Rogue Squadron (Team Alpha)

Dragonfly Unmanned Aerial Vehicle Development

Unmanned Aerial Systems (UAS) are a blooming technology.  Aerospace companies around the world are realizing the countless applications of this technology and it is said to be growing faster and faster. The importance of UASs have been recognized by the Association for Unmanned Vehicle Systems International (AUVSI) in their Student Unmanned Aerial Systems (SUAS) competition. This competition asks student teams to create their own UAS. For our senior design, we propose to design an airframe and validate the flight controller for the 2016 AUVSI SUAS competition. SUAS would not only give us a chance to work with UASs hands on, but would also open up the possibility of innovation. In addition to these benefits, our experiences would ready ourselves more fully to make solid and meaningful contributions to the industry after graduation as engineers and/or researchers.

The AUVSI-SUAS Competition is held near the end of June at the Patuxent Naval Air Base in the Maryland Peninsula. Last year, teams from all over the world came to signed up to compete with only a little more than half being american teams. With such a diverse array of peoples it was very interesting to see the systems and the software that others had developed to complete the same missions. The innovation sparked by this competition was apparent in the eyes of the many industry representatives who would walk up and shake our hands, ask us about our plane, and then leave a business card. Their interest in the development of our UAS was apparent.

The competition employs a similar ruleset from year to year.  The biggest and most important missions include the ability for the UAS to takeoff, navigate, and land autonomously. Other missions involve the systems ability to search for targets. These targets vary in type and size from alphanumeric targets placed on a shape then laid on the ground, to infrared targets. Once these targets are found they can be classified based on colors, alphanumeric characters, location, and orientation. Teams can attempt to find and classify targets autonomously to achieve a higher score.  In addition to these missions there are other more intensive missions such as sense, detect, and avoid obstacles, interoperability, and many others. Due to the difficulty and number of missions, the majority of teams do not attempt every available mission. This leads to a robust competition with a large amount of diversity in the missions that each team is attempting.

The competition employs a similar ruleset from year to year. The biggest and most important missions include the ability for the UAS to takeoff, navigate, and land autonomously. Other missions involve the systems ability to search for targets. These targets vary in type and size from alphanumeric targets placed on a shape then laid on the ground, to infrared targets. Once these targets are found they can be classified based on colors, alphanumeric characters, location, and orientation. Teams can attempt to find and classify targets autonomously to achieve a higher score. In addition to these missions there are other more intensive missions such as sense, detect, and avoid obstacles, interoperability, and many others. Due to the difficulty and number of missions, the majority of teams do not attempt every available mission. This leads to a robust competition with a large amount of diversity in the missions that each team is attempting.

Project Overview

Senior design team Alpha has been hired by the Saint Louis University Association for Unmanned Vehicle Systems International (AUVSI) Student Unmanned Aerial Systems (SUAS) team to design and build an airframe capable of supporting its flight control and imagery systems. The AUVSI competition is primarily a systems engineering competition that does not revolve around the design of an aircraft for each competition. Because of this, the AUVSI team has the desire for an airframe capable of supporting the team over the next five years of competition.

The task of team alpha was to design and build an airframe capable of allowing the AUVSI team to compete in the next five years of competition. This airframe will facilitate the use of an internally mounted gimbal from which the team may collect data and images. The plane will have ample payload volume to facilitate all flight controls and allow for expansion over the next five years. Additionally, Team Alpha has designed the Dragonfly to be easily manufactured and recreated by the team. This will ultimately allow the AUVSI team to be competitive at the AUVSI competition.

Manufacturing Materials

• • This will be a general overview of the materials required and utilized in the production of composite components for the aircraft. Along with the description of each item, suggestions or advice for proper use may also be found. It is suggested that before producing any of the components, individuals review each item to avoid any issues experienced by Rogue Squadron members upon their first exposure to composite material construction methods.

    • A.              Epoxy Resin and Cure

    • The epoxy resin and cure shown are mixed at a ratio, by weight, 100:27. Typically, applications require hundreds of grams of epoxy, so it is important to adhere to the weight ratio in order to use a consistent product. Mix for at least a minute. Apply epoxy to material liberally. If the materials are not saturated properly, the component will develop bubbles when cured which will weaken the piece. Once the epoxy is mixed, there is a two hour working life.

    • B.              Peel Ply

    • A crucial component for the vacuum bagging process, peel ply is wrapped around the composite material components that have had epoxy applied. During the bagging process, excess epoxy is drawn out through the fabric. Once the bagging process is complete, the epoxy will have hardened and the peel ply will be removed from the material. To limit epoxy issues, it was ensured that every wet surface was covered with peel ply.

    • C.              Breather Cloth

    • Breather cloth serves two main purposes. It is porous enough to allow the vacuum process to evenly apply pressure across the surface. Secondly, the cloth absorbs excess epoxy that is drawn out during the vacuum bagging process, thus must be placed on top of the peel ply. For most applications, two layers of breather are used to ensure that all epoxy is absorbed. If excess epoxy is not absorbed, components will retain the epoxy, increasing weight and fragility.

    • D.              Bagging Film and Sealant Tape

    • Used in the creation of vacuum bags, the bagging film is capable of stretching 200 times its size. The sealant tape is used to attain a vacuum tight seal around the bag. Components that have been epoxied are placed inside the bag created. Sealed up and vacuum sealed. This process adheres the composite materials and ensures smooth finishes. Bagging film is easily pierced during the bagging process and therefore must be handled with care.

    • E.              Fiberglass Cloth (8 oz)

    • Lightweight, strong, and resilient, the fiberglass cloth is woven to provide equal strength in both horizontal and vertical yarn directions. Used for both fuselage and wing surface construction the fabric easily molds into desired shapes. Excessive handling will rearrange the fibers, creating non-uniform yarn directions. Cut the material with a rotary cutter to insure that weave will not be disturbed.

    • F.               Carbon Fiber Weave

    • Similar to fiberglass cloth, the carbon fiber cloth is lightweight and is woven to provide equal strength in both horizontal and vertical yarn directions. However, the material is stronger in terms of tensile strength, but more ductile, increasing the likelihood of fracturing. For the fuselage, this material is paired with layers of fiberglass to combine the favorable qualities of each material. Similar to fiberglass the weave can be disturbed, therefore one must be cautious when working with carbon fiber weave.

    • G.             Unidirectional Carbon Fiber

    • This fabric shares the material properties of the carbon fiber weave, but all fibers are laid in one direction, providing strength in only one direction. For this reason, this fabric is used in the creation of the landing gear, so that it will deflect along one direction. When applying epoxy, only brush strokes along the direction of the fibers were used to limit the disturbance of the fibers.

    • H.             Foam Insulation

    • Foam insulation is used as the structure for the wings and tail surfaces. The foam wire cutter is used to cut the foam into the desired shapes. The foam comes in panels of 2 in. x 4 ft. x 8 ft. Waste can be eliminated by adding  and providing a backup wing. Cut using a painter’s tool or with the hot wire foam cutter.

    • I.                Honeycomb

    • Honeycomb structure provides high amounts of resistance to compression and shear forces. For these reasons, it is used as the internal structure for the main landing gear. Honeycomb is filled with foam to stop epoxy from filling in the voids and hardening, which would increase fragility of the structure and increase weight.

    • Manufacturing Process

Wings

It goes without saying that by design the wings of an aircraft support the majority of structural weight in flight. For this reason, it is important that the wings of an aircraft be strong enough to facilitate aerodynamic loading but light enough as to not surpass the MTOW requirement of 25 pounds. Because there is no 'one size fits all' solution for the amount and type of material that is necessary for construction it was prudent that we test various material configurations and determine which worked best for our design.

Insulation foam was chosen to be the internal material for the wing as it is readily available at any hardware store, reasonably lightweight and it can be easily cut to shape using the hotwire CNC machine available in the Saint Louis University composites lab. The CNC machine cut the insulation foam to the shape of a FX 63-137 airfoil at a 14-inch cord length. Once cut, the foam was sanded to allow a smooth surface to lay composite materials. When sanding was complete, the foam trailing edge was removed and replaced with a 1/16-inch strip of balsa wood which was then sanded to give a thin trailing edge. This allowed the composite material to end at the tip of the airfoil in the thinnest possible fashion when laid over the wing section.

Sandwich configurations of woven carbon fiber and fiberglass were tested and compared based on weight and strength. Upon comparison, it was determined that a single layer of fiberglass laid on top of the foam would be sufficient to support the aircraft and minimize the weight. Next, a sandwich structure of unidirectional carbon fiber and 1/32-inch balsa was tested within the wing structure to see if the resulting strength benefits would be enough to justify increased weight. The wing was determined to gain significant strength by adding this sandwich construction to the airfoil's upper and lower surfaces but was ultimately determined to be over-engineered and detrimental to the weight budget. The final configuration for the wing was decided to be made using one layer of six-ounce fiberglass with balsa trailing edges as mentioned earlier as well as basswood ribs laser cut to fit each airfoil.

The strength and rigidity of the wings come from the main spar and anti-torsion spar embedded within the wing. For ease of construction and quality consistency it was decided to purchase our spars from a carbon fiber supplier. The size of the main spar was determined using FEA software (CREO) and optimizing the size based on what the expected loads would be. The noteworthy aspect of this portion of the design lay in the method for introducing the spar to the wing and not the ability to place an order from a supplier. Each wing section was carefully positioned in a vertical position, resting on the ribs, while a metal rod was heated using a blowtorch. The rod was then inserted into the foam which ultimately disintegrated around the metal. Speed of insertion and the heat of the rod were experimented with in order to obtain proper circumference and ultimately the optimal fit for the main spar. In similar fashion, the anti-torsion spar holes were made. the anti-torsion spars were designed to be 6 inches in length and allowed to be inserted 3 inches deep into each airfoil for a total of five wing sections across the span. The horizontal and vertical tail sections were also designed and manufactured in this fashion. It is important to note that the center wing section was cut in half and angled on the adjoining edge to allow for the five-degree dihedral. Following this the total mass of wing was centered and permanently attached to the fuselage using composite seam tape. 

Fuselage

    Due to the requirements presented by SLU's AUVSI team, it was important for the fuselage to be easily replicated. To meet this requirement, the fuselage was created in a three step process: creating a plug, mold, and assembly. The plug is an intermediate step used to create the shape of the fuselage and assure a smooth surface to the exterior of the fuselage. From the plug, the mold can be created. In our project the plug was not able to be saved but it is desirable to keep the plug if at all possible. 

    To meet our design of the fuselage, a right and left plug had to be created. The plugs were created from Wick's Aircraft foam. This foam was chosen due to its availability and its capability to be cut by the Vertical Milling Center (VMC) here at Saint Louis University. To begin the creation of the plug, the Haas VMC was prepared. This was done by covering the coolant channels inside the VMC to prevent small particles of foam from collecting in the channel and slowing coolant flow and quality. After the VMC was prepared, the foam block was placed into the VMC and anchored down. The program was then loaded into the VMC and cutting began as can be seen in Figure 35. Periodically, the VMC was paused and a vacuum was used to clean up foam particles left on the plug. This prevented these particles from obstructing the cutter and providing a better cut. 

    Completing the cuts in the VMC, allowed the excess foam to be cut away leaving material for a flange to be added later. The freshly cut plug was then ready to be layered with Bondo. Bondo reinforces the plug and fills the small gaps that are created from the foam. Applying Bondo to the plug was an iterative process that included applying Bondo and sanding it down. This was done until a desired finished was obtained.  

    At this point in the plug creating process, the flange was added to the plug. The flange material was found to be too low. To increase the height in the flange material, pressboard was cut in the shape of the flange material. It then had to be determined whether or not the gel coating, used to create the mold, would react well to the pressboard. A practice piece was prepared and completed. The result can be seen in Figure 36.  As it can be seen, the pressboard came up with the gel coating and resulted in a poor surface to create a part from. This lead to further testing that can be seen in Figure 36 as well. This demonstration shows that the foam and bondo resulted in a smooth surface ideal for creating parts. The results led us to keep the pressboard side down and foam on the top. Bondo was added to the foam flange and sanded down. 

    Once the entire plug was layered with Bondo and sanded to the desired finish, the plug was spray painted with primer. This allowed us to see where there were small imperfections in the plug and sand down the material to obtain an even smoother finish. The plug was then ready to be waxed. Waxing was done by applying the wax to the plug and wiping away excess wax. The wax would then sit for about five minutes and then be polished. This assured that a thick layer of wax laid between the Bondo and the epoxy. This prevented the epoxy from sticking to the Bondo and foam.  

    When the plug for both of the two molds had been waxed with about ten to twelve layers the next step toward creating the molds was to apply a thin layer of release agent to the surface of the plug.  This release agent was a special chemical which dries into a thin film within about five minutes and it was applied directly before the mold process was begun.  With the release agent dry on the plug, it was time to apply the gel coat to the surface of the plug.  With the wax coat and the release agent on the surface of the plug, the gel coat tended to bubble and leave open spots.  It was very important for to inspect every inch of the applied gel coat before it became too stiff and to make sure that the entire plug was covered. 

    With the gel coat applied over the surface of the plug, the next step was to provide the structure of the mold.  This came in the form of fiberglass matte which was placed directly on the now tacky exposed surface of the gel coat.  Two layers of fiberglass were applied to every inch of the exposed gel coat and at the same time a wooden box was also attached to the fiberglass matte to act as a support for the mold once removed from the plug.  This process was done for both of the plugs with the second time proceeding in a much smoother fashion.  Each mold was given at least two days to fully cure before the process of removing it from the plug was begun. 

    After the two days had elapsed since the mold was applied to the plug, the process for removing the mold was initiated.  The idea was to be able to pry the plug from the mold in one piece and thus to be able to reuse the plug to make new molds if necessary.  After trying to carefully pry at the mold with wedges, and not much progress was made, the decision was made to destroy the plug in order to remove it from the mold.  While paying careful attention to the surface of the mold, the foam plug was hacked and shredded out of the mold.  The molds were in very good shape but the plugs were not going to be able to be reused as was the initial plan. Despite that unfortunate happening both the molds' excellent shape was enhanced by a light wet-sand with 1000 grit sand paper and the application of several coats of wax. 

    The fuselage was ready to be laid into the molds and it was decided to start with one and see what the result would appear like.  The four layers of the fuselage were laid into the mold in the order of light fiberglass, carbon fiber, fiberglass, carbon fiber.  This layering was maintained on the future iterations of the fuselage layups which were done because the first one did not turn out satisfactory.  The first side of the fuselage did not turn out as nice as one done in the future and this is because relief cuts in the material was not as highly emphasized on that first iteration and the result was that there were sections that pulled off the mold due to the exaggerated contours of the fuselage design.  With this lesson learned, relief cuts were precut into the material as a preparation before the layup was even begun.  This lead to much more satisfactory results and both sides of the fuselage were laid-up in each respective mold. 

    After removing the sides of the fuselage from their molds, the excess was trimmed off the flanges and the two sides were clamped together using rubber bands.  This allowed for the application of a special 1-inch-wide fiberglass tape which was run around the complete inside and outside of the seam.  This was allowed to dry for about two days and after the rubber bands had been removed, the fuselage was now a single complete piece. 

Landing Gear

    Landing gear for small aircraft can be typically bought from an online store. The landing gear from online stores come in various weight classes. In the design of our aircraft we needed a main gear and nose gear capable of withstanding a 25-pound aircraft. When searching for a both a nose gear and landing gear for this weight class, the height of the two never aligned. The tallest nose gear we found for RC aircraft was roughly 6.5" tall. Whereas the landing gear was about 8.5". This meant we would be short in the front by about 3" with wheels. This meant that we would either have to order custom nose or landing gear, or create our own landing gear. Since the design of our aircraft is meant to last at least five years, it was decided that the best course of action would be to create our own custom landing gear. 

    To create the landing gear, the step by step process recommended by Fabricating Composite Landing Gear by William Koster was used and slightly modified. Similar to the creation of the fuselage, a mold was created. Unlike the fuselage the mold was created from pressboard. The shape of the landing gear was cut into two pieces of pressboard. These two pieces were then offset from one another and bridged together using pressboard and five-minute epoxy. Due to the curvature desired at the corners of the landing gear, the pressboard was covered with a thin layer of Bondo. This was then sanded down to a desired smoothness. The mold was then spray painted with primer. The primer allowed us to visually see small imperfections and sand it down smoother. Now the mold could be prepared for the part. 

    To create the part, it was decided that a combination of unidirectional and bidirectional weave carbon fiber would be used alongside a plastic honeycomb filler. It was found that in many store bought composite landing gear, some form of filling material was used. To assure the composite material would not sink into the holes of the carbon fiber, expandable foam was laid into the honeycomb. Once the honeycomb was filled, the excess foam was sanded flat and cut so that it could form the shape of the landing gear. 

    The first iteration of our mold can be seen in Figure 37. This mold was found to fail after multiple uses. To strengthen the mold, the material was changed from pressboard to aluminum. The aluminum was cut to size and welded together. Excess oil left over from the manufacturing process was cleaned off the mold. Clay was used to fill the gaps in the welded joints. Creating the part was then done as follows. A layer of carbon fiber weave was lain onto the mold and thoroughly soaked with epoxy. Next a layer of uniaxial carbon fiber was placed over the weave. A second layer of carbon weave was added to the top layer. To strengthen the bends in the landing gear, three layers of carbon weave were added at the joints.  The center of the landing gear was strengthened using a plastic honeycomb material. The honeycomb was filled with an expanding foam to prevent epoxy from filling the gaps. Layering in the carbon fiber in the reverse order from above, the part was ready to be vacuum bagged. Each layer was coated in epoxy before additional layers were added. Sealant tape was then applied to the outer surface of the mold. Bagging film was then carefully secured to the sealant tape to ensure there would be no leaks. Finally, a vacuum pump was attached to the vacuum bag and activated.

    After twenty-four hours, the vacuum pump was removed from the bagging film since at that point the majority of the excess epoxy had been removed. The part was then allowed to sit for another 24 hours in air to ensure the epoxy was dry. Once the part was complete excess carbon fiber was removed using a cutting wheel tool. An axle was then added for mounting the wheels. With the wheels in place, the landing gear were finished and ready to be mounted.

Design Recommendations

With the completion of two tests and the resulting undesired ground contact of the aircraft, key areas for improvement were identified. Addressing the likely cause for the undesired ground contact, it is suggested that both the horizontal stabilizer, as well as the elevator be increased in size. It would be possible to utilize the initial design of the tail configuration by scaling up the surface sizes, with some minor adjustment in the control surfaces sizing. Another possible contribution to the elevator issue were two underpowered servos that deflected the control surface. It is suggested to mitigate all sources of failure from the tail configuration that these servos be increased in size, along with the resizing of the surfaces.

The fuselage molds manufactured were based off the initial Dragonfly design which was prone to pitch instability. For the benefit of future fuselage manufacturing, these molds would need to be rebuilt, following the final fuselage design implemented for the second test flight, which featured a fuselage length increase of 14 inches. Another recommendation would be to manufacture these molds utilizing a different process followed than the original manufacturing process which was both time exhaustive, as well as work intensive, and also offered no real benefit over other methods. The original method produced a fiberglass mold from the creation of a male plug, constructed of blue foam and car body filler. Producing the mold resulted in the complete loss of the plug as it deteriorated in the process of separating the two. It is suggested that the future plug be made of layered particle board, which would then be placed into the CNC machine to produce the fuselage shape. This plug should then be covered in epoxy and sanded extensively to produce a smooth finish. This would yield similar results to the foam plug, however, the wood plug would be much more resilient in the separation process, as well as less time exhaustive to produce.

Additional recommendations would revolve around the reduction in weight of the aircraft through means of different material choices for connection methods, namely the pine blocks used to connect the tail booms to the wing could be switched out with balsa blocks. Vacuum bagging processes should also be completed in separate vacuum bags, as it was noted that some wings that were vacuum bagged together were heavier than those vacuumed separately. 

Gantt Chart

Mission Profile

Design Configuration

Team Contact Information:

Colin Klemstein: Manufacturing Lead, Structures cklemste@slu.edu

Tom Oven: Stability and Control Lead, Recorder toven@slu.edu

Peter Rackovan: Structures Lead, Manufacturing prackova@slu.edu

Chris Russell: Team Captain, Propulsion Lead, Aerodynamics, Archivist crusse18@slu.edu

James Shields: Aerodynamics Lead, Propulsion, Webmaster, jshield9@slu.edu