Êtes-vous un étudiant de l'EPFL à la recherche d'un projet de semestre?
Travaillez avec nous sur des projets en science des données et en visualisation, et déployez votre projet sous forme d'application sur Graph Search.
Continuous fiber thermoplastic composites are potential substitutes of steel in structural automotive components, since they combine fiber strength and rigidity with ductility, potential recyclability and low density of the thermoplastic (TP) matrix. Resin transfer molding (RTM) is a process for composite production, involving direct impregnation of a reinforcing fabric in a closed rigid mold with a liquid resin. The duration of the impregnation step is directly proportional to the resin viscosity and inversely proportional to the fabric in-plane permeability. As a result, the high viscosity of melt thermoplastic resins (mTP) generally precludes their use in RTM, in particular to reach production volumes in line with the automotive sector. However, recent development of high-fluidity (HF) TPs, with medium range viscosity (10-50 Pa s) opens new possibilities for mTP-RTM. The goal of this thesis is to explore strategies to enhance the impregnation kinetics in mTP-RTM with HFTP by further increasing the fabric in-plane permeability. The effect of modified fabric architectures was investigated through experiments allowing visualization of flow propagation and permeability measurement using model fluids on a modified non-crimp fabric. A saturated permeability of ca 10-9 m2 for 46 vol% fibers was measured. The mesoscopic pore-space of the compacted fabrics was imaged with X-ray Tomography, and analyzed to propose permeability predictions based on the large channels geometry, which agreed well with the experimental results. Capillary effects were assessed over a broad range of capillary numbers (from 4.10-5 to 0.4) by direct comparison of unsaturated and saturated permeability. Permeability in these dual-scale fabrics is governed by viscous flow in the meso-channels. As a result, provided that the capillary number exceeds a threshold value, their permeability can be rather accurately measured by carrying out unsaturated measurements, neglecting micro-flow and capillary effects. Further improvement up to 10-8 m2 for 46 vol% fibers was achieved by introducing a flow-enhancing structure, made by polymer 3D-printing in the core of a woven fabric stack. The resulting flow was markedly dual-scale, going first very quickly in the core and later through the fabric. Optimized impregnation time was obtained by pulling vacuum prior to injection, letting the fluid flow within the core to the outlet, which is then closed, and finally continue the injection through thickness to complete saturation. A proof of concept with epoxy resin showed that a poly(latic acid) spacer led to an increased bending stiffness of 20%, but strong reinforcement gradients. Using a sacrificial spacer made of polycaprolactone, which melted during resin cure (80°C) allowed full fabric relaxation and a more homogeneous fibers distribution. Finally, these findings were applied to the development of a mTP-RTM mold to produce plates with HF polyamide 6 matrix. An 11 cm-long plate with 46 vol% glass fibers was impregnated in 7 min (1 min in the core, 6 min through the fabric) using high temperature spacers compared to more than 45 min without, also with improved bending stiffness by 26%. This concept is conceivably scalable to larger parts, since the bottleneck of the whole process is transferred to the saturation step. A preliminary cost analysis was carried out to evaluate the scalability of this technique as compared to alternative compression TP-RTM or a baseline epoxy RTM process.
François Gallaire, Edouard Boujo, Yves-Marie François Ducimetière
Véronique Michaud, Baris Çaglar, Helena Luisa Teixido Pedarros