Optimization of the Ratio of Induced Flow Velocity to Electrical Power Consumption in the DBD Plasma Actuator Using Artificial Neural Network and Genetic Algorithm

Dielectric barrier discharge (DBD) plasma actuators are one of the new devices for active flow control, which has received substantial attention during the last decade. The performance of the actuator is optimum when it induces the highest velocity per unit of power consumption. Since the induced velocity and the power consumption of the actuator depend on many different variables, finding the optimal set, which results in the best performance, is of immense importance. In this paper, in order to optimize the performance of these actuators, at first, by using full factorial design of experiments the effect of electrical variables (including voltage and frequency) and geometrical variables (including the gap between electrodes, dielectric thickness, and covered electrode width) on induced flow velocity and power consumption in steady actuation is experimentally investigated. Then, by using the multi-layer perceptron neural network, a model is created for the ratio of induced velocity to power consumption. The model is validated both statistically and experimentally. The results indicate that the coefficient of determination for training and test data is higher than 95 percent. Finally, the surrogate model is optimized by genetic algorithm and the optimal value of electrical and geometrical variables is determined. In order to validate the result, an actuator is designed based on the optimal set of variables and it’s ratio of velocity to power is measured to be 29.71 (m/s)/(kW/m). The difference of 3 percent between the measured and the predicted value demonstrates high accuracy and correctness of the proposed model and method.