Experimental investigations on the repeatability of real scale fire tests

• We examined the repeatability of design fire scenarios in three test series containing 65 single experiments.
• Measurements included temperature, smoke obscuration, gas quantities and mass loss rate.
• It turned out that temperature and smoke density are the most reliable properties to be predict.
• Gas quantities are highly uncertain properties.

A large number of publications dealing with experimental fire investigations can be found in the literature. Predominantly, results and conclusions are based on single experiments and the influence of stochastic effects, uncertainties as well as error propagations are neglected or had not been discussed. Thus, the validity of the conclusions is not proven and the verification is hindered. In this paper, real scale fire tests had been performed with respect to repeatability and error propagation using statistical analysis. The design of the test series was to minimise systematic errors on the trials in a series, but demonstrate the impact of involved uncertain variables which are random by nature.