Flow control with synthetic jets on two tandem airfoils using machine learning

Active flow control was applied to a tandem configuration of two SD7003 airfoils. The tandem configuration consisted of an upstream airfoil (forefoil) with a pitching motion at a fixed frequency and a downstream airfoil (hindfoil) that was not moving. Synthetic jet actuators (SJAs) were applied on both airfoils to control the flow fields at the low Reynolds number of 30 000. The flow physics inherently involved three different frequencies: frequency of the pitching forefoil and two actuation frequencies of the two of SJAs. In this study, we kept all three frequencies fixed at 5 Hz. However, we allowed for phase differences between them. An optimization study was conducted in order to improve total aerodynamic performance defined as the combined total time-averaged value of lift-to-drag ratio of both airfoils (L/D) tot Injection angle of the two SJAs, phase differences between each SJA frequency, and frequency of the pitching motion in addition to vertical spacing between the airfoils were considered as design variables of the optimization study. Optimization algorithm was coupled with a machine learning method to reduce computational cost. We found that lift coefficients were enhanced, and drag coefficients were reduced for the optimum controlled case in comparison with the uncontrolled case, which led to an aerodynamic performance improvement of 304%. However, drag force was the dominant parameter in determining final performance value. For all design variables, drag force determined the final optimum value