Optimizing reduced frequency using genetic algorithms for plasma flow control to achieve drag reduction on a circular cylinder

A closed-loop parameter optimization system around a cylinder is built by integrating the plasma actuation and genetic algorithms in this research, employing numerical simulations and experimental methods. The study aims to minimize the total drag on the cylinder by optimizing the reduced frequency. A pair of surface dielectric barrier discharge plasma actuators, powered by alternating-current high-voltage sources, is symmetrically positioned at ±90° azimuth angles on the two sides of a circular cylinder, and the Reynolds (Re) number is 1.5×104 based on the cylinder diameter. Numerical simulations were first used to determine the optimization space for the reduced frequency, followed by wind tunnel experiments to further search for the optimal research within this space. Particle image velocimetry and hot-wire anemometry were used to investigate the flow field's instantaneous and time-averaged characteristics. Ultimately, the optimal reduced frequency was identified based on duty-cycle frequency, free-stream velocity, and cylinder diameter. The results show that the optimal duty-cycle frequency obtained through genetic algorithm optimization in numerical simulations and wind tunnel experiments is the same, at 140 Hz, corresponding to a reduced frequency of approximately 1.372. The drag reduction rates are also similar, at 73.9% and 73.6%, respectively. During plasma flow control with the optimal reduced frequency, the dominant frequency of the overall motion of the separated vortex field is no longer the natural shedding frequency of the baseline flow. Still, it is instead controlled by the plasma duty-cycle frequency. Compared to the baseline flow, the plasma flow control at the optimal reduced frequency transforms the large-scale alternating vortices into small-scale shedding vortices, resulting in a time-averaged narrow and stable velocity deficit region, leading to reduced energy loss and significantly lower time-averaged drag coefficient. Meanwhile, the interaction between plasma-induced vortices and the Kármán vortex street in the cylinder wake enhances mixing, significantly suppressing turbulence intensity. The results demonstrate the effectiveness of genetic algorithms in identifying the global optimal reduced frequency of plasma actuation, achieving maximum drag reduction.