Adaptive composite materials with machine learning and 4D printing – REDeSIGN4D

The development of the industry of the future is based on innovation vectors such as eco-design, digital engineering, artificial intelligence and new materials obtained through innovative processes. The synergies created by the interaction between these different technologies are poorly understood and represent a disruptive potential which is the base for REDeSIGN 4D.
As part of défi 8 «Domaines transversaux», axe 13 « Industrie et usine du futur : Homme, organisation, technologies», REDeSIGN 4D explores technological bricks to create new adaptive bio-inspired structures with a controlled environmental impact. They will be eco-designed and produced from renewable and local resources, combined with the use of innovative 4D printing, functionalization, multi-physical digital simulation and machine learning processes.
Labeled by the EMC2 cluster, we have chosen as a case study hygromorphic biocomposites made from flax fibers, targeting promising benefits in several industrial sectors in demand (building, defense, energy, etc.). A second innovative axis lies in the systematic Life Cycle Analysis (LCA) to assess the environmental impacts of the proposed innovations.
Inspired by functional biological structures such as the pine cone, hygromorphic biocomposites break with the current paradigm which consists in fighting against the hygroscopic properties of plant fibers such as flax rather than seeking to benefit from them. These are architectural materials, both sensors and actuators that change shape (actuation or morphing) autonomously in the presence of a humidity gradient. At the same time, the advent of 4D printing, an extension of 3D printing dedicated to stimulable materials, opens up a field of possibilities in terms of the architecture of hygromorphic biocomposites.
The state of the art highlights 4 obstacles that REDeSIGN 4D will remove: 1) the mastery of the relationships between the parameters of the 4D printing process, 2) the slowness of the stimulated response and 3) the controlled prescription of the movements of the hygromorphic biocomposites and finally 4) the predictability of their multi-scale and multi-physical behavior.
REDeSIGN 4D is divided into four operational Work-Packages (WP1-4) corresponding to the 4 scientific and technical obstacles to be removed. The scope of skills is covered by the three partner laboratories (IRDL, INRAe BIA and PIMM) while relying on the recruitment of 3 doctoral students and one post-doctoral fellow. The environmental issues will be assessed throughout the 48 months by setting up an LCA approach, transversal to the WP.
First, at the mesoscopic scale we will apply statistical learning methodologies such as neural network (machine learning) to understand the effect of the process on the morphing capacity of hygromorphic biocomposites (WP1 months 0-36). At the same time, REDeSIGN 4D will propose on a microscopic scale an original functionalization of flax fibers to make them electrically active, in order to generate functionalized biocomposites whose potential for hygromorphism will be greater while benefiting moreover from the control of morphing by Joule effect (WP2 months 0-36). Based on this knowledge, using parametric and topological optimization processes, REDeSIGN 4D will propose, at the mesoscopic scale, architectural configurations of an optimized hygromorphic biocoposite fold allowing optimal and amplified morphing (WP3 months 0-36).
The last stage (WP4 months 30-48) makes the link between the work carried out in WPs 1, 2 and 3, via the realization of a structure known as "proof of concept" allowing to amplify, by structural effects, the morphing proposed by the hygromorphic biocomposite previously functionalized. We are therefore here at the largest scale, assembling the scales of the previous WPs.