During my Erasmus studies at The Netherlands, I participated in the DBF Minor (Design, Build & Fly) at HvA. There, we were asked to build a drone within teams up to five people according to our customer requirements, and following a series of deadlines fixed by himself.
Most of the decisions regarding the design process rely on successfully accomplishing the mission: the UAV has to fly the DBF letters, within a predetermined airfield; drop a payload, in a preset specific location and without suffering any damage; it has to be safely hand-launched and fly in automatic mode, including safe landing.
Besides, the fact that the airplane must fulfill a list of requirements has settled many other design outcomes during this phase. Otherwise, the airplane won't have clearance to fly. There are 2 types of requisites: the DBF Minor requisites stated by the HVA, such as a maximum weight of 4 kg or a maximum wingspan of 2 meters; and the CS-23 requirements stated by EASA, like the maximum limit and ultimate loads the wing must withstand. All these are conditions sine qua non to fly the mission.
Some of the most relevant design choices have been: a trapezoidal wing, due to the lift distribution that this shape generates; a Clark-Y profile, due to its lift to drag ratio and stall entry characteristics; a T-tail, which reduces the chance of damage during landing; a smoothly-curved fuselage shape with a low lateral surface, improving the aircraft's stability and making it possible to comfortably install all the components inside the plane; a dropping mechanism, including a drag parachute that reduces the chance of damage of the payload; and finally, a landing gear, located underneath the propeller in order to protect it and the rest of the drone during landing or a crash.
The choices of the materials have been also constrained by the mission requirements. Since the wing skin must contain composite materials, it will be built with glass fiber. lts structure will be made out of carbon fiber and wood. The tail and the fuselage's skin will contain the same materials excluding the carbon fiber, due to requirements regarding weight and cost limits. The landing gear will be constructed in stainless steel to provide suficient strength. Since the nose cone is a hard-to-build shape, it will be 3D printed in PLA plastic. Finally, the control surfaces of the aircraft will be made of foam and laminated in glass fiber in order to keep them as light and strong as possible and easily replaceable. The production methods will include laser-cutting, 3D printing and composite materials laminates. Nevertheless, there are parts of the drone that will be bought and not built. Including parts made of wood, carbon fiber or stainless steel.
The aircrafts’ flight capabilities were tested in the Merlin Simulator. After achieving a stable flying UAV after several tests, the data of the parameters determining its flight performance such as the turning radius or stall speed among others, will be exported from the simulations and compared to the theoretical results obtained.
The mission execution was located at the NLR (Netherlands Aerospace Centre) in Emmeloord. They execute national and international assignments in the field of aircraft development, specially those regardings with UAVs. On Monday 11th of June the mission execution of the DBF minor took place under NLR's responsibility.
Before the official mission, the aircraft had to be tested during the pre-mission (test) flight. This flight was for NLR to test if everything was working like it should and to minimize the chance of something going wrong during the mission flight.
Before the aircraft could take off, the signals were tested on the ground. This was done by taking the transmitter about 500 meters away from the aircraft and then the staff members of NLR checked if the control services were still working from a distance. Everything still worked, so this aircraft could fly.
After a correctly performed take-off, the aircraft had a stable flight. During this flight, the maximum speed, climb angle and stability were tested and measured by the flight instruments. Then, after it flew for about 5 minutes, the engine support of the drone broke off. After breaking off, the propeller hit one side of the landing gear and those parts fell together with the bottom part of the nose cone and the engine. Even though those parts fell off mid-air, the aircraft was able to make a safe landing afterwards without any more damage.
The cause of this error was the weakness of the 3D-printed part which held the engine and propeller. The thrust was too high, and the density of the 3D-printed part was too low, which caused a tear in this engine connection. Due tO "the crash", this aircraft was not able to fly the mission. That's why the aircraft has not flown in automatic mode and the package with the parachute wasn't dropped either. Also, the shape of the DBF letters weren't flown in the air. Although, there was no mission flight, several tests were executed during the test flight. The data of these tests was downloaded and was used and processed in the evaluation report.
After 270 seconds of the test flight, the engine suddenly detaches from the drone's body. The first conclusion is simple: the 3D printed engine mount broke, triggering the contact of the propeller at maximum RPM speed with the landing gear that instantly provoked the complete detachment and wires disconnection of the engine from the rest of the fuselage.
What caused this failure that forced the drone to perform an emergency landing, avoiding to perform the mission? First, a component failure was considered as the issue: the metallic part of the engine that should be screwed to the engine mount was slightly bent due to hard landing in a previous DBF mission. This component could note be replaced on time and the drone had to fly with it.
Nevertheless, after meticulously analyzing the fissure the main problem was enlightened: the 3D printed engine mount was not hard enough. Despite that it was a 3D printed piece instead of a wood piece, the reason of the crack wasn't the weakness of the PLA plastic, but the fact that this piece was a TEST part: its filling density was only 5%, unacceptably below the correct calculated value: 75%.
These "test parts" were printed to check the fitting and sizing of the 3D printed parts before printing the final piece saving time and money in case of failure. In this case, epoxy gluing the test piece (by mistake) to the fuselage made it impossible to unmount the part and install the correct one for the flight day on time.
Although there was no mission flight, the previous test flight gave enough data to analyze the aircraft’s final performance, comparing it to the calculated values during the design phase. The main aspects analyzed were:
Although this project was not without its problems, it was a success. Despite not being able to fiy in the flight of the mission, the aircraft could at least fly for about 5 minutes without any problems and according to the design calculations.
The production methods used were a hit and miss. Some methods were very poor due to not using the workshops correctlly. This caused big delays in the construction process. h,moreover, the design was also very complex, which added even more delays in the construction process.
The mission execution ended with mixed feelings, because the aircraft flew on one side but on the other side the actual mission flight could not be performed due to the emergency landing. The structural failure that occurred was caused by a weak PLA parte that was used by mistake on the flight. The filling density was way too low, and this should have been checked.
For the flight dynamics most data matched the calculated performance and Merlin Simulator's data. The flight dynamics also show that the aircraft would have been able to fly the mission because the turning radius was as expected and as other systems like manual control, Return to Launch and live stream of data to ground station worked perfectly.
So, despite everything, the drone was able to meet all the established requirements and have its systems installed and ready for flight on the day of the mission. And, although it was not able to perform the mission completely, all the analysis of telemetry data and its close value to the calculated ones suggest that, finally, it would have been able to fulfill the mission without any problem if it had not been for the failure of the engine support.