Optimization control of the D-shaped cylinder flow with genetic algorithm and Coanda pulsation jets

Flow around the D-shaped cylinder is one of the basic models in aerospace, civil, and marine engineering applications. This paper reports an experimental work on the optimization control of the D-shaped cylinder flow using genetic algorithm (GA) and Coanda pulsation jets. Experiments are conducted in a wind tunnel at a Reynolds number Re = 2.0 × 10⁴, based on free stream velocity U_∞ and the cylinder height H. The Coanda pulsation jets are realized by issuing pulsation jets over the surface of a one-quarter cylinder (radius r = 0.2H), which is attached on the base and near the trailing edge of the D-shaped cylinder. Control parameters of the Coanda pulsation jets include momentum coefficient (C_μ), nondimensional pulsation frequency (f_j*), duty cycle (DC), and phase shift (Δϕ) between the lower and upper jets. GA is adopted to seek optimal working parameters of the Coanda pulsation jets for the maximum recovery of base pressure (indicating drag reduction). The near wake flow of the D-shaped cylinder under optimal control is measured using particle image velocimetry, and then, thoroughly examined using proper orthogonal decomposition (POD), to investigate underlying mechanisms of drag reduction. Two sets of optimal control parameters differing in f_j* and Δϕ are identified by the GA-based optimization control, corresponding to the maximum base pressure recovery of 40.7%-43.3%. Significant modifications in the near wake flows are observed in the presence of the optimal control, based on instantaneous, time-mean and phase-averaged flow structures, Reynolds stresses, as well as the POD analyses. It is found that the Coanda pulsation jets working at the different sets of optimal parameters render distinct perturbations to the near wake flows, weakening the large-scale Karman vortices while energizing the small-scale structures.