Final Design

    The final design was composed of three major components: electronics enclosures, composite airframe, and the float system. The configuration of the multicopter was identical to the current hexacopter model with six motors. The composite airframe consisted of the six carbon fiber tubes as arms and two carbon fiber plates. The above figure shows the CAD image of the entire structure of the multicopter. The float was oriented such that two cylinders mounted to the four landing gears parallel to the direction arm, which was shown in green. The upper and lower enclosures were attached to the center plates of the composite airframe.    

Water Protection

    In terms of water protection, the electronic speed controllers (ESCs) were coated with thermally conductive epoxy as shown in the figure below. With that, it was possible to mount the ESCs externally to keep them cooled and easily accessible. Furthermore, the enclosures protect the autopilot, battery, power distribution board, radio and all other electronics. For the brushless motors, a WD40 treatment was applied.

Thermally potted ESC

Key Components

1.Enclosure for Electronics

    The design of the upper and lower enclosures was chosen to keep the overall weight to a minimum as well as keeping it aesthetically appealing. The upper and lower enclosures were part of the central structure and provide the needed water resistance to protect the delicate electronics of the multicopter. The enclosures were made of carbon fiber which was made off a fiberglass mold of the rapid prototype version of the enclosures. The only wiring that would exit the lower enclosure lead to the ESCs which were sealed using a soft closed cell foam sandwich between the enclosure and the lower plate. The designed enclosures fulfill one of the primary requirements of the craft being recoverable from the surface of the ocean and experiencing limited salt water damage as the electronics are protected from the elements. The plates had a pass through for wires to connect with the electronics in the upper and lower housing.

Lower (left) and Upper (right) enclosure rapid prototypes

Carbon plate with pass-through

    Since the enclosures cover the electronics and access to the battery, easy access was a definite consideration when designing this part. To provide quick and tool free access the enclosures were held on using quick release fasteners. By using a tool free fastener it reduced the number of tools needed to work on the craft and make the overall process easier and less time consuming.   

Quick Release Fasteners

2.Composite Airframe

    The structure of the hexacopter is vital to its primary functionality: being flight worthy.  As such, the structure must ensure that the motors are held in place relative to one another, and there must be a central area to mount the electronics.  With a star patterned shape, the arms must be capable of sustaining any thrust loads from the propellers or loads during landing, and the center section has to hold all of the arms in place. 

   In order to fulfill the first overall objective of the project, decrease the weight of the multicopter through the use of composite materials, carbon fiber was utilized throughout the entire airframe structure. These carbon fiber parts mimic the design concept utilized by 3DR, a modular structure of arms mounting to a center plate system.  The original 3DR Hexa-B multicopter was composed of a 2mm thick, G10 composite, dual center plate system with six 19.5mm x 240mm, 1 mm thick square tube aluminum arms. The final design of our structure used 1.6mm thick carbon fiber plates for the dual center plate system and six 8mm x 240mm, 1mm thick square unidirectional carbon fiber tube arms.

Composite arms (left) and plates (right)

A primary concern when weight optimizing the composite structure was the sizing of the composite tube arms.  By constraining the carbon arm material selection to off-the-shelf, square cross-section carbon tubes, the largest readily available tube was the 8mm square, 1mm thick tube.  To ensure that this would provide sufficient strength for the multicopter’s design, a Solidworks Simulation was performed on the arms.  To define the material, composite shell elements were used with all fibers aligned along the axis of the arm since the chosen tubes were unidirectional.  This analysis consisted of two load cases, one to analyze ascent, or when the motors provide maximum thrust, and one for landing, where a free-fall with only one arm contacting the ground first similar to the landing from an inexperienced operator.

            To define the load on the arm due to thrust, the combination of weight of the structure and drag while flying were used to calculate the vertical force.  Since the thrust is due to a spinning propeller, the reaction torque due to the drag of the propeller was also considered.  This combined loading was applied at both the motor clamp mount faces and the through holes supporting the fasteners.  Finally, fixed geometry was defined at the surfaces where the center plates clamp the arm in place.  These constraints can be seen in the following Figure.

Annotated Image of FEA simulation

For the load case to simulate the multicopter ascending, the maximum load along the axis of the fiber was 114.4 MPa compared to the ultimate tensile strength of 1500 MPa, while the maximum stress 90o to the fiber’s orientation was 33.9 MPa compared to the UTS of 50 MPa.  This shows that in the weakest orientation, there is still a factor of safety of nearly 1.5.  The simulation for an extreme landing yielded similar results, with the primary difference being a maximum stress 0o to the fiber’s axis of 608.2 MPa.  This analysis gave preliminary confidence in the selection of the arm size and was followed with bench testing of the composite arm to verify its ability to withstand a constant fluctuating load similar to that which is seen during flight.

    In comparison the arms were considerably smaller than the original design. The weight savings of the composite arms was considerable, a set of 6 composite arms weighed in at 52g compared to 6 aluminum arms weighing 258g. In addition the smaller arms also minimized the surface area thus decreasing the potential drag experienced by the craft. Since the arms of the final design were considerably smaller than the previous design the motors had to be mounted utilizing  cross shaped motor mounts which provided the necessary mounting area for the motors.

Comparison of 1 Aluminum arm to 6 Carbon Fiber arms

Cross shaped motor mount

It was determined that switching to a completely composite frame would offer significant weight savings. By reducing as much weight in the structure as possible it would make it possible to add the other design features without surpassing the previous crafts overall weight. Also by utilizing the same motor orientation and dimensions of the current model flight capabilities and response would remain unaltered.

3. Float System

   The purpose of the float system was to allow the multicopter to land on the surface of the water. With water landings, the copter can be easily recovered while experiencing minimal salt water submergence. There were several functional requirements for the float system. First, the floats should not significantly affect the aerodynamics which would minimize the effects of power consumption of the current model. In addition, the floats should not add significant weight to the multicopter.

    For the final design, polyethylene closed-cell foam tube was used with a diameter of 7.62 cm (3 inches). There were many options but this was commercially available at a low cost. A simple buoyancy analysis was performed to find out the volume of float materials needed for the multicopter assuming the float materials will be half submerged, assuming the Factor of Safety to be 2. Based on the definition of buoyancy force, the following governing equation was obtained:

which then derived to be:

from that the expression for the total volume of the float materials is:

Assuming the float materials to be closed-cell polyethylene with a density of 16.02kg/m3 and the salt water density to be 1027kg/m3, the final result came out to be 3600 cm3.

    With the total required volume of floats and the given diameter of the polyethylene tube, the total length of the tube was calculated to be 78.94 cm. In terms of the orientation of the floats, we had two cylinders attached to the landing gear in parallel shown in Figure 10, making each cylinder half of the total length of the tube. The advantages of this orientation are to maximize stability and surface area.

Float Orientation

    To further predict the behavior of the float design, we did a displacement analysis to ensure that the multicopter would not sink with the floats. In this case, we assumed the float to be in simple harmonic motion once it contacted the water surface and the float would hit the water surface at an angle of 90 degrees. The equation of motion was derived to be:

The variable of this model was the vertical displacement of the horizontal cylinders. Implementing the equation of the cylinder volume, the second-second ODE came out to be:

After transforming into first-order ODE, the solution was obtained using MATLAB with the measured parameters. In this case, the mass included both the multicopter and floats, which was 1.859 kg. Figure 14 shows the displacement vs. time graph.

Float Displacement

As it was seen, the float goes down to about 7 cm, compared to the depth of the tube of 7.62 cm. This was an overestimate of the prototype due to negligible surface tension of water and its properties by modeling a simple harmonic motion. Air resistance was included to compensate the loss of energy during the harmonic motion.

Finally, several drop tests were performed using the multicopter without electronics but with added weights to compensate for the electronics weight. The multicopter was dropped at different heights up to 1.5 m and angles up to 60 degrees. For results, there was no significant submergence of the multicopter, referring to Figure 13. Because of that, we assumed that the multicopter would land on water at an angle less than 60 degrees. In other words, this float system allowed “safe” landing with angle of approach upwards of 60 degrees. 

 Aerodynamic effect of floats

    Then a test flight was performed done with and without the float system to examine if the aerodynamic effect of the added float system significantly lowered the performance of the current model. In addition, it should be noted that the test without floats would not have equal mass to that with floats. To ensure that the primary change for the multicopter was the geometry of the floats, a ballast was added to the multicopter when the floats were removed. The current draw from the battery is shown in Figure 22, which represents an increase in power of 3.3% with floats. The increase was significant but acceptable with the battery it carried. The battery would provide sufficient amount of power for one mission even with the floats.

Float test flight data with and without floats

Prototype Performance

Test Conditions

    The basic flight tests were performed on an overcast spring day with slight breeze (2-3 m/s) along Warren Mall at University of California San Diego. Multicopter is complete with upper and lower composite enclosures, fitted with 7800mah LiPo battery and weighing a total of 1.859kg with float system attached (1.780 kg w/o floats). Test flight of craft is being performed by an experienced pilot.

    The water resistance testing was performed at Scripps Pier in La Jolla. The multicopter is in the same configuration as the basic flight tests with the addition of a GoPro Hero 3 Black (73g) camera mounted to the lower portion of the craft bring the total weight of the craft to 1.932kg.

Results

    After calibration and syncing of the ArduPilot with the ground station the preliminary test flight was conducted to verify the craft was flight capable. The craft was flown for 2 minutes without the float system and was successful. Upon landing the float system was attached to test any adverse effects of the floats on the multicopter. Again the craft was flown for 2 minutes and despite the addition of the floats did not exhibit any adverse behavior during flight. Each flight tested the basic function of the craft up to an altitude of approximately 80-90 meters and performed flawlessly. 

In summary, the multicopter's capabilities were determined to meet or exceed that which was defined by SIO:

Water capabilities:

The multicopter proved able to successfully land on the ocean surface in flight tests performed at the Scripps Pier. Furthermore, all systems remained operational, showing no influential water damage to the aircraft.

Flight capabilities:

By loading the composite arms in excess of what is seen during flight, the structural integrity was proven. The flight time is also increased through a 5% decrease in overall mass.  

Final Product