The optimal stiffnesses and spatial locations of contact-aided compliant mechanisms in a dynamic flapping wing structure are found using a numerical dynamics model and multi-objective genetic algorithm. Using mathematical descriptions of how wing shape change impact pitch agility in a flapping wing mechanical bird, an optimization problem was formulated to find compliant joint stiffness and location parameters which induce desired shape change. Specifically, the goal of the shape change was to induce forward sweep at the upstroke to downstroke transition while otherwise remaining stiff in an effort to move the aerodynamic center ahead of the center of gravity. A single compliant joint in the leading edge spar of an ornithopter wing was considered. A multi-objective genetic algorithm was used to solve the optimization problem, generating 3892 unique designs over 20 generations. Machine learning visualization and regression was used to better understand the data set. The data set was narrowed using higher level decisions, and one optimal design which satisfied the design requirements was chosen based on its relative performance and design parameters. The design was able to achieve the desired forward sweep while only allowing small bending and twist motion compared to the wing structure without a compliant joint inserted.
Local Search Based Enhanced Multi-objective Genetic Algorithm of Training Backpropagation Neural Network for Breast Cancer Diagnosis