Abstract
General linear regression models were used to determine the relationship between thermal conductivity and specific air permeability of 48 heterogeneous, needlepunched, nonwoven samples that were made from ceramic and glass webs. Parameters analyzed included number of needle barbs, fabric weight, thickness, and porosity. Other factors considered were fabric layering structure, temperature drop across the fabric and specific air permeability. Of the linear regression models examined, three models were found to be significant at greater than 95% confidence. These models had r2 values of greater than 97%. Factors that proved to be greater than 95% significant in predicting the effective thermal conductivity of the samples tested were fabric thickness and weight, fabric porosity, fabric mean pore size, and specific air permeability.