Machine learning is increasingly used for active flow control. In this experimental study, alternating-current dielectric barrier discharge plasma actuators are deployed for the closed-loop intelligent control of the flow around a cylinder at a Reynolds number of 28 000 based on the velocity feedback from two hot-wire sensors placed in the wake. Variations in the cylinder drag are monitored by a load cell, and the temporal response of the wake flow field is visualized by a high-speed particle image velocimetry system working at 1 kHz. The high-speed control law is operated using a field programmable gate array optimized by genetic programing (GP). The results show that the peak drag reduction achieved by machine learning is of similar magnitude to that of conventional steady actuation (∼15%), while the power saving ratio is 35% higher than with conventional techniques because of the reduced power consumption. Analysis of the best GP control laws shows that the intensity of plasma actuation should be kept at a medium level to maximize the power-saving ratio. When compared with the baseline uncontrolled flow, the best controlled cases constrain the meandering motion of the cylinder wake, resulting in a narrow stabilized velocity deficit zone in the time-averaged sense. According to the results of proper orthogonal decomposition and dynamic mode decomposition, Karman vortex shedding is promoted under the best GP control.
Particle image velocimetry measurement on shock wave/boundary layer interaction in a Mach 2.0 supersonic wind tunnel is performed to quantitatively reveal the plasma flow control effect in this paper. The typical flow structure is produced by a 24-degree compression ramp model and the streamwise plasma actuator array with five pulsed spark discharge plasma actuators is adopted as the control device. In the midspan plane, the results show that although the separation region exhibits an obvious extension, the foot of the separation wave moves upstream and the shock wave angle decreases from 41.6° to 22.3°, proving the decline in shock intensity. The shock wave drag is estimated to be reduced by 45%. According to the phase-averaged velocity field, the reason that the high-frequency actuation plays a key role in achieving the continuous control effect is revealed through the temporal evolution of the separation region area. Also, another interesting phenomenon that the flow deflects when passing through the actuation region is found, which may induce the upwash and downwash motions of the boundary layer and further reduce the flow separation on both sides of the actuation region. At last, a preliminary conceptual model is proposed to reveal the probable flow control mechanism.
Hypersonic boundary layer transition is a hot yet challenging problem restricting the development and breakthrough of hypersonic aerodynamics. In recent years, despite great progress made by wind tunnel experiment, transition mechanism and transition prediction, only partial knowledge has been gained so far. In this paper, firstly, the specific scenarios of hypersonic boundary layer transition control are clarified. Secondly, the experimental research progress and mechanism of passive control and active control methods under different hypersonic transition control demands are summarized, with their advantages and disadvantages being analyzed separately. Plasma actuation is easy to produce controllable broadband aerodynamic actuation, which has potential in the field of boundary layer transition control. Hence, the following part of the paper focuses on plasma flow control. The feasibility of plasma actuation to control the hypersonic boundary layer transition is demonstrated and the research ideas are presented. Finally, hypersonic boundary layer transition control methods are summarized and the direction of future research is prospected.
In this paper, a continuous Q-learning algorithm is deployed to optimize the control strategy of a trailing-edge airfoil flow separation at a chord-based Reynolds number of 2×105. With plasma synthetic jets issued at the middle chord and a hot wire placed in the separated shear layer acting as the actuator and sensor, respectively, a high-speed reinforcement learning control at an interaction frequency of 500 Hz is realized by a field-programmable gate array. The results show that in the Q-learning control, the controller only needs several seconds to elevate the instantaneous reward to a level close to the final mean reward, and convergence of the control law typically takes less than 100 s. Although the relative drag reduction achieved by Q-learning control (10.2%) is only slightly higher than the best open-loop periodical control at F∗=4 (9.6%), the maximum power saving ratio is improved noticeably by 62.5%. Physically, Q-learning control creates more turbulent fluctuations, earning more rewards by increasing the transition possibilities toward high-value states. With increasing penalty strength of plasma actuation, the final control laws obtained from Q-learning exhibit a decreasing number of active states. Detailed comparisons between the open-loop and Q-learning control strategies show that the statistics of the controlled velocity fields remain similar, yet the turbulent fluctuations contributed by the vortex shedding mode are reduced by constant-frequency plasma actuation.
A streamwise array of pulsed spark discharge operating at a high frequency of 10 kHz is used to control the shock wave/boundary layer interaction (SWBLI) induced by a 24° compression ramp (CR) in a Mach 2.0 flow. High-speed schlieren imaging at 50 000 frames/s is deployed for flow visualization. The schlieren snapshots, as well as statistics of the schlieren sequence, show that the intensity weakening effect on the shock wave can be achieved for a long time up to 20 ms, which is the first experimental demonstration of stable control on the CR-SWBLI with plasma actuators. Then, based on fast Fourier transform analyses, the low-frequency unsteadiness of the shock wave is also significantly modified and the dominant oscillation frequency is switched to the high frequency St = 0.32. The mechanism of the encouraging control outcome is finally discussed that the strong separation wave induced by a CR is replaced by a multistage weak compression wave produced through a high-frequency actuator array.