Synthetic nonwoven fibrous layers are widely used in insulation industries such as residential buildings and piping insulation. An accurate estimation of thermal diffusivity of bulk complex fibrous material has not been available in transient conditions by commonly used methods. In this research attempts was made to estimate transient heat transfer rate of synthetic nonwoven barriers by partial differential equation (PDE) in numerical aspect. An intelligent modeling method based on training of a propagation neural network was applied to estimate thermal diffusivity in transient conditions. Nonwoven samples of PAN, PET, PA 66 and PP were prepared in different weights and thicknesses. Temperatures of inner and outer sides of layers were measured in a purpose-designed apparatus with precision thermometers. The intelligent model is validated by experimental data with a very good accuracy. The thermal diffusivity results demonstrated that the PA 66 layers have high insulating resistance with slow heat flow because of their low thermal diffusivities. Also, the thick layers have less insulating ability in which porous layers with more air permeability insulated heat well. The main mechanism of heat transfer in low temperatures is thermal conduction where thick layers with low porosity have more interaction and more conduction between fibers. Keywords : Synthetic Nonwoven, Transient Heat Transfer, Thermal Insulation, Intelligent Modeling
