Structural Optimization for Automated Fiber Placement

Abstract

Structural optimization has long been used with isotropic materials to provide a preliminary design of components for revision toward a final part. Generative solutions closer to this final state can be obtained by implementing additional constraints in the optimization process. These constraints can be manufacturing, functional, or otherwise. For example, in a printed structure, a manufacturing constraint might limit the minimum member cross-section in the print plane. Implementing constraints for use with fiber-resin composite materials has been a more recent development, and an active area of research. Fiber-composite laminates manufactured using the process of Automated Fiber Placement (AFP) have their own unique manufacturing requirements which existing commercial optimization suites do not adequately address. This results in optimal” ply shapes which cannot be realized. This thesis describes and demonstrates a manufacturing constraint within an existing optimization process, such that the modified process results in plies which meet the minimum tow length constraint of AFP. This novel constraint was then demonstrated on a test panel. Designed using T800/3900 carbon fiber prepreg, it features stress concentrations in the form of rounded corners and a 787-window cutout. The performance characteristics of the modified and current state-of-art panels were evaluated using finite element models. Finally, the modified ply boundaries were interpreted via AFP CAM software to validate their manufacturability. This demonstration resulted in a panel that was 8g (1.6%) heavier than the standard optimization process with 1% increases to stiffness and strength. While still maintaining a 40% mass reduction over a comparable strength “black aluminum” panel. Crucially, this panel also has plies which can be accurately manufactured via AFP directly from the optimization output, as verified via CAM software. This makes the design process easier and faster.