRE model were in the range of one to two times the values of measurement. Thirdly, calculation by means of the JCV model widely used in Japan was done under the assumption of there being five types of leakage distribution in order to clarify the effect of leakage distribution on the accuracy of estimation. With the JCV model the internal pressure and the air infiltration are calculated by the Newton Raphson method using an equation in which the total infiltrationis zero. Before this calculation, the pressure difference due to the combination of the wind and stack effects is introduced into the leakage equation for each leak. As a result , the best estimation is yielded by the uniform distribution as opposed to the other distributions . So, it can be said that the assumption of uniform distribution of leakage can be accepted in the case of a house which is not so much airtight .