Charring rates for large‐section timber based on experimental data have been generally established. The established rates may not be appropriately used for the prediction of failure times of lumber members which are small by comparison. It is questionable whether a constant rate can be safely assumed for lumber members since the rate is likely to increase once the centre‐point temperature of the members starts to rise. This paper presents an empirically based model of charring rates for Spruce‐Pine‐Fir (SPF) Machine‐Stress‐Rated (MSR) 2×4 lumber subjected to a constant‐temperature exposure of 500°C, on the basis of test results on 55 specimens. In order that the model can be used with reliability analysis, one of the two model parameters was treated as a Lognormal random variable to explain variations observed in the charring rates. Furthermore, the model has been extended to permit evaluations for other lumber sizes and under exposures such as ASTM E119 fire conditions. The model was compared with existing models and used to predict char data found in the literature. Copyright © 1999 John Wiley & Sons, Ltd.

Summary Recently, measured flame spread indices on commercial wood species tested per ASTM E84 were found to be lower than previously published data. One reason for this may be that the hygrothermal conditioning of the red oak calibrant required by the test standards for measuring flame spread was changed between 1973 and 1981. This paper examines how much variability there is in the moisture content of commercially important wood species at 50% relative humidity by collecting water vapor sorption isotherms. Additionally, the effect of moisture content on the flame spread was evaluated after conducting 14 tests with eastern white pine in accordance with CAN/ULC‐S102 and four in accordance with ASTM E84 at four commercial test laboratories. For the sorption isotherms, it was found that the moisture contents at 50% relative humidity ranged from 6.8% to 11.4% moisture content and depended on the species and whether the specimens had been conditioned in absorption or desorption. The flame spread indices, as measured as different laboratories, also varied from 37% at 10.4% reported moisture content to 200% at 6.5% reported moisture content. The findings suggest that the wood moisture content and conditioning requirements of the standards are important test variables that affect the flame spread results.

No abstract