GDR CNRS 2024
Multiscale Mechanics of fibrous media

The next MECAFIB days will focus on the variability in fibrous media:
 

• At the fiber scale, morphological characteristics and mechanical properties of fibres vary drastically. This is the case, for examples, of  carbon or glass fibre bundles, biosourced fibres, collagen, elastin, cellulose fibrils...
• At the scale of the representative elementary volume (if it exists), the microstructural and mechanical properties of fibrous media are also dispersed whether the fibrous media are quasi-ordered (woven, braided , knitted, cables, bundles ...) or very disordered (nonwoven, mats, papers, filters ...).
• At a larger scale, i.e., that of the fibrous macrostructure, undesirable property gradients may also be observed due to bad control of processing conditions.

Thus, whatever the considered scale, assessing this dispersion and evaluating its impact on the mechanical properties of fibrous materials raise particularly difficult experimental and modeling problems. These points will be adressed during this workshop. For example, associating and quantifying pertinent topological and geometric descriptors to fibrous microstructures and their variability constitutes a major issue. Developping relevant experimental procedures to characterise the mechanics of fibrous material is also very important within this context. Accounting for the propagation of this variability on the macroscopic properties of fibrous media using dedicated modelling techniques constitutes another bottleneck.

To help us to tackle the aformementionned difficulties, the worshop will begin with three lectures proposed by experts in the field :

• Dominique Jeulin from Mines ParisTech (Centre de Morphologie Mathématique, Fontainebleau), will talk about (i) morphological modelling in random fibrous materials, (ii) a statistical approach to determine Representative Elementary Volumes in these media, (iii) statistical models for the failure in fibrous materials,
• Maurice Lemaire from Univ. Clermont Auvergne (Institut Pascal), will emphasise that every time a numerical model is implemented, questions about the uncertainties about data and the model must be raised,
• Christian Soize from Univ. Paris Est (laboratoire de Modélisation et Simulation Multi-Echelle) will talk about stochastic modelling of fields, inverse statistical identification, probabilistic learning on varieties, and design optimization under uncertainties in numerical mechanics.

This meeting will continue with oral presentations of your works mainly focusing on this theme, whether the approach is experimental, theoretical or numerical or a mixture of three.