Search

Lux06b; International audience; A general method combining the volume averaging technique and image analysis is proposed to determine the effective thermal conductivity tensor of real fibrous materials featuring local anisotropic thermal properties. The application of mathematical morphology tools on 3D images of wood based fibrous insulators allows a thorough investigation of the microstructure of these materials. A representative elementary volume is determined and the geometrical structure and local anisotropy are studied and quantified. The classical closure problem coming from the one equation model is solved on the 3D thermal conductivity tensor field and the effective thermal conductivity is computed. Good agreement with available experimental data is achieved.

The microstructure of fibrous networks in low density wood-based fibreboards elaborated with various processes is observed by X-ray tomography. Image analysis tools derived from mathematical morphology are used to process 3D images obtained and to recover 3D information on the fibres and the fibre network. A 3D model of random fibre networks developed by means of experimental information is presented. The parameters of the model correspond to internal architecture properties. Their influence on a local thermal conductivity is studied through finite element method.