The report accounts for judgements and tests that are carried out as a background for the revision of method NT Build 307 "Roof Coverings; Wind load resistance" originally published in November 1986. The goal of the revision was to define the method more precisely, to give a better description of the test rig and the routines concerning applying and testing.
The load diagram for alternative A) static pressure test, and especially alternative B) cyclic pressure test are evaluated and adjusted. Eventually some tests have been carried out according to the revised method. On this basis some conclusions have been drawn. Most of the conclusions from the project work have been taken into account proposing a revised method, but it is indicate that the effect of temperature has to be investigated further before at revised method can be accepted as final.
It has been considered whether the load acceleration of each gust should be increased. The Nordtest-method, load-alternative B, include cyclic load, but each gust have so long time of insert that the dynamic effect is very small. To make changes here were regarded as a big interference and should have been subject to a thorough examination before it could be included. This was considered to be outside the framework of this project, and therefore left out. (An investigation of the effect of the dynamic load can how ever be done as a separate study later). Because of this fact the methods ability to reproduce self-unscrewing of screws mounted in the profiled steel decks is not improved. This is a pity since this was one of the goals regarding this revision work. Here we make reference to the revision work concerning method NT Build 306 which also includes reproducing self-unscrewing.
The Nordtest-method "NT Build 307, Roof Coverings; Wind load resistance – Rev. 2002", load alternative B) Cyclic pressure test as proposed, was tested with a limited number of experiments. The results from these experiments show a very good repeatability.
The revised method now more than earlier absorbs several of the relationships that affect the results. As mentioned before this has given a method with good repeatability, but it has also resulted in lower capacities than before. The capacity at failure for bituminous roof membranes seems to have been reduced more than the corresponding capacity for polymeric membranes.
This can be explained by the change of test temperature; now both group of product (bituminous and polymeric membrane) have been tested at room temperature. The effect of different test temperature should have been investigated more. It is proposed to perform an evaluation of what is the correct testing temperature, or if a factor of temperature correction should be introduced.