Propeller Design.doc

Report as inappropriate Your report has been received

Shared by: wangnuanzg

Tags

Stats

views:: 44
posted:: 5/23/2012
language:: English
pages:: 14

Public Domain

Document Sample

Background Work:
An initial project that had to be completed early on in the project was a complete literature review
and cataloging of as many Micro Air Vehicle Resources as possible. This was accomplished by gathering
past Purdue University Research Reports into MAVs as well as those available from other Universities
online. Next information on power sources had to be gathered and cataloged to compare the energy density
of various types of energy systems to allow for an unbiased analysis of their respective effectiveness. This
goal was partially completed by gathering data about various battery system chemistries and their
respective characteristics, as found in Appendix A. Finally a comprehensive listing of many MAV
resources and projects currently available on the Internet was made and categorized by the area of MAV
technology or research they represent such as engines, controllers, batteries, propellers, etc. This data can
be found attached in Appendix B. The following materials were collected in this process:

 A binder containing many commercial batteries, including technical data.
 A comprehensive listing of many compressed gasses and their properties.
 A list of URLs for MAV related areas, with descriptions and broken down by area (Appendix B).
 A table showing a comparison between various battery technologies (Appendix A).
 A binder of past Purdue MAV research project papers.
 A binder containing other Universities’ and private companies’ MAV research papers.

After reading various project research papers on MAV design and testing, including past Purdue
projects, there are several problems that the researchers have had in common. One is the problem of
excessive torque caused by the engine, since the vehicles are so small and the propellers spin at such high
RPM an excessive amount of engine torque is developed. One area worth researching to combat this would
be the use of multiple engines with counter-rotating props. Another problem commonly run into is the
power source. While both ICE and electric have been used there hasn’t been much experimentation beyond
these power sources. Research into flywheels, fuel cells, and other power sources would be advised.
Another important need is for a highly efficient propeller. The design and manufacture of a highly efficient
propeller would ensure good power translation from the motor.

Propeller Design: Electric:
The propeller design took into account a 10 m/s flight speed and a need to generate .3 Newton of
Thrust to allow the aircraft to overcome drag, with a diameter of 3.5 inches. Taking this into account two
propellers were optimized each to operate at three different RPM settings. The first propeller was designed
to optimize the efficiency at 15,000 RPM, and a final at 20,000 RPM. The following data was collected for
the three designs, which were then used to generate the AutoCAD files for Surf CAM and the mold milling
procedure.

 Number of Blades 2
Diameter 0.089 [m]
Velocity 10.00 [m/s]
Rotational Speed 15000.0 [1/min]
Power absorbed 4.116 [W]
Thrust delivered 0.300 [N]
Power Coefficient Pc 1.0826 [-]
Cp 0.0387 [-]
Thrust Coefficient Tc 0.7891 [-]
Ct 0.0627 [-]
Advance Ratio Lambda 0.14 [-]
v/nD 0.45 [-]
Efficiency Eta 72.89 [%]
3.50 x 2.19 [inch]
Propeller size
8.9 x 5.6 [cm]
Airfoil at root E 193, Re=100'000
Airfoil at tip E 193, Re=100'000
15,000 RPM Optimized Propeller
Isometric view of the 15,000-RPM Propeller Blade in cross-section form (root is at left, tip at right)

 Number of Blades 2
Diameter 0.089 [m]
Velocity 10.00 [m/s]
Rotational Speed 20000.0 [1/min]
Power absorbed 4.195 [W]
Thrust delivered 0.300 [N]
1.1034
Power Coefficient Pc [-]
0.0167
Cp [-]
0.7891
Thrust Coefficient Tc [-]
0.0353
Ct [-]
Advance Ratio Lambda 0.11 [-]
0.34
v/nD [-]
Efficiency Eta 71.51 [%]
3.50 x 1.69 [inch]
Propeller size
8.9 x 4.3 [cm]
Airfoil at root E 193, Re=100'000
Airfoil at tip E 193, Re=100'000
20,000 RPM Optimized Propeller
Isometric view of the 20,000-RPM Propeller Blade in cross-section form (root is at left, tip at right)
Propeller Design: ICE:
Next an efficient propeller was to be designed for use with the Cox .010 internal combustion
engine. For this engine flight speed to be attained was approximately 15 m/s, absorbing 20 Watts of power,
at 20,000 RPM. After using these variables the following information was generated.

 Number of Blades 2
Diameter 0.089 [m]
Velocity 15.00 [m/s]
Rotational Speed 20000.0 [1/min]
Power absorbed 20.0 [W]
Thrust delivered 0.926 [N]
Power Coefficient Pc 1.5587 [-]
Cp 0.0794 [-]
Thrust Coefficient Tc 1.0827 [-]
Ct 0.1089 [-]
Advance Ratio Lambda 0.16 [-]
v/nD 0.51 [-]
Efficiency Eta 69.47 [%]
3.50 x 2.60 [inch]
Propeller size
8.9 x 6.6 [cm]
Airfoil at root E 193, Re=100'000
Airfoil at tip E 193, Re=100'000
Isometric view of the ICE Propeller Blade in cross-section form (root is at left, tip at right).

All of the propeller designs were contingent upon the use of a cambered low Reynolds Number
airfoil design, which caused the Eppler 193 airfoil to be chosen. It is a 10.22% thick airfoil, which is
designed to operate at low Reynolds Number of approximately 100,000 to 300,000. Below is a picture of
the E193 airfoil, which was used to generate the cross-section of the various propellers.
Eppler 193 Airfoil Cross-Section used for Propellers.

Propeller Fabrication:

Before the fabrication of the propeller designs was to precede the airfoil and the radial
distributions of the propeller had to be used to create AutoCAD files. Taking the cross-sectional data of the
Eppler 193 airfoil, obtained at UIUC Airfoil Data Site (http://amber.aae.uiuc.edu/~m-selig/ads.html), the
data was used to create a scalable cross-sectional representation of the propeller blade. This was then scaled
at each radial position to the correct chord length and rotated to the correct Beta angle at that radial station.
This created a single blade element for the each of the 2 blade designs. Next the designs were saved in dxf
format compatible with the Surf CAM software to allow for milling out of a material of our choosing.
The actual fabrication process of the propeller would be a difficult task and would have to be a
careful undertaking. At first it was believed that 2 molds could be made of each propeller that would be
negatives of the top and bottom shapes. When fitted together each of these pairs of molds would have a
cavity, which would exactly form the propeller shape. These molds would have thin strips of a light carbon
fiber layer in them and then be filled with resin to create the total airfoil shape. When it came to trying to
use this process though there were several problems including the ability of Surf CAM to convert the model
to two separate negative molds so for time constraints this approach was abandoned.
The second option that we had would be to create the propeller completely out of the material that
the mold would be milled out of. The CNC machine would cut the outer surface of one side of the prop into
the material. This would then be turned over and the CNC machinery would proceed to mill out the
opposite side of the propeller, leaving us with a 3 dimensional propeller of the material of our choosing.
The propeller material would have to be light yet durable and also be able to be milled into airfoil shapes
with a minimum thickness at some points of only a millimeter or less for this approach. Several materials
were tried and had varying degrees of success.
The first material chosen was a light fiberboard, which could not be cut into thin enough cross
sections without the mill breaking the tip of the blade off. Next acrylic was chosen for its durability, easy
cutting, and low density. This material too seemed to have problems breaking when the mill was cutting the
thin cross sections. Finally aluminum was chosen because of its relatively low density and the ability to cut
it very thin and still have it retain shape and not break. This too delivered a propeller that was prone to
breaking at one of the blades. This indicated a problem with the CNC machinery or Surf CAM and the way
in which it was milling the material. One blade was grossly thicker than the other blade, which was prone
to breaking. After examining the model and making adjustments the model was redesigned to include a
slightly thicker, 20 percent thick, airfoil right at the root, NACA 66-021 shown below, of the propeller
blade to keep it securely fastened to the small hub machined with it. Acrylic was then chosen again as the
material of choice for this prototype blade. The propeller was successfully machined out of this material
and was then balanced by taking material off of the blades until they were the same weight.

Custom Propellers: At Left Aluminum Prototype, at Right the Final Version from Acrylic

NACA 66-021 Airfoil Cross-section used for root of propeller blades.

Propeller Results Analysis:

After fabrication and testing of the 4 propellers for the electric engine the data had to be taken and
compared to the theoretical curves generated for the custom propeller. This testing data was to be used to
calculate the advance ratio versus coefficient of pressure, Thrust, and propeller efficiency, by using the
following formulas.
v
Advance Ratio =
n*D
T
Ct 
rho * n^ 2 * D ^ 4

P
Cp =
rho * n^3 * D^5

Ct
Eta = AdvanceRatio *
Cp

v Velocity (m/s)
T Thrust (N)
P Power (Watts)
rho Density (kg/m^3)
D Diameter (meters)
n RPS (1/sec)
The resulting curves generated for each plot are shown next versus the theoretical custom propeller curves,
in black, and show some significant results. It should be noted that no data was able to be collected between
0 and approximately 1.45 m/s as 1.45 m/s was the lowest possible wind tunnel speed we could achieve, and
wind tunnel speed was what was used to control Advance Ratio thus the graphs have gaps at low advance
ratios until static conditions were added in.

Coefficient of Power versus Advance Ratio for all Propellers

The custom propeller above had a significantly higher coefficient of power versus the B2 propeller, which
was specifically designed for use with the Wattage B2 motor used during testing. Yet it also had a
considerably smaller peak Cp than either the U80 or K&P propellers, which were tested. Peak coefficient
of power for the custom propeller was around .17 at an advance ratio of .54, while peak Cp for the U80
hovered near .048 with the K&P around .032. More significant was the fact that it was so far below the
expected theoretical results generated for it that had a peak Cp at and advance ratio of .3 of approximately
.048. These results were unexpected but did follow the trend predicted by the theoretical curve and thus
should help in the future to predict actual performance from propellers designed using the Adkins method
on Martin Hepperle’s Internet site.
Next the thrust and coefficient of thrust generated by each of the propellers is to be analyzed. The
goal when designing the electric motor propeller was to get roughly .3 Newtons of thrust out of the design,
which would allow the MAV to fly. The actual peak thrust generated by each of the propellers is in the
table below. As can be seen utilizing a single cell Lithium battery to power the B2 engine the maximum
thrust that could be gained would be approximately 15.56 N, well below what we would need to fly. Thus
further experiments should be conducted using more batteries to produce more thrust using the current B2
engine.

Propeller Maximum Thrust (N)
U80 15.56
K&P 8.97
B2 6.84
Custom 8.53
Maximum Thrust for each Propeller Using a Single Cell Lithium Battery

Coefficient of Thrust versus Advance Ratio for all Propellers
Again it can be seen in the above graph that there was a great deviation between the expected CT
curve for the custom propeller and that which it actually generated. Only at higher advance ratios of greater
than .62 did the actual results meet or exceed the expected results for the custom propeller. Again the U80
and K&P propellers exceeded the results achieved by the custom propeller. Also the Custom propeller
exceeded the results of the B2 propeller designed for this engine. Next the data for the propellers was used
to generate a propeller efficiency graph versus the advance ratio, the result are presented in the graph
below.
Propeller Efficiency, Eta, versus Advance Ratio for all Propellers
Once again the experimental results of the custom propeller differ greatly from the theoretical
curve generated earlier. AT a peak efficiency of only 40 percent it is approximately half as efficient as the
maximum 79 percent of the theoretical curve. Yet it does off some advantages over the other propellers,
especially in that it is efficient longer than both the U80 and B2 at developing power into thrust at higher
advance ratios. This means that, in the case of our wind tunnel testing, when flight speed is increased the
custom propeller continues to generate thrust while the B2 and U80 do not. This can be useful since to
generate enough lift the engine needs to not only generate enough thrust but the MAV must fly at a high
enough speed while doing it to generate enough lift to keep the MAV in the air.
Finally the Cox .010 internal combustion engine was tested in the wind tunnel using its stock
propeller to develop Cp, Ct, and Eta curves for it. Unfortunately the custom propeller designed for use with
this engine was not ready at time of testing so data for it is not included. Thus these graphs are included
purely to give a comparison versus the electric engine and its performance.

Cp versus Advance Ratio for the Cox Stock Propeller versus the Theoretical Custom Curve
Ct versus Advance Ratio for the Cox Stock Propeller versus the Theoretical Custom Curve

Ct versus Advance Ratio for the Cox Stock Propeller versus the Theoretical Custom Curve

The previous curves seem incomplete beyond and advance ratio of .43 due to the fact that advance
ratio was controlled by wind tunnel speed and a researcher had to sit within the tunnel constantly refueling
the engine as data was taken. Thus beyond a 19m/s tunnel velocity was a bit excessive. This lack of data at
higher speeds can be overcome in the future through the fabrication of a larger external fuel tank for the
Cox engine so that longer running times can be achieved. Just looking at the data of the stock propeller
versus that of the custom’s theoretical shows a large gap, but this may not have held up during actual
experimental testing of the custom designed propeller as before with the electric engine.
Several conclusions can be drawn from the results obtained above. Since the custom propeller did
not come close to matching the expected result it might be inferred that the theoretical results generated
were incorrect. While the trends presented by these theoretical curves are indeed followed fairly well by the
experimental results the actual custom propeller curves do not come close to the values generated in the
curves. When designing this propeller it was chosen to have approximately 4.2 Watts of power absorbed by
the propeller. This would have been the equivalent of 2 single cell Lithium batteries in series. In testing it
was found that this would overheat the engine too much so only static data was taken for this power setting
as it could not be kept cool long enough for dynamic testing. Thus only a single cell Lithium battery was
used for testing cutting the power provided to the cell approximately in half. This reduction in power might
account for quite a bit of the loss in thrust and power that the custom prop had versus its theoretical curve.
Steps must be taken in future projects to allow cooling of the engine both during testing and in flight to
have the required levels of power.

Engine Heating:

Engine overheating at the high voltages experienced for such small electric engines will be a
problem, thus ways to cool then engines had to be invented and fabricated. Two ideas advanced for this
were a radial heat sink sleeve, which would fit around the airfoil and allow for greater surface area to
conduct the heat away from the motor. The second method was the use of a perspiration system using a
small reservoir of Ethanol or other such liquid along with a permeable membrane to soak the liquid up and
conduct it around the engine casing for evaporative cooling. Using dimensions for the electric engine which
is to be used on the MAV a sleeve was fabricated with 8 evenly spaced fins on it to help conduct the heat
away. Each of these fins extends approximately 15 mm away from the surface of the sleeve and are 18 mm
long. This essentially increased the surface area of the engine 6.5 times its original value. Tests on how
well the heat is conducted away from these fins and away from the engine body itself must be performed to
gauge its effectiveness.

Isometric View of the Heat sink shroud for the electric engine.

The second system for cooling the engine does not rely on conducting the heat away but rather
allowing an evaporative cooling process using a liquid with a low boiling point. Ethanol, or some other
liquid, was chosen since it is relatively easy to obtain and harmless to people. A small plastic surret with a
cap was chosen for the reservoir to allow easy refill during testing and flight conditions. Into the side of this
container were drilled 3 small evenly spaced 1 mm holes to allow the fluid to seep out and into the
membrane through capillary action. The membrane used to shroud the engine and absorb the Ethanol was
to be a simple cotton/polyester gauze pad again due to the relative ease of obtaining. Wetting the gauze pad
and then filling the reservoir with water for testing conducted initial testing of this system. Then a fan was
placed on the system at 15 m/s air velocity to simulate airspeed and the system was timed as to how long it
would take the reservoir to drain. Initial tests show that the reservoir would drain in 9 to 10 minutes with
water, but this could be prolonged by the use of a larger reservoir or the use of smaller holes in the
reservoir. Further testing of this engine cooling method will continue.

Isometric View of the Perspiration Cooling System.

Shared by: wangnuanzg