Task 4

Aircraft characteristics : Theory vs. Experimental data

Shane Hills, Jae Hwan Choi, Jean-claude Angles, and Matthew Berk

1. Theoretical Flight performance:

Jae Hwan Choi, Shane Hills, and Jean-claude Angles

AVL was used to estimate aerodynamic performance of the Spaero2. Major differences between the Spaero1 and the Spaero2 are aspect ratio of wing and fuselage wing, dihedral angle in the outboard section.

The AG 35 airfoil stalls when its lift coefficient exceeds 1.235. From the AVL analysis, about 50% of the main wing section has a lift coefficient bigger than 1.235. Therefore, C_Lmax = 1.224. The overall weight and reference area of the Spaero 2 are

W = 0.668kg * 9.81m/s² = 6.5531N

S = 0.19528m²

which yields a stall speed

Note that Spaero2 shows better lift coefficients for whole range of angle of attack.

2) Drag polar

From above drag polar, it can be seen that Spaero2 exhibits better aerodynamic performance compared to the former design in terms of lower drag for high speeds.

3) Lift to drag ratio vs C_L

Lift to drag ratio of Spaero2 increased for speed range from 9m/s to 17m/s compared to the Spaero1.

4) Required power in level flight vs airspeed / Required power in climb vs climb rate

Using the L/D data from AVL, we can use the propulsive model we've developed since the previous problem set.

From the above table, turning radius of 5.14 m can be achieved with 80.09 degrees of bank angle, theoretically. However, such high value of bank angle is unrealistic. Therefore, turning radius is determined as a function of airspeed with the bank angle of 45 degrees.

Note that all lift coefficients are below the C_Lmax. Therefore, minimum turning radius is about 6.5 m.

Ⅱ. Flight dynamics

1) Stability derivatives

These are some selected stability derivatives for the Spaero2. Note that C_Lα is larger than the Spaero1 which agrees with the aerodynamic performance analysis.

2) Frequencies and damping ratios

The Spaero2 has lower damping ratio for the Phugoid mode compared to the Spaero1.

- Optimal cruise condition : V = 12m/s

- Optimal climb rate : V = 12m/s, AOA = 4.75deg, Vz = 1m/s

Ⅲ. Weight and CG statements

The driving factor in improving on Spaero 1 was to shave as much weight from the airframe as possible in order to provide as much ballast to the mission planning team as possible. To that end, without ballast (i.e. including airframe, avionics, and all other components necessary for flight, but without water weight), Spaero 1 measured approximately 630 g. Considering the somewhat chunky design of the Spaero 1 prototype (basswood avionics board, larger than necessary fuselage dimensions, etc.), the design team expected to be able to shave off approximately 100 g of weight in iterating on the design.

In order to estimate the weight of Spaero 2, the fuselage, boom, and avionics were modeled and assembled in SolidWorks in order to gain an estimate of the total weight and CG location. The main wing, vertical, and horizontal tail were modeled in DevWing, a CAD modeling software specifically for designing spar and rib-type model wings. Unfortunately, DevWing is not able to return an estimated weight based on the density of balsa, and is not importable into SolidWorks, so theoretical weight estimates were not achievable for the airfoil components of the aircraft. In order to get an estimate of our constructed airfoil weights, wings with a similar construction approach (i.e. balsa D-box design of this type) were analyzed. At a total planform area of 525 in², the Allegro-E Lite main wing weighs 269 g, corresponding to 794.21 g/m². Given that the planform area of Spaero 2 is .195 m², the weight is estimated to be 155 g. Similar estimates for the vertical and horizontal tail are made.

The weights of the avionics were all manually overridden in SolidWorks, such that the total estimated weight of the aircraft is the result of true avionics and boom weights, estimates for the fuselage in SolidWorks from the default balsa density, and airfoil weights outlined above. The total estimated weight was 535 g. In considering the underestimation of weight from Spaero 1 that was the result of neglecting glue, covering, and other miscellaneous weights, the error on the weight estimate is approximately 5%.

Furthermore, the arrangement of avionics was done by approximately splitting the weight evenly across the longitudinal center line of the aircraft. By arranging the components appropriately in the SolidWorks assembly model the CG location was determined to be approximately 15.8 cm from the fuselage tip.

2. Experimental Flight Performance:

Matthew Berk

Various flight tasks were performed using the Spaero2 and several flight trials were conducted. Unfortunately, on the only low wind day available to fly after completing the model, we had a firmware issue fail to log data. Thus the subsequent trials gave useful (but noisy) data which must be taken with at least a 15% error margin. Hopes are high for another good weather day in order to finalize the data to obtain control trims and other useful information for mission planning.

The plots of the obtained data are shown below:

Figure 1: For level flight power vs velocity the curve seems to make sense. The upward trend follows quite closely with the predicted model

Figure 2: While the L/D generally follows what was predicted, the L/D increase at high speed/low CL does not make much sense at all. This is likely the result of some of the aforementioned error in the data.

Figure 3: CL vs alpha is slightly nonlinear, but mostly appears to be about consistent with what one would expect.

Figure 4: The drag polar again seems fairly reasonable, though it does not match the theoretical values exactly.

Figure 5: In the climb rate vs power analysis it is difficult to pull out a detailed plot because there is very little excess power available. The one above came under automated control and was the closest to a consistent climb rate as could be obtained.

Figure 6: An aggressive phugoid mode was observed in flight, with a period of about 4 seconds (0.25Hz) at 10m/s. This is much faster than predicted, but could be accounted for by low Re dynamic effects which AVL has not accounted for. The other eigenfrequencies were not able to be accurately determined from experimental data. This phugoid mode made determining experimental values difficult in many cases.

Note that experimental turning performance data can be found in Task 5.

3. Comparisons

Matthew Berk, Jae Hwan Choi, Shane Hills, and Jean-claude Angles

Below is a table comparing a variety of expected and final physical measurements for Spaero 2, including weight, area, and positions measurements. Note the dashes indicate negligibly different values.

In most aspects then, particularly in the dimensions of the aircraft, the theoretically expected measurements agreed very closely with the measured. This is expected considering the precision of our manufacturing process, which is outlined below.

One of the more deviating measurements was in the actual weight of the airfoils. The construction of the airfoils for Spaero 2 were lighter than expected from comparison to the Allegro glider aircraft, which is easily attributed to rib spacing, spar construction, and webbing constructions variations.

4. Design and Manufacturing Process

Shane Hills, Matthew Berk, Jean-Claude Angles, Jae Hwan Choi

As was mentioned above, the main driving factor in revamping the design of Spaero 1 was in reducing the total airframe weight and improving the aerodynamic performance. The following is a breakdown of Spaero 2's components and the improvements made from Spaero 1. Note that a large amount of information is available in Problem Set 2, Task 4, where much of the design philosophy has been covered. In summary, our general approach was to take inspiration from the RC glider community and construct an aircraft with high-mounted, high aspect ratio wings and "suspended" fuselage. By constructing the fuselage in an airfoil shape it is possible to house the many avionics components in a shape that is aerodynamically efficient. In iterating on the design it was possible to reduce the total weight of the aircraft by approximately 100 g, and improving aerodynamic performance considerably.

I. Main Wing

One of the primary drawbacks of Spaero 1 were the compromises made on the main wing in pursuit of rapid prototyping for initial flight. Spaero 1 featured a foam wing of approximately the same area, but with much smaller aspect ratio. In order to cut the foam wings in a single section (improving structural rigidity in the span-wise direction) it was necessary to reduce the span and increase the aspect ratio, increasing drag and reducing overall aerodynamic efficiency. By moving to a balsa wing with smooth plastic heat wrap for Spaero 2 it was possible to increase the aspect ratio and decrease the effects of parasitic drag terms contributed by the roughness of the foam.

Spaero 2 featured the same airfoil (AG 35) as Spaero 1, but at much larger aspect ratio (10 compared to ~7). The larger aspect ratio was made possible by constructing the main wing using a D-box design that employs a more common spar and rib design. Using materials like 1/8in balsa and carbon rods provided the ability to create highly structurally rigid airfoils at a fraction of the weight of a foam wing of the same size. By using a convenient modeling program called DevWing the design team was able to rapidly and effectively design all components of the main wing, produce templates for laser cutting, and assemble the wing using the materials mentioned. DevWing prints out a convenient template over which to lay the laser cut ribs for easy assembly. Below is a model view of the half span DevWing balsa wing design.

Spaero 2 also moved the dihedral outboard to a second wing section, rather than implementing a dihedral angle on the entire wing section. The inner wing section A consisted of a flat section that made attaching to the fuselage considerably easier. The second wing section B featured a dihedral angle of 5 degrees. Furthermore, the ailerons were located on the outboard dihedral section to improve control surface effectiveness. The control surface above is depicted in section B in blue. A picture of the main wing and horizontal tail construction is below.

More on the assembly can be found below.

II. Horizontal and Vertical Tail

Spaero 1 featured solid 1/4'’ balsa sections for both the horizontal and vertical tail components. Having simply sanded down the leading edges to create a slightly more airfoil-like shape, in the initial prototype phase the tail existed and was sized exclusively to provide stability to the aircraft, once again with the priority being rapid prototyping. Spaero 2 provided the opportunity to greatly improve on the tail design by using a similar approach to the main wing sections. By eliminating the flat plate balsa tail and using a symmetric airfoil constructed using balsa ribs, carbon spars, and plastic shrink skin, the drag could be greatly reduced. Taper was also added to the tail to improve aerodynamic efficiency.

As was the case with the main wings, the ribs, tip coverings , and shear webs were constructed using balsa, while the spars were done using a double carbon rod. The control surface above is depicted in red.

III. Fuselage and Tail Boom

The original Spaero 1 fuselage was conservatively sized in order to ensure that there was ample room to house all the necessary avionics components. The original fuselage design used a Clark Y airfoil, mainly for it's common use at low Reynolds number and flat bottom, making manufacturing a smoother process. However, considering the unused space that remained in the fuselage, it was decided that drastically downsizing the fuselage would result in large weight savings. However, by reducing the size of the fuselage airfoil the aspect ratio decreases considerably, reducing the effective lift and greatly increasing draft. As a result, it was determined that the savings in weight that could be achieved by drastically reducing the size of the fuselage was preferable to the small component of lift achieved by the Clark Y airfoil, and the fuselage airfoil shape was revamped to a NACA 0015, providing no lift at 0 AoA, but having excellent drag qualities and large thickness for housing the avionics.

The actual construction of the fuselage was similar to that outlined in Problem Set 2, Task 4 for Spaero 1. The tail boom runs through the center of the fuselage and is sandwiched on either side by balsa ribs. A balsa plate reinforced with carbon strips connects the outer and inner fuselage ribs and provides a surface to mount the electronics. Finally, the motor is mounted to the end of the carbon boom and the fuselage is covered with the same plastic shrink skin that is used on the airfoil sections. To provide accessibility to the electronics a simple thin sheet of balsa is taped in place as a "hatch". Images of the balsa fuselage CAD model and construction are shown below.

In the above image it is possible to see the pylons that attach the main wing to the aircraft. These pylons are glued into the fuselage above the boom and serve as the attachment points for the main wing. Since the center section of the wing is flat, nylon screws are used to bolt the center section of the wings to the pylons. In the event of a hard landing the nylon screws shear and save structural components of the airframe. There is also space in front, behind, and between the pylons that is used as space for the water ballast. Below are photos of the fuselage construction

IV. Assembly

Final assembly of Spaero 2 took place over multiple days. CAD rendering of the final assembly are shown below. Note that since the wings were modeled in DevWing, only a crude model that did not include the dihedral angles were included in the SolidWorks model, since only the weight distribution was required to obtain a CG location estimate. Below are 3D renderings, multiview drawings, and true photos of the construction.

Assembly of the fuselage and main wing is discussed above. The tail is attached to the boom via a system of glue and carbon rods that run through the vertical and horizontal tail and boom.

Flight Performance Comparison

Matthew Berk

In general the flight performance seemed to meet the values expected by theory. While there is a limited amount of experimental data due to time constraints and some software issues, that which exists does indicate that reasonable predictions were made. In particular, power usage, CL vs alpha, and L/D give quite good correlations to the predicted values. The drag polar indicates slightly more drag at intermediate values, but it is within the 15% error tolerance which has been previously mentioned. Turn performance is quite good if one looks at Task 5, and though more aggressive turns than 45 degrees have not yet been attempted, the values at 45 degrees seem to indicate a very good correlation to the predicted values, hopefully meaning that sub-10m steady turns are possible.

Subjectively, the flight characteristics of the model are that it is very stable and friendly to fly, with a gentle stall. It does have a strong phugoid mode which may be partially due to a bit of overstability. The phugoid mode is not difficult to fly with however it does have a tendency to make certain data difficult to obtain accurately. Turns are easy and stable, and the model seems to have solid spiral stability preventing it from 'tucking in' to turns in a bad way. Some experimentation is being done with intentionally tip stalling turns in order to have low radius turns which may aid the mission algorithm - the altitude loss associated with this maneuver seems like it may be worth it.

Overall flight performance is more than satisfactory and in terms of flight characteristics there is little that seems to need improvement at this time.