Towards silent and efficient flight by combining bioinspired owl feather serrations with cicada wing geometry

The study addresses the challenge of reducing rotor/propeller noise while maintaining or improving aerodynamic performance for aerial vehicles. Conventional propeller design, often based on B-spline methods, encounters tradeoffs between aerodynamic efficiency and passive noise reduction when employing 2D edge serrations. Drawing inspiration from biology, especially the stealthy flight of owls (leading-edge serrations, trailing-edge fringes, and downy coatings) and the aerodynamically efficient planforms of cicada forewings, the authors propose integrating these morphologies. The research goal is to create and evaluate a 3D sinusoidal serration topology on a cicada-inspired planform that can concurrently suppress noise and enhance efficiency across relevant Reynolds numbers for drone propellers. The importance lies in mitigating urban aerial vehicle noise pollution without sacrificing energy efficiency and thrust.

The paper situates its work within bioinspired aeroacoustics and propeller design. Prior research on owl-inspired devices has identified key features (leading-edge serrations, trailing-edge fringes, porous/velvet-like coatings) and demonstrated various noise control approaches using 2D serrations and porous materials. However, these often degrade aerodynamic performance. Cicada wing planforms, despite operating at lower Reynolds numbers than owls, have been shown via simulations and experiments to confer aerodynamic benefits across a range of Reynolds numbers relevant to UAV propellers. Conventional design frameworks (e.g., Bézier/B-spline techniques) excel at shaping but do not inherently resolve the noise–efficiency tradeoff. The literature indicates a gap: combining complex 3D bioinspired surface topographies with beneficial planforms to achieve both noise suppression and aerodynamic gains.

Design and topology: The authors digitized cicada forewing planforms and owl feather serration morphologies to generate a 3D sinusoidal serration across the blade surface (not just edges). The cicada-like chord distribution is defined by fifth-order polynomial fits for leading and trailing edges (C_l, C_t), with span b=3 in (rotor radius). A sinusoidal surface waveform C_s(x)=A·sin(2πx/λ) is superimposed, with wavelength λ parameterized as λ=A·A+λ0. The 2D airfoil section is NACA 8412. Reynolds number is defined as Re=(ρ ω x / μ)·cos(α). Pitch angle for cicada-based prototypes is fixed at 15° to prevent separation at the operating Re. Prototypes and benchmarks: Five designs were fabricated (6-inch diameter): 3D-SC (3D sinusoidal serrations on cicada planform), B1 (smooth cicada planform), B2 (conventional planform with serrations), B3 (conventional planform without serrations), and B4 (DJI Phantom 3 industry benchmark, obtained via 3D scanning/rescaling). For fairness, the 3D-serrated prototype used amplitude A=0.04 in and wavelength λ=0.4 in; leading-edge serrated controls spanned a range including this baseline. Additional 12-inch versions of 3D-SC and B4 were fabricated for scalability tests. Fabrication: PolyJet 3D printing (Stratasys) with digital ABS material; ~32 μm layer thickness. Post-processing included wet sanding, thin epoxy coating, and spray painting with care to preserve serration geometry. Aeroacoustic experiments: An omnidirectional USB microphone (miniDSP UMIK-1) recorded SPL at 48 kHz. Measurements were made at radial distances of 0.1 m and 5 m, at multiple circumferential angles; OASPL computed by integrating spectra (0–2 kHz cited), using a 2nd-order Savitzky–Golay filter (333-point window). A 100 Hz high-pass cutoff was applied to remove DC bias. Rotor actuation used a thrust stand with brushless motor and ESC; RPM measured by a digital tachometer. Each condition was measured in triplicate; uncertainty quantified (e.g., OASPL SD ~0.37–0.79 dB; CV ≤0.86% at 5 m). Aerodynamic experiments: TYTO 1585 thrust stand measured thrust, torque, electrical power, and RPM at 40 Hz. Each plotted data point averaged 100 samples; uncertainty quantified over 1000-point datasets (thrust SD ~0.29–0.66 gf; power SD ~0.032–0.097 W). Tests covered 2000–6000 RPM and thrust levels 10–50 gf for 6-inch blades; 12-inch blades tested up to 3000 RPM with results at 150 gf. Parametric study: Sixteen 3D-SC variants with amplitudes A=0.01–0.04 in and wavelengths λ=0.1–0.4 in were tested at 2000 and 5000 RPM. High-order surface interpolation produced contour maps of OASPL and thrust versus waveform parameters. Observed diagonal equipotential trends tied to a slope (aspect ratio λ) and intercept λ0, with local OASPL minima identified at specific (λ, λ0) pairs. Computational simulations: CFD used LES for aerodynamics and Ffowcs Williams–Hawkings (FW-H) for acoustics, coupling velocity/pressure fields to compute sound pressures each timestep. Domains comprised a rotating inner region around the propeller with inflation layers and a large stationary outer region (radius ~450 mm, inlet/outlet distances 500/600 mm). Boundary conditions included pressure inlet/outlet (0 atm gauge), no-slip blade surfaces, slip outer wall, and frozen-rotor interfaces. Simulations were run at 2000 and 5000 RPM to probe low- and high-Re regimes. Post-processing analyzed streamlines, helicity iso-surfaces, swirling strength iso-surfaces (color-coded by vorticity), and surface dipole/quadrupole source strength contours to link flow structures to tonal/broadband noise mechanisms.

• Noise reduction vs industry benchmark: 3D-SC reduced OASPL by up to 5.5 dB at 5 m and 30° (50 gf) compared with B4; A-weighted maximum reduction 5.2 dB(A).
• Noise reduction vs controls: 3D-SC showed 8.3 dB lower OASPL than B1 at 0.1 m, 50 gf. Conventional serrated B2 had 8.8 dB lower OASPL than conventional non-serrated B3. Cicada planform alone (B1 vs B3) yielded modest OASPL reductions (up to 1.6 dB at 0.1 m, 50 gf; 1.9 dB at 5 m).
• 3D vs 2D serrations: At identical amplitude/wavelength (0.04×0.4 in), a 3D-serrated propeller’s OASPL was 3.63 dB lower than a leading-edge serrated counterpart; 3D serrations consistently lowered acoustic emission across the spectrum versus various 2D leading-edge serrations.
• Spectral characteristics: At low thrust (15 gf; Re ≈1.01×10^4), tonal (loading) noise dominates; 3D-SC reduced multiple harmonic SPL peaks versus B1–B4. At higher thrust (50 gf; Re ≈1.82×10^4), broadband noise dominates and 3D-SC reduces SPL across the spectrum relative to smooth or conventional designs.
• Aerodynamic performance (6-inch): At 5000 RPM, B1 vs B3 increased thrust by 14.8 gf (+39.2%) due to the cicada planform. 3D-SC produced 20.3 gf more thrust than B4 at 5000 RPM. Thrust coefficient for 3D-SC was ~0.04 higher than B4 (+55.6%).
• Propulsive efficiency (6-inch, 50 gf): 3D-SC used 0.17 W less power than B1 (4.1% improvement). B1 vs B3 reduced power by 1.49 W (26.7%). 3D-SC reduced power by 20.2% versus B4 at equal thrust (concurrent thrust and efficiency improvement).
• Scalability (12-inch, 150 gf): 3D-SC reduced mechanical power by 2.29 W versus B4, a 22.6% improvement.
• Parametric insights: Densest serrations (A=0.04 in, λ=0.1 in) minimized OASPL at 2000 RPM but increased OASPL at 5000 RPM, indicating Re-dependent optimality. OASPL minima identified at 2000 RPM (79.7 dB) for λ=16.0, λ0=−0.223 and at 5000 RPM (91.4 dB) for λ=16.2, λ0=0.073. Reduced serration intensity (larger λ, smaller A) tended to increase thrust.
• Mechanism (CFD): 3D-SC generated spanwise flow and coherent vortex structures (CVS) across the span, especially near trailing edge, elevating surface vorticity and helicity compared to smooth B1. At low Re, CVS reduced dipole pressure sources (tonal noise). At high Re, CVS presence suppressed quadrupole sources near the trailing edge (broadband noise), aligning with measured spectral shifts.

The findings demonstrate that combining a cicada-inspired planform with 3D sinusoidal serrations achieves dual objectives: substantial passive noise mitigation and improved aerodynamic performance. The cicada planform primarily boosts thrust and efficiency, while the 3D surface serrations drive broadband and tonal noise reductions beyond 2D serrations. CFD elucidates that valley-like 3D topography induces pressure distributions that promote coherent vortex structures, which mitigate harmonic loading interactions at low Re (reducing dipole sources) and stabilize larger-scale flow at high Re (suppressing energy cascade to small eddies and hence quadrupole sources). The performance sensitivity to serration amplitude and wavelength indicates design tunability to target specific operating conditions, enabling application-dependent optimization. These synergistic geometric effects are not replicated by either feature alone, underscoring the importance of integrated bioinspired design for quiet, efficient aerial systems.

This work introduces a 3D sinusoidal serration propeller topology on a cicada-inspired planform that reduces overall sound pressure levels by up to 5.5 dB relative to an industry benchmark while simultaneously increasing thrust and propulsive efficiency (over 20% power reduction at equal thrust). Experiments across distances and thrust settings, combined with LES–FW-H simulations, verify that 3D serrations create coherent vortex structures that suppress tonal noise at low Re and broadband noise at high Re. Parametric studies reveal that serration amplitude and wavelength critically govern aeroacoustic and aerodynamic outcomes, enabling targeted optimization for specific operating regimes. Future research directions include refined cicada planform tuning, broader parametric exploration of 3D serration fields, in-flight and multi-rotor studies, and scaling to larger rotors and different fluids (e.g., wind/hydro turbines) to generalize the multifunctional benefits.

• Planform model: The cicada wing geometry used was a preliminary model and not optimized; fine-tuning for improved performance is deferred to future work.
• Scale and materials: Most experiments used 6-inch rotors due to the limited strength of digital ABS material; only 12-inch 3D-SC and B4 were tested for scalability.
• Test conditions: Measurements were primarily static bench tests (not in-flight), with acoustic data taken at 0.1 m and 5 m; far-field community noise implications beyond 5 m were not assessed.
• Spectral and parameter bounds: Acoustic spectra and optimization covered specific RPM ranges (2000–6000) and serration parameters (A=0.01–0.04 in, λ=0.1–0.4 in); generalization outside these ranges requires further validation.
• CFD assumptions: Simulations used specific meshing, boundary conditions, and solver settings (LES with FW-H, frozen rotor interfaces); results may vary with alternative turbulence models or higher-fidelity transient coupling across domains.