Optimization of imperfection-insensitive continuous tow sheared rocket launch structures

Geometric imperfection sensitivity is the largest influencing factor that limits the design of thin-walled monocoque cylinders. Current generation cylindrical architectures, such as those found in rocket launch vehicles, rely on the use of sandwich or blade-stiffened structures to reduce the imperfection sensitivity of the cylinder. Whilst much research has focused on the creation of new knockdown factors that relate to the modern architectures used, this paper focuses on reducing the imperfection sensitivity of a monocoque cylinder from a design perspective. Variable-angle composites offer an opportunity to design the load paths of structures, thus reducing the effective area over which imperfections initiate buckling. Continuous Tow Shearing (CTS) is one such variable-angle manufacturing technique. It does not cause common in-process manufacturing defects associated with Automated fiber Placement such as fiber wrinkling or fiber buckling. In addition, there is a shearing angle-thickness coupling that results in a local thickness build-up, which, whilst increasing the mass of the structure, enables embedded stiffeners to be created by shearing the tow. Three genetic algorithm (GA) optimizations are carried out to maximize the imperfect mass-specific buckling load to investigate the efficacy of CTS and tow-steered designs in reducing imperfection sensitivity. The first optimization considers idealistic manufacturing capabilities with a random geometric imperfection and results show that whilst the imperfection sensitivity has decreased considerably, the GA-optimum result does not have general imperfection insensitivity. The second and third optimizations consider current manufacturing capabilities and are compared against one another to analyze the use of a evolutionary hybrid GA and a probabilistic, reliability-based GA. In all three optimizations, the GA-optimum laminate demonstrates imperfection insensitivity.