(Received 15 April 2011; accepted 19 July 2011)
Published Online: 01 November 2011
| ||Format||Pages||Price|| |
|PDF (1.1M)||9||$25||  ADD TO CART|
Cite this document
Repeatability and reproducibility are typically determined for standard fire test methods on the basis of an interlaboratory test program (round-robin) to define their precision. Significant differences in precision were found in the past between the repeatability and reproducibility of small-scale and intermediate- and large-scale fire test methods that all utilize oxygen calorimetry to measure heat release rate. Repeatability of heat release rate related measurements with small-scale apparatus has been found to be significantly better than the repeatability of the intermediate and large-scale apparatuses. In this paper, the results of two round-robins are compared on the basis of relative repeatability standard deviations and relative reproducibility standard deviations calculated for individual materials for individual test parameters. The first one was a cone calorimeter round-robin and the second was an intermediate-scale calorimeter (ICAL) round-robin. The objective and subjective factors that might have contributed to the differences between the two test methods in both repeatability and reproducibility were analyzed. The most important factors that caused the differences in the ICAL round-robin were higher theoretical uncertainty of the ICAL, inadequate pre-round-robin calibrations, the small numbers of participating laboratories and samples tested, prevalence of fire-retardant-treated materials, and possible failure of some of the laboratories to fully comply with the standard. These factors, especially the fact that the testing procedures and apparatuses in the participating laboratories may not have been in complete accordance with the standards, indicate that the results of the round-robins did not reflect the real precision of the test methods. However, the terminology using the two components of precision, repeatability and reproducibility, is maintained in this paper except where a specific distinction is noted. Recommendations are made on how to improve the results of future fire test round-robins.
Univ. of North Carolina at Charlotte,
Stock #: JTE103916