The long-term goal of this ONR funded project is to facilitate the design of architected composites that play a key role in damage tolerant and resilient structures. The main emphasis is on developing new composite structures with improved performance and durability as compared to conventional structural composites. To that end, we will present our work in detail on the following within the realm of sandwich composites along with a novel Machine Learning framework for stress prediction in composites: 1) Novel recoverable sandwich composite structures: Traditional sandwich cores such as foam core or honeycomb structures are good options for enabling lightweight and stiff structures. Although, these cores are known to dissipate energy under extreme conditions such as impact loading, they experience permanent damage. Here, our goal is to design core structures that undergo substantial deformation without accumulating damage and recover their original geometric configuration after the loading is removed. In contrast to a traditional foam or honeycomb structure, we have developed a multi-layer architected core design that facilitates significant deformation beyond the initial peak load, yielding a larger energy dissipation during impact and other extreme loading scenarios. We utilize the concept of pseudo-bistability of truncated cone unit cells to achieve elastic buckling for energy dissipation and shape recovery of core structures. 2) Tailoring of sandwich composite facings: Our objective is to establish the influence of fiber architecture on moisture diffusion pathways in FRPC facings for enabling damage tolerant facing designs. To that end, we have evaluated the moisture kinetics in FRPCs by developing micromechanics based computational models within FEM. We have explained the effect of tortuous diffusion pathways that manifest within FRPCs due to internal fiber architectures. Finally, we established the relationship between tortuosity and diffusivity that can be used for studying moisture diffusion in other FRPCs.
