1. Introduction
Fixed-wing drones have become indispensable in applications such as aerial surveillance, disaster communication, and precision agriculture due to their long endurance, high maneuverability, and large payload capacity. However, low-altitude complex wind shears and strong convective atmospheric motions pose serious challenges to the flight stability of these platforms. Stall, triggered by flow separation at high angles of attack, is a critical issue that degrades lift and increases drag. Inspired by avian flight, natural selection has equipped birds with specialized feather structures—coverts—that can passively deploy under separated flow conditions to suppress stall. By mimicking these biological features, we aim to enhance the aerodynamic performance of fixed-wing drone wings.
Recent studies have demonstrated that flexible materials, when applied as biomimetic devices, exhibit superior flow control capabilities compared to rigid counterparts. The interaction between flexible flaps and the surrounding fluid can transform large-scale turbulent structures into smaller scales, thereby reducing unsteady loads and noise. However, the placement of such devices on the wing surface is crucial to maximize their effectiveness. In this work, we experimentally investigate the effect of bio-inspired flexible serrated coverts installed at different chordwise positions on a straight wing at a high angle of attack. The experiments are conducted in a low‑turbulence wind tunnel using a NACA0018 profile. The focus is on the wake flow characteristics, including mean velocity recovery, turbulence intensity reduction, and spectral energy redistribution. The results provide insights into the underlying mechanism and practical guidelines for the design of fixed-wing drone flow control systems.
2. Experimental Setup
2.1 Wind Tunnel and Wing Model
All measurements are performed in a low‑speed, closed‑loop wind tunnel at our laboratory. The test section dimensions are 2.3 m (length) × 1.0 m (width) × 1.0 m (height). The free‑stream turbulence intensity is about 0.1 % for the velocity range of 5–60 m/s. A straight wing with a NACA0018 airfoil section is used, having a chord length c = 300 mm and a span s = 1.0 m. The wing is mounted vertically at the centre of the test section, spanning the full tunnel height to ensure quasi‑two‑dimensional flow in the mid‑span region. The free‑stream velocity is set to U∞ = 25.0 m/s, corresponding to a chord‑based Reynolds number
$$Re = \frac{U_{\infty} c}{\nu} = 5.1 \times 10^5.$$
The angle of attack is fixed at α = 15°, which causes a fully separated stalling flow over the upper surface.
2.2 Bio-Inspired Flexible Coverts
The artificial coverts are fabricated from a 0.5 mm thick silicone flexible membrane, cut into a serrated shape. Each covert consists of a 20 mm long base (attached to the wing surface) and a 30 mm long sawtooth with a width of 15 mm. The base prevents reverse flow behind the covert, mimicking the overlapping region of bird coverts. The flexible coverts are installed at six different chordwise positions on the upper surface: 10%c, 20%c, 40%c, 60%c, 80%c, and 100%c. A clean wing (without coverts) serves as the baseline. In total, seven configurations are compared.
2.3 Hot‑Wire Anemometry
A constant‑temperature hot‑wire anemometer (IFA300) with a single‑wire probe (tungsten, diameter 5 μm, length 2 mm, over‑heat ratio 1.5) is used to measure the streamwise velocity in the wake. The probe is traversed by a three‑axis computer‑controlled system. Measurements are taken at x/c = 0.7 downstream of the trailing edge, along the y‑direction from y = −80 mm to +180 mm, in 20 steps. The coordinate origin (y = 0) corresponds to the mid‑chord of the wing. Sampling is performed at 4000 Hz for 65.5 s at each point, providing statistically converged data.
3. Results and Discussion
3.1 Time‑Averaged Wake Velocity and Turbulence Intensity
The time‑averaged streamwise velocity U normalized by U∞ at x/c = 0.7 is plotted as a function of y/c for all cases in Table 1. The clean wing exhibits a pronounced velocity deficit at y/c ≈ −0.17, indicating a large separation bubble. When coverts are placed at 20%c and 80%c, the deficit is significantly reduced, and the wake recovers faster. In contrast, coverts at 10%c, 40%c, 60%c, and 100%c show only moderate improvement.
| Configuration | Umin / U∞ | Peak Urms / U∞ |
|---|---|---|
| Clean | 0.12 | 0.21 |
| 10%c | 0.18 | 0.16 |
| 20%c | 0.31 | 0.14 |
| 40%c | 0.15 | 0.18 |
| 60%c | 0.14 | 0.17 |
| 80%c | 0.28 | 0.09 |
| 100%c | 0.11 | 0.20 |
The root‑mean‑square (RMS) of velocity fluctuations, Urms / U∞, reveals the presence of two shear layers: one originating from the leading edge (negative y) and one from the trailing edge (positive y). For the clean wing, two distinct peaks appear. Installing coverts reduces both peak magnitudes, with the most dramatic suppression observed for the 80%c case, where the two peaks become very close and the turbulence level drops below 0.1. This indicates that the flow is re‑attached over most of the upper surface, leaving only a thin separated region near the trailing edge. The 20%c case also shows a significant reduction, but the peaks remain separated.
3.2 Frequency‑Domain Analysis
Power spectral density (PSD) of the velocity fluctuations is computed at the peak RMS locations (leading‑edge shear layer and trailing‑edge shear layer). The PSD is dimensionless using
$$\frac{P}{c\,U_{\infty}} \quad \text{versus} \quad \frac{f\,c}{U_{\infty}}.$$
Results are summarized in Table 2. The clean wing shows no clear spectral peak; energy is broadly distributed at low frequencies (f c / U∞ < 1). When coverts are placed at 20%c, a peak appears at f c/U∞ ≈ 0.3–0.4, indicating a large‑scale coherent motion induced by the flapping coverts. For the 80%c case, the peak shifts to f c/U∞ ≈ 3, demonstrating that the coverts break down large structures into small‑scale, high‑frequency fluctuations. Similar trends are observed in the trailing‑edge shear layer.
| Configuration | Leading‑edge peak f c/U∞ | Trailing‑edge peak f c/U∞ |
|---|---|---|
| Clean | – | – |
| 20%c | 0.35 | 0.75 |
| 80%c | 3.0 | 3.2 |
The shift towards higher frequencies is beneficial for fixed-wing drone applications because it reduces unsteady loads and noise. A coherence analysis between the leading‑edge shear layer signal and the trailing‑edge shear layer signal is performed using
$$\gamma^2 = \frac{|P_{xy}(f)|^2}{P_{xx}(f) P_{yy}(f)}.$$
For the clean wing, the coherence is low across all frequencies. In the 20%c case, a strong peak appears at f c/U∞ ≈ 4.5, indicating a coupling between the two shear layers with a preferred scale. For the 80%c case, multiple coherence peaks are observed both in the low‑ and high‑frequency ranges, suggesting a more distributed interaction that enhances the mixing and reduces the overall separation.
3.3 Multi‑Scale Coherent Structures via Wavelet Analysis
To further reveal the temporal evolution of the coherent structures, we apply a discrete wavelet decomposition to the velocity signals. The fluctuating velocity is decomposed into ten frequency scales (Table 3).
| Scale s | Frequency range (Hz) | f c / U∞ |
|---|---|---|
| 1 | 1.95–3.91 | 0.023–0.047 |
| 2 | 3.91–7.81 | 0.047–0.094 |
| 3 | 7.81–15.63 | 0.094–0.188 |
| 4 | 15.63–31.25 | 0.188–0.375 |
| 5 | 31.25–62.50 | 0.375–0.750 |
| 6 | 62.50–125.0 | 0.750–1.500 |
| 7 | 125–250 | 1.5–3.0 |
| 8 | 250–500 | 3–6 |
| 9 | 500–1000 | 6–12 |
| 10 | 1000–2000 | 12–24 |
Wavelet coefficient maps (time versus normalized frequency) show that the clean wing contains large, low‑frequency “U‑shaped” patterns indicative of bursting events. The 20%c coverts reduce the number of such patterns from three to two, and the 80%c coverts nearly eliminate them, leaving only small‑scale, high‑frequency fluctuations. This confirms that the adaptive fluttering of the coverts at 80%c effectively breaks down the large‑scale separation structures into smaller eddies, thereby suppressing stall. The energy is concentrated at f c/U∞ ≈ 3, consistent with the PSD results.
4. Conclusion
We have experimentally investigated the flow control capability of bio‑inspired flexible serrated coverts on a straight wing at a high angle of attack, relevant to fixed-wing drone stall mitigation. Key findings are:
- Coverts installed at 20%c and 80%c both reduce the mean velocity deficit and turbulent fluctuations in the wake. However, the 80%c location yields a more profound effect, nearly eliminating the velocity deficit and reducing turbulence intensity by more than half.
- Frequency analysis shows that coverts at 20%c produce a dominant low‑frequency peak (f c/U∞ ≈ 0.35) associated with large‑scale flapping, while coverts at 80%c shift the energy to high frequencies (f c/U∞ ≈ 3), indicating a transition from large‑ to small‑scale structures.
- Wavelet analysis reveals that the 80%c coverts effectively suppress the low‑frequency coherent bursting events and promote the generation of fine‑scale turbulence, which enhances mixing and delays separation.
- The underlying mechanism is the adaptive vibration of the flexible coverts: when placed near the trailing edge (80%c), the coverts attain a quasi‑equilibrium position with small‑amplitude oscillations, which efficiently break down the unstable shear layer. In contrast, coverts near the leading edge (20%c) undergo large‑amplitude flapping that, while still beneficial, is less effective in destroying large coherent structures.
Our results demonstrate that bio‑inspired flexible coverts offer a passive, adaptive, and highly efficient solution for stall control in fixed-wing drones. The 80%c placement is recommended for practical implementations. Future work will focus on the effects of multiple coverts in tandem and the integration into morphing wing concepts for real‑world fixed-wing drone applications.