- Program Areas
- Installation Energy and Water
- Environmental Restoration
- Munitions Response
- Resource Conservation and Resiliency
- Weapons Systems and Platforms
- Energetic Materials and Munitions
- Noise and Emissions
- Surface Engineering and Structural Materials
- Fuels and Greenhouse Gases
- Lead-Free Electronics
- Waste Reduction and Treatment in DoD Operations
Optimization of Self-healing Fiber-reinforced Polymer-matrix Composites via Convolutional Neural Networks
Jason Patrick | North Carolina State University
Internal delamination damage in fiber-reinforced composites is difficult to detect and nearly impossible to repair by conventional methods. To date, this failure mechanism remains one of the most significant factors limiting the reliability and leads to wasteful design/replacement of composites in lightweight structures. Drawing upon inspiration from biology, self-healing polymers and composites have emerged to combat inevitable degradation from in-service operation and/or unavoidable damage from unexpected overload. The project team plans to develop a sustainable self-healing composite system capable of complete restoration in interlaminar fracture resistance without compromising in-plane mechanical properties.
This project takes a collaborative, interdisciplinary approach by combining polymer mechanics/chemistry, emergent manufacturing, advanced computing and deep learning to accelerate the development of such self-repairing structural composites. This synergistic experimental-computational strategy relies on: (i) a newly realized in situ self-healing platform in epoxy-matrix fiber-composites via thermal remending of an environmentally inert three-dimensional (3D) printed thermoplastic interlayer; and (ii) novel microstructural material optimization using automated finite element simulations and deep learning via convolutional neural networks. A visual summary of the research project is presented in Figure 1.
Figure 1: Project concept - (left) scanning electron microscopy image of 3D printed thermoplastic (blue) on woven glass; (middle) Self-healing thermal remending via in situ resistive heating; (right) Material microstructure optimization via auto-modeling and deep learning.
The envisioned self-repairing composite and the complementary computational design platform will benefit Department of Defense (DoD) and the environment by eliminating the need for costly inspection, reduce overall maintenance and replacement, and enhance structural fiber-composite reliability and performance. The developments have game-changing potential in a variety of defense sectors including aerospace, civil, and naval. In particular, infrastructure in corrosive, erratic, and inaccessible environments would greatly benefit from improved durability and eliminate the need for manual repair intervention. In addition to enhancement of defense assets, critical life protection systems (e.g., body/vehicle/structure armor) could endure higher loadings and increased number of impact/overload events. The unique innovation for this research is an interdisciplinary combination of state-of-the-art experimental, computational, and artificial intelligence techniques with practical composite fabrication pathways to produce a cross-cutting multifunctional composite. The resulting self-healing, structural composites will increase safety, resilience, and reliability. This breakthrough technology will also enhance tactical capabilities for competitive advantage in DoD operations while helping to preserve the as-built environment.