CONTINUOUS FIBER-REINFORCED COMPOSITE STRUCTURE DESIGN USING THE ML BASED DESIGN METHOD
Abstract
Recently, researches have been actively conducted to find fiber-reinforced composite structures that improve structural rigidity by printing textile materials on existing structures. However, most of those studies focus on the manufacturing feasibility such as the printability and toolpath, or the mechanical performances such as the structural stiffness. Therefore, studies considering both performances are rare. This research aims to derive functionally graded structures which show both the macro-scale topological configuration and the micro-scale unit structures. Instead of using the asymptotic homogenization theory (AHT) to obtain composite material properties, machine learning (ML)-based design method will be used for the micro-scale design for the reinforcement path because it has more advantages over the AHT-based method. Based on the design considering the mechanical stiffness under static load, the printing path for the fiber reinforcement will be determined based on micro-scale design using such as de-homogenization for manufacturing. In the first year, macro- and micro-scale design on statically loaded, single-layer solid-state models will be conducted to maximize the structural stiffness while it will be extended to design on solid-state models consisting of multi-layers under static load. The design process along with the design results will be offered to AOARD-AFRL, and the conceptual models and related material properties will be provided by AOARD-AFRL. However, the source code of the design process will not be provided to AOARD-AFRL in this research.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Jan 04, 2023
- Source ID
- FA23862214038
Entities
People
- Jeonghoon Yoo
Organizations
- Air Force Office of Scientific Research
- United States Air Force
- Yonsei University