Modeling the response of blended PVC/PET fabrics to fire environments

Abstract

The fire behavior of fabrics with uncertain composition of polyvinyl chloride and polyethylene terephthalate blends presents numerous kinetic and flame spread modeling challenges. Such fabrics have extensive usage in light-weight shelters or for protective-coverings. This study details thermogravimetric analysis and one-sided heating experiments carried out for three different PVC/PET blended fabrics. A challenge with characterizing these materials is the relative contributions of PVC and PET are unknown. Only approximate ranges are often provided by manufacturers. In addition, the char/species yields are not known since blended materials (with additives) perform very differently from homogeneous PVC or PET samples. To unravel initial mass fractions and char/species yields, a new non-linear regression analysis is developed to process TGA mass loss data. This analysis is verified by Scanning Electron Microscopy/Energy Dispersive X-ray Spectroscopy to confirm relative amounts. The output of the regression analysis provides important kinetic rate parameters for Arrhenius modeling used in FEA. To verify this approach, FE model predictions are compared to experiments showing good agreement to measured fabric temperature and mass loss. These results confirm the utility of using non-linear regression analysis for providing kinetic and composition information for constructing thermal degradation models of blended PVC/PET fabrics subject to fire environments.