Control of reverse flow over cantilevered swept blades using passive camber morphing

Abstract

One of the primary limitations experienced by high-speed rotorcraft is reverse flow that occurs on the retreating blade. Reverse flow is characterized by flow traveling from the sharp geometric trailing edge of the rotor blade to the blunt geometric leading edge. The reverse flow region is characterized by the generation of negative lift, a large increase in drag, and a sharp pitching moment impulse as the blade's orientation changes. Furthermore, due to the complex dynamics associated with rotorcraft, the cyclic pitch control can lead to reverse flow dynamic stall which can lead to pitch-link failure due to the large cyclic loads. A proposed way of mitigating the adverse effects of reverse flow is by introducing a trailing edge camber to the final 25\% of the rotor blade chord. By angling the sharp trailing edge of the blade into the flow, the size of the separation is reduced, leading to a corresponding reduction in the adverse loads. Previous studies have shown that this was an effective form of flow control on 2D and 3D unswept blades. The present work looks to extend that to cantilevered swept blades with sweep angles $\Lambda\pm20^\circ$ to assess the effect of spanwise flow on the proposed control of the reverse flow. At a Reynolds number $Re_c=1.79 \times 10^5$, aerodynamic load measurements were taken across a range of angles of attack $\alpha=175^\circ-205^\circ$ for three blades each with an aspect ratio of 3 and sweep angles $\Lambda=0^\circ$ and $\pm20^\circ$ in the presence or absence of trailing edge camber. Additionally, Stereoscopic Particle Image Velocimetry (SPIV) flow volumes were collected over the baseline and cambered swept blades at $\alpha=190^\circ$. Finally, specific spanwise locations over all blades were selected for SPIV measurements across a range of angles of attack. While the spanwise flow led to differences in the size and shape of separation along the span of the blade, the net loads were fairly similar across the three sweep angles. Furthermore, the introduction of trailing edge camber significantly reduced the size of the separation in all cases, leading to a reduction in drag of up to 40\% for the unswept and sweptback blades. To evaluate dynamic stall, the same blades were tested under a range of sinusoidal pitching motions. Cycle-averaged aerodynamic loads were acquired for the blades pitching sinusoidally about mean angles of attack $\overline{\alpha} =190^\circ$ and $\overline{\alpha} =195^\circ$, with pitch amplitudes $\alpha_A =5^\circ$ and $\alpha_A =7.5^\circ$. Note that these combinations of angles were chosen to assess the effect of shallow and deep dynamic stall on both the baseline and cambered blades. Additionally, to evaluate the effect of pitch rate, reduced frequencies of $k_f = 0.019$, $k_f = 0.037$, and $k_f = 0.075$ were tested for each pitching motion. From these measurements, a subset of pitching motions was chosen for phase-locked SPIV measurements acquired at the same spanwise locations as the static pitch measurements. The baseline blades all exhibited large hysteresis in the lift and pitching moment, as well as smaller hysteresis in drag. The phase-locked SPIV measurements showed differences in the size of the separation during pitch up as opposed to pitch down. The differences in these flowfields lead to the large hysteresis seen. By comparison, these differences over the cambered blade were reduced, leading to reductions in hysteresis in $C_L$, $C_D$, and $C_M$ of up to 82\%, 51\% and 71\%, respectively, for the sweptback blade. Similar results were seen for the blades at the other sweep angles.