Taxa de carbonização de madeiras de espécies folhosas e características mecânicas do angelim-vermelho em altas temperaturas

Abstract

For the structural fire design of timber structures, two parameters are considered basic: the charring rate and the depth of layer with assumed zero strength and stiffness. Both parameters contribute to the determination of the effective charring depth of a given structural timber element. Samples of thirteen hardwood species with specific gravity between 500 and 1300 kg/m3, at 12% moisture content, were subjected to carbonization with exposure to the ISO 834 standard fire. It was found that wood with the same specific gravity may present different charring rates, depending on the type of obstruction in the vessell. This finding guided the grouping of species according to the characteristics of obstructions in the vessels, being: cluster 1, for species that have all or some vessels obstructed by deposits associated or not with tylosis, and cluster 2, for species that have obstructed vessels only for tylosis. Models for prediction of charring depth were determined for each cluster and, by derivation, models for prediction of charring rate. The model developed for cluster 1 presents, as predictor variables, the specific gravity and the time of exposure to fire while the model developed for cluster 2, presents only the time of exposure to fire as a predictor variable. Models for temperature gradients within the specimens were proposed for both groups. It was found that for periods of 20 and 30 min under exposure to standard fire, temperature stabilization occurs at a depth of 50 mm. Exploratory calorimetry tests were also carried out to verify the thermal stability of the wood of the species angelim-vermelho (Dinizia excelsa Ducke) as well as the evaluation of the influence of temperature on the mechanical characteristics of the same species, in the range of 40 to 240 ° C. As a consequence, factors for modifying the mechanical properties of Dinizia excelsa Ducke in hot and cool conditions are proposed. Finally, by correlating the temperature gradient model developed for cluster 1 with the second average glass transition temperature, determined for the species Dinizia excelsa Ducke, the depth of layer with zero strength was calculated, confirming the value adopted in the technical standards.