Optimal versus Nonoptimal Flapping Dynamics for Thrust Production

Whilst the usage of flexible propulsors is ubiquitous in nature, the significance of the resultant fluid-structure interaction is not yet well-defined. Previous research has focused on the relative efficiencies of a solid and uniformly flexible propulsor, showing that when optimized for a side-force, the flexible fin converges to the same trajectory as the solid fin but with a lower geometric efficiency. This work aims to compare the vortex dynamics created by optimal and nonoptimal fin trajectories that are instead derived from optimization of thrust production. To determine optimal and nonoptimal trajectories, a robotic fin mounted to a spherical parallel manipulator (SPM) performs experimental function evaluations and converges on an optimal trajectory through use of a covariance matrix adaptation evolutionary strategy (CMA-ES). The loss function, a measure of optimality, is used to select the optimal trajectory and certain nonoptimal cases. Finally, digital particle image velocimetry (DPIV) is used to illuminate the underlying vortex dynamics that distinguish optimality.