Active control assessments towards optimizing the performance of a cycloidal rotor at hover

Abstract

The present study demonstrates an improved performance of cycloidal rotors by actively controlling the pitching oscillations and rotational speeds. The computational fluid dynamics (CFD) coupled with artificial neural network (ANN) were the methodologies used in the optimization analysis for the hover-state operation rather than the take-off mode under ground effects [1]. The former is carried out to obtain numerical predictions at various operating conditions for an UAV-scale cyclorotor. The oscillating-rotating blades and the corresponding flowfield is computed unsteadily along the complete circular trace for performance considerations. From CFD simulations, the optimum operational state is predicted for a 30∘ and 500 (rpm) pitch angle and rotating speed, respectively. On a second step, by training the ANN with the CFD database at various operating conditions and parameters, the ANN was then capable of analyzing the optimum states for operating at different conditions. The pitching oscillation schedule is then optimized for each rotational speed by using ANN and for each azimuthal location over the traversing trace. This will imply to perform on-board control in active mode for the blades, rather than assigning constant pitching oscillations for all operating states. This active control concept showed to be a potential approach to enhance the cyclorotor efficiency by 12 percent in average.