In recent years, the study of drone formation flight has garnered significant attention due to its potential to enhance operational efficiency, expand surveillance coverage, and improve mission success rates in both civilian and military applications. As a researcher focused on control theory and engineering, I have been particularly interested in the aerodynamic interactions that occur when multiple drones fly in close proximity. These interactions, known as aerodynamic coupling, can lead to reduced drag and increased lift for trailing drones, thereby improving overall formation performance. This article presents a comprehensive investigation into aerodynamic coupling in drone formation flight, employing a methodology that combines vortex model analysis with computational fluid dynamics (CFD) simulations. The goal is to determine optimal spacing for drones in tight formations and quantify the resulting aerodynamic benefits, with a focus on maximizing the lift-to-drag ratio—a key metric for energy efficiency and endurance.
The concept of drone formation flight is inspired by natural phenomena, such as bird flocks, where individuals exploit updrafts from leaders to conserve energy. In aerial systems, similar principles apply: the leading drone generates wingtip vortices that create flow fields, which trailing drones can harness to reduce induced drag. However, achieving these benefits requires precise control of relative positions, as improper spacing can lead to detrimental effects like increased turbulence or loss of stability. Previous research has explored this through theoretical models, such as lift-line theory and horseshoe vortex representations, or via expensive wind tunnel tests. For instance, studies have shown drag reductions of up to 25% in formation flight, but these often lack practical validation or involve high costs. In my work, I address these limitations by integrating vortex modeling with CFD simulations, offering a cost-effective and detailed approach to analyze aerodynamic coupling in drone formation. This method not only validates theoretical predictions but also provides insights into real-world applications, such as aerial refueling or coordinated missions.
To begin, I developed a geometric model for a two-drone formation, representing the leading and trailing drones in an inertial coordinate system. In this model, the relative positions are defined by longitudinal spacing \(x_i\), lateral spacing \(y_i\), and vertical spacing \(z_i\), where the subscript \(i\) denotes the trailing drone. The primary objective was to identify optimal spacing that maximizes the trailing drone’s lift while minimizing its drag, thereby enhancing aerodynamic efficiency. The core of the analysis lies in understanding how the leading drone’s wingtip vortices induce velocity changes on the trailing drone. These vortices result from pressure differences between the upper and lower wing surfaces, causing airflow to roll up at the wingtips and create swirling motions. As the trailing drone encounters these vortices, it experiences an upwash velocity, which alters its effective angle of attack and, consequently, its aerodynamic forces.
Based on vortex theory, I derived mathematical expressions for the induced effects. The upwash velocity \(W_U\) leads to an increment in the trailing drone’s angle of attack, denoted as \(\Delta \alpha_i\). For small angles, this can be approximated as:
$$\Delta \alpha_i = \arctan\left(\frac{|W_{U\alpha_i}|}{V_i}\right) \approx \frac{|W_{U\alpha_i}|}{V_i}$$
where \(V_i\) is the trailing drone’s velocity and \(W_{U\alpha_i}\) is the dimensionless average upwash induced velocity. This change in angle of attack directly impacts lift and drag. The lift coefficient increment \(\Delta C_{L_i}\) is given by:
$$\Delta C_{L_i} = \Delta \alpha_i c_i = \left(\frac{|W_{U\alpha_i}|}{V_i}\right) c_i$$
where \(c_i\) is the lift curve slope of the trailing drone. Similarly, the drag coefficient increment \(\Delta C_{D_i}\) due to induced effects can be expressed as:
$$\Delta C_{D_i} = -C_{L_i} \frac{|W_{U\alpha_i}|}{V_i}$$
These formulas highlight how aerodynamic coupling in drone formation flight depends on relative positions and drone parameters. To quantify this further, I modeled the vortex strength \(\Gamma_{i-1}\) of the leading drone using circulation theory:
$$\Gamma_{i-1} = \frac{2}{\pi \lambda_{i-1}} C_{L_{i-1}} V_{i-1} b_{i-1}$$
where \(\lambda_{i-1}\) is the aspect ratio, \(C_{L_{i-1}}\) is the lift coefficient, \(V_{i-1}\) is the velocity, and \(b_{i-1}\) is the wingspan of the leading drone. Applying the Biot-Savart law, the average upwash induced velocity on the trailing drone was derived as:
$$W_{U\alpha_i} = \frac{\Gamma_{i-1}}{4\pi b’_i} (-z_i) \times \left[ \ln \frac{(\bar{y} – \pi/4)^2 + z’_i^2}{y’_i^2 + z’_i^2} – \ln \frac{y’_i^2 + z’_i^2}{(y’_i + \pi/4)^2 + z’_i^2} \right]$$
Here, \(b’_i\) is the effective wingspan, with \(b’_i/b_i = \pi/4\) based on elliptical lift distribution assumptions, and \(y’_i = y_i/b_i\), \(z’_i = z_i/b_i\) are dimensionless coordinates. Substituting these into the earlier equations yields comprehensive expressions for lift and drag increments in terms of formation geometry. For example, the drag coefficient increment becomes:
$$\Delta C_{D_i} = \frac{2}{\pi^3 \lambda_{i-1}} C_{L_{i-1}} C_{L_i} \frac{V_{i-1}}{V_i} \frac{b_{i-1}}{b_i} \times \left[ \ln \frac{(y’_i – \pi/4)^2 + z’_i^2}{y’_i^2 + z’_i^2} – \ln \frac{y’_i^2 + z’_i^2}{(y’_i + \pi/4)^2 + z’_i^2} \right]$$
Similarly, the lift coefficient increment is:
$$\Delta C_{L_i} = \frac{2 c_i}{\pi^3 \lambda_{i-1}} C_{L_{i-1}} \frac{V_{i-1}}{V_i} \frac{b_{i-1}}{b_i} \times \left[ \ln \frac{(y’_i – \pi/4)^2 + z’_i^2}{y’_i^2 + z’_i^2} – \ln \frac{y’_i^2 + z’_i^2}{(y’_i + \pi/4)^2 + z’_i^2} \right]$$
These equations form the basis for optimizing drone formation spacing. By analyzing them computationally, I determined that the optimal configuration occurs when the lateral spacing \(y_i = \pi b’_i/4\) and vertical spacing \(z_i = 0\), with longitudinal spacing \(x_i = 2b’_i\). This corresponds to specific distances that maximize lift and minimize drag for the trailing drone, leveraging the beneficial effects of aerodynamic coupling in drone formation flight.
To validate these theoretical findings, I conducted CFD simulations using a detailed model of the X-47B drone, a tailless, flying-wing design known for its stealth and efficiency. The process involved three-dimensional reconstruction of the drone’s geometry based on technical drawings and parameters, followed by mesh generation and flow solving. Using CATIA software, I created an accurate 1:1 scale model of the X-47B, ensuring that all aerodynamic features were captured. The mesh was generated with ICEM CFD, producing a structured grid around the drone surfaces to resolve flow details effectively. For the simulation, I considered a two-drone formation with the optimal spacing derived from the vortex model: longitudinal distance \(x_i = 33.11\) m, lateral distance \(y_i = 14.85\) m, and vertical distance \(z_i = 0\) m, based on the X-47B’s wingspan. The flight conditions were set to a cruising altitude of 2,000 m, Mach number 0.5, atmospheric pressure 79,495 Pa, and temperature 275.15 K, with the solver iterating to a residual tolerance of \(10^{-5}\).
The CFD simulations provided detailed aerodynamic data for the trailing drone in this tight drone formation. I compared these results with those from a single drone flying solo and from a formation with larger spacing, where the lateral distance was increased to 37.84 m while keeping other parameters constant. This allowed me to assess the impact of aerodynamic coupling across different configurations. The key metric was the lift-to-drag ratio \(L/D\), which indicates aerodynamic efficiency. For the tight formation, at an angle of attack \(\alpha = 0^\circ\), the trailing drone exhibited a lift coefficient \(C_L = 0.0874\) and a drag coefficient \(C_D = 0.0121\), resulting in \(L/D = 7.2231\). In contrast, for the single drone, \(C_L = 0.0850\), \(C_D = 0.0153\), and \(L/D = 5.5556\). For the large-spacing formation, \(C_L = 0.0849\), \(C_D = 0.0155\), and \(L/D = 5.4774\). These values clearly demonstrate that the tight drone formation enhances aerodynamic performance, with the lift-to-drag ratio increasing from 5.4774 to 7.2231—a significant improvement of approximately 32%. To illustrate this further, I performed simulations at other angles of attack, as summarized in the table below:
| Formation Type | Angle of Attack \(\alpha\) (°) | Drag Coefficient \(C_D\) | Lift Coefficient \(C_L\) | Lift-to-Drag Ratio \(L/D\) |
|---|---|---|---|---|
| Tight Drone Formation | 0 | 0.0121 | 0.0874 | 7.2231 |
| 1 | 0.0129 | 0.1442 | 11.1783 | |
| 2 | 0.0150 | 0.2005 | 13.3667 | |
| Large-Spacing Drone Formation | 0 | 0.0155 | 0.0849 | 5.4774 |
| 1 | 0.0163 | 0.1434 | 8.7975 | |
| 2 | 0.0183 | 0.2020 | 11.0383 | |
| Single Drone | 0 | 0.0153 | 0.0850 | 5.5556 |
The data shows that across all angles of attack, the tight drone formation consistently yields higher lift-to-drag ratios compared to both the large-spacing formation and solo flight. For instance, at \(\alpha = 1^\circ\), the ratio increases from 8.7975 to 11.1783, and at \(\alpha = 2^\circ\), from 11.0383 to 13.3667. This underscores the effectiveness of optimal spacing in harnessing aerodynamic coupling benefits. The improvement stems from the trailing drone positioning itself within the upwash region of the leading drone’s vortices, which reduces induced drag and enhances lift. In contrast, the large-spacing formation places the trailing drone outside the vortex influence zone, negating these advantages and resulting in performance similar to solo flight. This highlights the critical role of precise positioning in drone formation flight for energy savings.
To delve deeper into the vortex model’s predictions, I analyzed the sensitivity of aerodynamic increments to spacing variations. Using MATLAB, I computed \(\Delta C_{D_i}\) and \(\Delta C_{L_i}\) for a range of \(y’_i\) and \(z’_i\) values, with fixed \(x’_i = 2\). The results confirmed that the minimum drag increment and maximum lift increment occur at \(y’_i = \pi/4\) and \(z’_i = 0\), aligning with the optimal spacing derived earlier. This optimization process can be formalized as finding the extrema of the functions:
$$\min \Delta C_{D_i}(y’_i, z’_i) \quad \text{and} \quad \max \Delta C_{L_i}(y’_i, z’_i)$$
subject to constraints such as \(y’_i > 0\) and \(z’_i \geq 0\). The solutions provide a guideline for formation control systems in drone formation flight, enabling autonomous adjustments to maintain efficient spacing. Additionally, I explored the impact of drone parameters like wingspan and aspect ratio on coupling effects. For example, the ratio \(b_{i-1}/b_i\) influences the strength of induced velocities, suggesting that formations with drones of similar sizes may yield more predictable benefits. These insights are crucial for designing heterogeneous drone formations where different models collaborate.
The implications of this research extend beyond basic aerodynamic analysis. In practical scenarios, such as aerial refueling, the optimal spacing identified here can guide the docking process between a tanker and a receiver drone. By positioning the receiver in the tanker’s upwash zone, fuel consumption can be minimized during low-fuel conditions, enhancing safety and mission endurance. Moreover, the combined vortex-CFD methodology offers a scalable framework for studying larger drone formations with multiple trailing drones. In such cases, interactions become more complex due to overlapping vortex fields, but the principles remain similar. Future work could involve dynamic simulations where drones adjust positions in real-time to maintain optimal spacing amidst disturbances like turbulence or maneuvers.
In conclusion, this study demonstrates that aerodynamic coupling in drone formation flight can significantly improve aerodynamic efficiency when drones are positioned at optimal intervals. Through a blend of vortex modeling and CFD simulations, I have shown that a tight formation with specific lateral, longitudinal, and vertical spacing boosts the trailing drone’s lift-to-drag ratio from 5.4774 to 7.2231 at zero angle of attack, with further gains at higher angles. The vortex model provides a theoretical foundation for spacing optimization, while CFD validates these predictions with detailed flow analyses. This approach not only advances understanding of formation aerodynamics but also offers practical tools for designing energy-efficient drone operations. As drone technology evolves, leveraging such aerodynamic synergies will be key to enabling sustained, coordinated missions in areas like surveillance, delivery, and environmental monitoring. The continued exploration of drone formation flight promises to unlock new levels of performance and autonomy in aerial systems.