Surrogate-based multi-disciplinary design optimization of passive morphing aeroelastic composite unmanned aerial vehicle wings

Abstract

This paper provides a multidisciplinary design optimization (MDO) approach for optimizing the performance of passive morphing composite wings in unmanned aerial vehicles (UAVs). The objectives considered in this study include minimization of the drag coefficient (C D ), structural weight(m), and root bending moment () and maximi - zation of passive aeroelastic twisting. Neural network-based surrogate modeling was adopted to develop a computationally efficient aerodynamic and structural response prediction tool. The surrogate model was incorporated into the multi-objective NSGA-II optimization algorithm, where composite laminate stacking sequences were treated as design variables in order to take advantage of passive bend-twist coupling. Excellent predictive capabilities of the surrogate model were confirmed in this study, with drag coefficient errors less than ±0.002 and passive twist angle prediction errors smaller than ±0.5°. The Pareto-optimal solutions obtained had a drag coefficient in the range of 0.028 to 0.030, structural weights between 4.1 and 4.5 kg, and root bending moments between 226 and 267 N·m. Passive twist angles ranged from 0° to 8°.