Premium
Correlation between cone calorimeter data and time to flashover in the room fire test
Author(s) -
Östman Birgit A.L.,
Tsantaridis Lazaros D.
Publication year - 1994
Publication title -
fire and materials
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.482
H-Index - 58
eISSN - 1099-1018
pISSN - 0308-0501
DOI - 10.1002/fam.810180403
Subject(s) - cone calorimeter , correlation coefficient , calorimeter (particle physics) , range (aeronautics) , materials science , ignition system , arc flash , linear regression , test data , environmental science , forensic engineering , mechanics , statistics , mathematics , composite material , engineering , chemistry , thermodynamics , combustion , physics , detector , optics , char , organic chemistry , software engineering , insulator (electricity)
Correlations based on linear regressions between data as time to ignition and heat release in the cone calorimeter and time to flashover in the room fire test have been developed. They are a further development of an earlier approach which has been modified and extended to a wider range of surface linings. The correlations apply so far only to surface linings on both walls and ceilings. When the density of the linings as a simplified measure of the thermal inertia is included, the correlations are improved significantly. The new correlations are based on data readily available from the cone calorimeter test at one heat flux level, 50 kWm −2 . The correlation coefficient for the basic relationship, including the density of the linings, is now 0.98 when applied to the 13 linings investigated earlier. This is slightly better than the previous study, in which the best correlation coefficient was 0.96. When applied to 28 linings, the correlation coefficient remains about the same (0.97). Very similar regression equations have been obtained when analysing only 13 products and all 28. This is a strong indication of the general predictive capacity of this approach. The inclusion of other data such as thickness of linings or mass loss during fire does not improve the correlation coefficients. The approach is quite straightforward and simple. However, it has provided a useful prediction which is also valid for an extended range of linings.