'Experimental Optimization by Genetic Algorithm for Flow Separation Control with Surface Plasma Actuator

TL;DRAbstract

The present investigation concerns the active control of a turbulent separated flow downstream of a backward-facing step by surface plasma discharge. A single-objective genetic algorithm is implemented in order to achieve the minimization of the recirculation length. For that purpose, a series of unsteady pressure sensors installed on the bottom wall can detect the mean reattachment location. The optimized variables are the voltage amplitude, burst frequency and duty-cycle of the applied signal. Here, single-objective evolutionary algorithm, usually coupled to computational fluid dynamics simulation, is coupled with real-time experimental data for the first time. It is shown that the genetic algorithm is successful at finding the optimum forcing conditions for a turbulent flow at Re h=30000 (Re θ=1650). Then the optimal conditions are explored by time-resolved PIV in order to give physical explanations of the best input signal identified from the evolutionary algorithm. In particular,

Chat with Paper

AI Agents for this Paper