M2 Internship – Ultrasonic waves propagation in thick, curved, and semi-immersed composite places: from numerical analysis to physically informed machine learning

ANR SMATSH project
Transition towards sustainable transport and more reactive and stealthy military vessels cannot be achieved without composite materials. However the use of Maritime Composites (MCs) for civilian ships or military vessels requires robust and efficient in-situ waves-based Structural Health Monitoring (SHM) algorithms to monitor them throughout their lifetime in a harsh environment. Unfortunately, waves based SHM algorithms for MCs are not yet available in a reliable and robust way. A first keylock is that MCs are complex (thick stacks, curved surfaces, salt immersion, and pressurization) and waves propagation within them is not well understood. Secondly, optimized transducers networks need to be designed given MCs specificities by combining promising technologies such as Flexible Polymer-Based Transducers (FPBT) and Fiber Bragg Gratings (FBG). Finally, associated waves-based SHM algorithms should be both data and physics driven in order to be as robust and as transferable as possible. The ANR funded SMATSH project thus aims to overcome these key limitations and to demonstrate that given a physical understanding of waves propagation in MCs, powerful dedicated French SHM solutions can be developed using FPBTs and FBGs coupled with physically informed (PI) machine learning algorithms. SMATSH scientific objectives are then to develop computationally efficient models to predict waves propagation in MCs, to optimize FBPTs & FBGs transducers networks for SHM purposes, and to design PI-SHM algorithms for MCs. The results will be illustrated by means of a semi-immersed smart ship hull part performing real-time SHM. SMATSH is structured around Naval Group (the European leader in naval defence) which will provide specifications, industrial demonstrator and validation as an end-user, PIMM laboratory (expert in wave-related problems and SHM) that will develop waves models and PI-SHM algorithms, and ARKEMA Piezotech (a spinoff of ARKEMA able to manufacture specific electroactive polymers) which will design and produce FPBTs.

More information on the link below.