Identification of microstructural descriptors characterizing the micro-behavior of heterogeneous random fibrous media
Résumé
This work is concerned with the multiscale prediction of the transport properties associated with thermocompressed
materials as recycled cotton felts bonded with petro-sourced fibers (Co-PET/PET). First, a geometric characterization is performed on the studied sample using scanning electron microscopy to identify the main microstructural descriptors (fiber angular orientation, fiber diameter polydiversity). Second, two representative volume elements (RVEs) of the sample are built: one for estimating the low-frequency transport parameters and one for estimating the the high-frequency transport parameters. Each RVE is built with rectilinear fibers parameterized by the probability density function of the fiber orientation
and an appropriate weighted diameter. For the low-frequency RVE, a volume-weighted mean diameter is used, and an inverse volume-weighted mean diameter is used for the high-frequency RVE. These two RVEs make it possible to estimate the transport parameters in low and high frequency asymptotic behaviors using numerical homogenization methods. Finally, the estimated
transport parameters are successfully compared to experimental measurements. The results demonstrate the role of
the diameter polydispersity on the transport properties of random fibrous structures.