Optimal structural design for a certain near-space composite propeller of airship using adaptive region division blending model

Near-space airship is a frontier and hotspot in current military research and development, and the near-space composite propeller is the key technology for its development. In order to obtain higher aerodynamic efficiency at an altitude of 22 km, a certain near-space composite propeller is designed as a long and slender aerodynamic shape with a 10 m diameter, which brings many challenges to the composite structure design. The initial design is obtained by the composite structure variable stiffness design method using based on fixed region division blending model. However, it weighs 23.142 kg, exceeding the required 20 kg. In order to meet the structural design requirements of the propeller, a variable stiffness design method using the adaptive region division blending model is proposed in this paper. Compared with the methods using the fixed region division blending model, this method optimizes region division, stacking thickness and stacking sequence in a single level, considering the coupling effect among them. Through a more refined region division, this method can provide a more optimal design for composite tapered structures. Additionally, to improve the efficiency of optimization subjected to manufacturing constraints, a hierarchical penalty function is proposed to quickly filter out the solutions that do not meet manufacturing constraints. The above methods combined with a Genetic Algorithm (GA) using specific encoding are adopted to optimize the near-space composite propeller. The optimal design of the structure weighs 18.831 kg, with all manufacturing constraints and all structural response constraints being satisfied. Compared with the initial design, the optimal design has a more refined region division, and achieves a weight reduction of 18.6%. This demonstrates that a refined region division can significantly improve the mechanical performance of the composite tapered structure.