An experimental particle swarm optimization-driven real-time control of Tollmien–Schlichting waves on a flat plate

Abstract

This experimental study demonstrates the feasibility of integrating particle swarm optimization (PSO) within a hardware-in-the-loop framework for real-time active flow control. The objective is the attenuation of Tollmien–Schlichting (TS) waves developing in an incompressible, two-dimensional laminar boundary layer, using dielectric barrier discharge plasma actuators. Single- and multi-frequency TS waves are artificially excited and amplify before reaching the control region. Control actions are optimized online using PSO, which iteratively updates a population of candidate solutions based on a pressure-based performance metric. In single-frequency scenarios, candidate solutions are parametrized as finite impulse response filters, convolved with an upstream pressure signal to generate the control voltage. In multi-frequency scenarios, the controller constructs a linear superposition of sinusoids at the triggered TS frequencies, identified online. The effectiveness of control is quantified by the attenuation of pressure fluctuations measured by wall-mounted microphones downstream of the actuator. Upon convergence, the best-performing control actions are re-tested in independent experiments, using two-component particle image velocimetry (PIV) to assess the effect of the controller on the velocity field inside the boundary layer. The results show that PSO consistently achieves appreciable TS waves suppression, as evidenced by pressure fluctuation levels reduced by 30%–40% relative to the uncontrolled case and by weakened phase-coherent velocity disturbances. Furthermore, phase-free PIV measurements demonstrate a downstream delay in the growth of velocity fluctuations of up to 15 δ₀*, where δ₀* is the displacement thickness at the control location. These findings confirm the feasibility of PSO-driven optimization for real-time flow control in transitional boundary layers.