Evolutionary Optimization of a Morphing Wing with Wind Tunnel Hardware-in-the-Loop

Active wing morphing is a fertile technology in the realm of unmanned and micro air vehicles: low flight loads and flexible wing materials allow substantial deformation with relatively low actuation power. Proper shape management of these wings for maximum performance is challenging: numerical optimization methods may struggle to accurately predict the three-dimensional flow separation, transition, and reattachment expected at the low operating Reynolds numbers. Conventional experimental optimization techniques (a response surface fit through a designed experiment) may be unable to represent the discontinuous objective functions commonly seen in aerodynamic applications, and the number of function evaluations can become overwhelming with high-fidelity actuation mechanisms. This paper outlines an effective experimental technique for optimizing a morphing airfoil: a genetic algorithm with wind tunnel hardware (strain gage sting balance) in the loop. Optimal wing shapes are found to maximize the measured lift or efficiency through a sweep of angles of attack, for a wing with a single actuation point (camber control) and two actuation points (camber and reflex control). Aerodynamic and electrical hysteresis limits the repeatability of sting balance measurements, complicating the convergence of the genetic algorithm, but not preventing adequate location of optimal designs.