Colin Genge
Bibliographic info:
32nd AIVC Conference " Towards Optimal Airtightness Performance", Brussels, Belgium, 12-13 October 2011

A building was tested the equivalent of over 1,000,000 times under windy conditions where each test satisfied the conditions of ASTM, CGSB, ISO, EN, ATTMA and USACE testing standards in every respect. The air flow measurements made at lower reference pressures, such as 4 and 10 Pa, varied over a wide range of +87% to -45% from the average, while the results at 50 Pa varied 15% from the average. The basis for a mathematical model was created to evaluate the error potential of existing standards and to advance new test procedures that would achieve the desired level of accuracy for results referenced to any pressures from 4 to 75 Pa, in the least amount of testing time over the widest possible range of test conditions.
Variations in flow measurements at 4 Pa were reduced to 7%, while at 75 Pa they were reduced to 5.7%. The 95% confidence interval calculation was checked against several batches of tests but the actual confidence interval was no better than a 60%. The new mathematical model could be used to accurately predict the potential variation for any result.