Surrogate Modeling for Optimizing the Wing Design of a Hawk Moth Inspired Flapping-Wing Micro Air Vehicle

Abstract

Proving the feasibility and overall efficiency of Flapping-Wing Micro Air Vehicles (FWMAVs) over other types of MAVs is vital for their advancement. Due to their complex aerodynamics and the difficulty of building accurate models of the flying animal, assessing the flight performance and efficiency of animals and FWMAVs mimicking those animals can be a challenging task. This paper investigates the hawk moth (Manduca sexta L.) forewing as inspiration for designing an optimal wing for a moth-scale FWMAV. Through a process of decoupling the flapping-wing kinematics from the aerodynamics, an experiment is designed to assess the variation in aerodynamic lift-to-drag ratio due to variations in the wing geometry parameters (i.e., wingspan, chord length, cross sectional geometry). Using the data from the experiments, a surrogate model is trained and serves as an optimization solution for determining a wing geometry configuration that maximizes the lift-to-drag ratio. The resulting trained surrogate model is a computationally inexpensive model that can rapidly evaluate the aerodynamic efficiency based on the wing geometry input parameters, thus identifying local extrema within the design space.