Embedded Lift: Novel VTOL Approach
VTOL drones can provide a valuable platform for several tasks that require both the flexibility of a vertical take off and landing and the capability of long range flight. This drone was specifically designed to be used in search and rescue operations, where the capability of hovering and vertical take off and landing is fundamental, while the long range can significantly increase the coverage area.
First the main requirements were established. Based on our specific mission, we were able to determine:
Range: 100 km
Cruise speed: 70 km/hr
Cruise Height: 50 m
Estimated weight: 6kg
Given those mission requirements and estimated weight of the drone based on similar VTOL drones, a complete aerodynamic, structural, sizing and electrical analysis were iteratively performed to retrieve the initial dimensions and characteristics of the drone.
After the dimensions were tested and improved utilizing aerodynamic and structural simulations, a final manufacturable model was created, built and tested.
Below you can find the detailed breakdown of the process.
The finalized design and performance characteristics of the drone are as follows:
- Wing Span: 1.55 m
- Chord Length: 0.30 m
- Aspect Ratio: 5
- Cl: 0.56
- Airfoil: NACA 4418
- Cd: 0.07
- Lift/Drag: 8
- Power: 330 W
- Weight: 6.45 kg
- Hovering Time: 10 minutes
- Cruise Range: 93.33 km
The design process began with initial mission requirements: target range, cruise speed, flight altitude, and an estimated total weight. Using these parameters, we conducted an aerodynamic performance analysis to guide the sizing of the drone.
To evaluate efficiency across different wing loading conditions, the drag equation was reformulated to express thrust loading as a function of wing loading. This allowed us to generate a thrust-to-weight vs wing-loading curve. The curve reveals how the required thrust, drag, varies with wing area: at low wing loading, induced drag dominates and therefore total drag is high; as wing loading increases, induced drag decreases while parasitic drag begins to rise. The optimal wing loading is found at the minimum point on the curve, where total drag (and thus thrust requirement) is lowest.
This analysis delivers an initial optimal wing area given the estimated weight and initial chosen characteristics. With this value a cyclical and iterative process can occur. Different characteristics of the drone can be updated to show the performance of different geometries, dimensions and profiles of the drone. These characteristics can be the aspect ratio, taper ratio, swept back angle, dihedral angle, parasitic drag (form drag, interference drag, skin friction drag) for different materials, finishes, NACA airfoils and fuselage-wing geometry.
Next, an electronics and power analysis is incorporated into the iterative process. This step is crucial because the power needed to achieve the required range and thrust depends directly on the total drag. As drag increases, more energy is required, which in turn affects the size and weight of the battery. This updated battery weight contributes to the total mass of the drone, which must be fed back into the wing loading and thrust loading analysis. The process is repeated, updating aerodynamic, structural, and electrical parameters, until the assumed initial weight used in the performance analysis aligns with the final calculated weight, ensuring the design can meet the desired range and flight conditions.
Once an optimal wing loading vs thrust loading condition has been found, which not only takes into account aerodynamic performance but also manufacturing capabilities and cost, the initial prototype drone can be designed.
One of the most critical aspects of the drone's design was determining the optimal placement of the vertical lift propellers. Achieving sufficient pitch control authority through thrust vectoring alone required a detailed stability analysis. To enable this, the wings were swept back significantly, which increased the longitudinal distance, or moment arm, between the front and rear vertical propellers relative to the drone’s center of gravity. This configuration ensured adequate pitch control during all phases of flight. Nonetheless it also decreased the perpendicular airflow velocity with respect to the leading edge, this decreases lift. Therefore the resulting changes in aerodynamic behavior were fed back into the performance analysis, and additional iterations were carried out to refine the drone’s final geometry and ensure it met the targeted range requirements.
Below are two reference graphs from General Aviation Aircraft Design: Applied Methods and Procedures by Snorri Gudmundsson, which were used to support the analysis above. The thrust-to-weight graph on the left served as a guideline for comparing various aerodynamic characteristics of the drone. The graph on the right, illustrating the relationship between aspect ratio, taper ratio, and drag factor, provided key data for the iterative design calculations.
Once the initial drone design was finalized, manufacturing considerations became a key focus.
It was first decided that the drone would be entirely 3D printed, primarily due to budget constraints. However, we had access to medium-sized, high-quality 3D printers at no cost, which made this method highly attractive. While 3D printing introduces several disadvantages, such as relatively high material density, rough surface finishes, inconsistent quality, and reduced mechanical strength, we conducted structural calculations and small-scale validation tests to confirm that the material properties would meet our load requirements for a functional prototype and proof of concept.
Due to the dimensional limitations of the 3D printer beds, a modular design approach was adopted. The wing was split into individual segments, each connected using self-aligning mechanical features and reinforced with an internal aluminum spar. This modularity also provided added flexibility: segments could be easily replaced if damaged or reprinted with different profiles or geometries during iterative testing and optimization.
The fuselage was designed using the same NACA 4418 airfoil profile to maintain aerodynamic continuity, minimize abrupt changes in geometry, and reduce interference drag. This streamlined shape also helped minimize form drag, improving overall aerodynamic performance. The fuselage was dimensioned specifically to house the key internal components, including the battery, power distribution board (PDB), flight computer, radio transmitter, and camera. The wing segments were designed to contain the motor controllers and associated wiring, allowing for clean internal routing and modular electrical integration.
Below, on the left, is the layout of the electronic configuration inside the fuselage and wings, along with the required wiring. On the right is the manufacturing process, which features self-locating components in the wing segments. Each segment includes a male connector on one side and a female connector on the other. These segments are pressed together using an aluminum spar, ensuring proper alignment and secure connections.
Once the drone design was finalized, simulations of the full prototype were conducted to gain deeper understanding into both its aerodynamic and structural performance.
Given the innovative integration of vertical lift propellers embedded within the wing, it was essential to assess how the actual aerodynamic behavior would differ from the assumptions made during the initial sizing and performance calculations. These simulations were especially important for understanding the drone’s behavior across different flight regimes, including hover, transition, and cruise. Additionally, structural simulations were used to analyze how the wing segment connections, material thicknesses, and the internal aluminum spar would respond to different loading conditions.
Simulation results, while powerful, must be interpreted carefully. Wrong parameter settings or coarse meshing can produce misleading results. However, after careful tuning and validation of the simulation parameters, the results revealed a clear recirculation region behind the embedded vertical lift propellers, on both the top and bottom surfaces of the wing. These regions are characterized by low pressure and low velocity, with swirling eddies dissipating energy through viscous forces. The disturbed flow significantly alters the local aerodynamics.
This effect was most pronounced during low-speed horizontal flight, where the vertical propellers deflect a larger volume of air downward, increasing flow separation and aerodynamic disruption. At cruise speeds, while the effect was still noticeable, it was less severe due to higher freestream velocity overcoming some of the separation effects.
After multiple simulation iterations, it was estimated that: At low-speed forward flight, total drag increased by approximately 75% compared to the fixed-wing non VTOL baseline. At cruise speed, drag increased by about 25%.
While these are substantial increases in drag, it's important to note that these values were compared against a conventional fixed-wing aircraft baseline. When compared to a traditional VTOL drone with exposed vertical lift propellers, the embedded design performed comparably, and in some scenarios, even showed considerably lower drag.
Below are two key aerodynamic behaviors observed under different cruise flight scenarios. In the left image, no vertical thrust is applied, resulting in recirculating air inside the duct. The incoming boundary layer detaches when it reaches the duct and looses energy with the recirculating air from the duct, it is then unable to reattach, leading to a low-pressure, turbulent recirculation zone above the wing, just aft of the duct.
In contrast, the right image shows the vertical propellers generating a small amount of thrust. This alters the flow behavior significantly. The vertical flow interacts with the incoming horizontal air, deflecting it downward. As a result, a low-pressure, turbulent recirculation region forms beneath the wing, behind the duct.
After completing manufacturing and applying the insights gained from simulations, the flight testing campaign began.
The first objective was to demonstrate a controlled takeoff and landing, which would validate the drone’s thrust-to-weight ratio, stability, and real-world performance of the electrical and energy systems.
Despite multiple testing sessions, we were unable to achieve this objective. Post-test analysis, combined with bench testing, showed a critical issue: the vertical lift motors were severely underperforming. At their rated current draw, they produced only about 60% of the thrust specified in the manufacturer’s datasheet. In addition, they could only sustain this current draw for 5–10 seconds before overheating, and burning up.
While the theoretical design had a thrust-to-weight ratio of 2.0, the actual performance was far lower. This discrepancy led to two major problems. The first was ground effect instability. When operating near full throttle, the downward airflow interacted with the ground and recirculated into the embedded propeller ducts. Given its inability to produce enough lift to quickly get out of the ground effect recirculating air, the drone would lose control and yaw or roll uncontrollably. Second thermal failure before liftoff. Just before the motors could produce enough thrust to lift the drone, they would overheat and burn out, preventing sustained flight.
This outcome was unfortunate, as it originated from unreliable motor specifications. Furthermore, due to budget and timeline constraints, we were unable to purchase and test a range of motors to find one that could reliably meet our thrust requirements.
From this testing campaign, we reached two important conclusions
1. Reduce Weight
The drone's weight must be reduced wherever possible. This includes using lighter structural materials (e.g., carbon fiber instead of PLA), and minimizing the size and mass of the electronic components.
2. Scale Up the Drone for Better Propeller Performance
Perhaps the most critical insight is that the drone’s overall size needs to increase, not decrease. While this may seem counterintuitive, it is backed by fundamental aerodynamic and scaling principles:
The weight of the drone scales approximately linearly with size in our design (since both structure and components scale roughly proportionaly). However, thrust scales faster than linearly, roughly quadratically with propeller diameter, assuming RPM and efficiency are similar. Larger drones can accommodate larger propellers, which are not only capable of producing significantly more thrust, but do so more efficiently. Larger props have a higher thrust-to-power ratio (more Newtons per watt) and produce less heat per unit of thrust.
This means that, for a larger airframe, it becomes easier to achieve a higher thrust-to-weight ratio using the same power system, or achieve the same performance with lower current draw, reducing overheating risk.
Therefore, scaling up the drone improves the physics: it allows the use of more efficient propulsion systems, enables better thermal management, and reduces the pressure on each individual component. With a larger platform, embedded propellers become far more viable, making this architectural concept both practical and scalable in future iterations.
This is going to be our next step, we will not give up this goal until we have achived all of our testing objectives!
Thank you very much for your attention.
