Analysis and interpretation of the plumbing materials file of the NYS combustion toxicity data bank

Abstract

New York State's Building and Fire Prevention Code includes a requirement that combustible materials used in electrical, plumbing and finishing systems be tested by the University of Pittsburgh protocol and that particular data from that testing be reported on standard forms to the Office of Fire Prevention and Control. The plumbing file in this data bank contains data on 41 products, excluding minor usages. In this paper, these data plus the data on Douglas fir are reported and analyzed. LC50 was found to have significant correlations with indices of thermal instability, of rate of mass loss, and of corrosive irritants. The most parsimonious model developed with stepwise regression correlated LC50 with three factors: %wt loss per degree of temperature rise, temperature at the beginning of rapid weight loss-temperature at which 1% of the sample weight was lost, and temperature at the end of rapid weight loss-temperature at which 1% of the sample weight was lost. LC50 was weakly correlated with maximum CO concentration and in the ‘wrong’ direction (high CO concentrations were associated with low toxicity). Log LC50 was significantly correlated with Log COmax, but also in the ‘wrong’ direction. LC50 was correlated with minimum oxygen concentration with a small R-squared and also in the ‘wrong’ direction. The question of the relevance of these test results to smoke exposures in real fires is discussed.