Windmill Optimization

Objectives:

1. Power Enhancement: Optimize DC-motor windmill power output.

2. Voltage Stability: Achieve consistent DC voltage through signal filtering.

3. Method Comparison: Evaluate optimization techniques against the baseline design.

4. Tool Proficiency: Gain hands-on experience with oscilloscopes and multimeters.

5. Design Iteration: Practice the iterative design methodology.

Materials Required:

• 1 DC Motor

• 1 Windmill PCB with:

• 4x Capacitors (1206 SMD)

• 2x Resistors (1206 SMD)

• 2x Connecting Wires

• 2x Alligator Clips

• Solder

• Hot glue

• Super glue

• 1 Windmill Stand

• Tools: Multimeter and Oscilloscope

• Soldering iron

• 1 3D Printer

Software Required:

• 3D Modeling (e.g., Fusion360, TinkerCAD)

• Circuit Simulation (e.g., Falstad)

• WaveForm

• Aerodynamic Analysis (e.g., NASA simulations)

1. Initial Design Experimentation with Provided Blades

To begin, I downloaded and imported the provided .stl files of the blades and the hub into PrusaSlicer. The components were already sized down, so all I had to do was slice the design. Then, to assemble, I removed all of the supports and super-glued the blades to the interior of the hub. I used a small drill to drill the hole on the other end, before I inserted the DC motor in.

Link to the  hub .stl file used

Link to the blade .stl file used

To test the horizontal windmill blades, I used alligator clips to connect to the scope channel 1 (positive and negative), or the orange wires to the motor. Then, plugging the micro-usb into the computer, I opened up WaveForm, which is an application that tracks the changing voltage levels of an AC signal. I spun the motor manually, watching the wave spikes on the screen. Because we were measuring AC voltage (and not DC), the waves were oscillating unevenly, as the electrons were constantly switching directions. This represents what the voltage looks like before its goes through a transformer; the transformer processes and filters the voltage, converting it to a steady DC.

2. Researching the Optimal Windmill

Websites Used:

• https://windandsolar.com/blog/blade-types-for-wind-turbine-users-the-complete-guide-/

• https://www.energy.gov/eere/articles/wind-turbines-bigger-better

• https://www.sciencedirect.com/science/article/pii/S221509861500155X

Conclusions:

• https://windandsolar.com/blog/blade-types-for-wind-turbine-users-the-complete-guide-/

Under this website, I learned a lot about the ideal number of blades and the effect that drag has on the speed (and ultimately efficiency) of the blades. The site compared industrial wind turbines vs. homemade, concluding that the number of blades was not just limited to three; it found that for smaller windmills, more blades resulted in greater stability, whereas less blades were more efficient. Besides that, it briefly mentioned how crucial the angle of attack, or the pitch, was for determining a windmill's effectiveness. If the design were too narrow, wind wouldn't "generate flow over the blade," and if the design were too wide, it would create too much drag.

• https://www.energy.gov/eere/articles/wind-turbines-bigger-better

Under this website, I read about how increasing the diameter of the rotor directly "allow wind turbines to sweep more area, capture more wind, and produce more electricity." Keeping this in mind, I tried to increase the length of the windmill blades in Fusion360 by changing the distance of the offset planes. This was particularly prominent in the first two prototypes.

• https://www.sciencedirect.com/science/article/pii/S221509861500155X

I briefly skimmed through this site, finding that the ideal angle of attack range was between 5º and 15º, depending on the environment.

3. Blade Design and Testing

Goals:

• Model new blades, maintaining hub compatibility.

• • Test and graph three different blade designs.

  • Select the most power-efficient blade design of the 3.

Methods:

• For this process, each group created at least three windmill models, adjusting factors such as angle of attack, blade length, and depth to form the most optimal design. After 3-D printing each model, we had to drill a hole in the bottom of the hub to fit the outer shaft in; we found that hot-gluing the two together was the best way to keep the windmill attached. Then, placing it in front of the fan, we utilized the oscilloscope to measure the millivolts generated. Similar to the initial assessment, we tested and analyzed 0º, 45º, 90º, 135º, and 180º to see which prototype was the most efficient. In the end, we found that each prototype had its strengths in weaknesses--for instance, certain designs were more consistent throughout all angles but had a lower average voltage, whereas others were less consistent but had a significantly higher average voltage for 0º.

Links:

• Link to a spreadsheet with graphs

• Link to prototype 1 .stl file

• Link to prototype 2 .stl file

• Link to prototype 3 .stl file

• Link to prototype 4 .stl file

1. First Prototype

For the first prototype, I attempted to create a modified version of the original, using the surface loft under the surface tab instead of the solid tab. First, I created two offset planes and used splines to draw the general shape of the turbines. Then, I applied the surface loft and the thicken command (1.5 mm). During my research process, I learned that having longer and angled blades was the ideal shape, so I tried to implement that in this design. I edited the dimensions so it would be longer than the original blades. Lastly, to ensure compatibility with the original hub, I used calipers on the original model to find the dimensions of the base; I used the two-point rectangle and extruded it by the measured distance.

2. Second Prototype

For the second prototype, I attempted a three-blade design with a hub similar to the original (except without the ridges). To make the blades, I used slots and three offset planes, before using the loft command to connect them (I had also created a spline connecting all three sketches to serve as the guide rail). Then, I created the center by extruding a circle and using the loft tool again to create the upper-cone shape. I applied the circular pattern command, selecting the singular blade as the profile body and the extruded circle as the axis. Finally, I extruded a smaller circle on the bottom of the hub for the motor to fit in. 

3. Third Prototype 

For the third prototype, I tried to take a different approach, adding shorter and flatter blades that angle only slightly away from the wind. To create the hub, I used the revolve command on a cone-shaped spline/rectangle. Then, for the blades, I first drew a horizontal line on the side of the hub, before creating an offset plane and drawing a spline. Next, I applied the loft feature connecting the two sketches. I thickened the lofted design and used the circular pattern feature to replicate the blade design (I selected the initial circle from the hub as the axis). I had initially created 10-15 blades, but there would be too much drag. Thus, I created four blades. While testing this design, I noticed that the blades were a bit heavy and could not rotate very fast as a result. 

Filtered vs. Unfiltered Signals: Below illustrates the signifcant difference in the signal spikes when we introduced capacitors to the motor. Capacitors filter signals, meaning that they decrease the ripple voltage components in the output, converting an initially varying signal to a stable one. Different capacitors filter different size signals (big vs. small), which is why we included a wide range of capacitors on our board. To observe this change in action, we conducted the initial performance test using the original hub and windmill blades. The only difference was that we connected the oscilloscope to the motor with the PCB soldered on. Turning the fan on, we noticed that the signals did not fluctuate at all; it was a lot more stable compared to the original.

4. Conclusions

• Comparison of performance of original output waveform vs final waveform:

• • Compared to the original output waveform, which was constantly rising and falling (unfiltered), the final waveform was relatively stable. There weren't any large jumps or spikes in waveform, and this was consistent with all prototypes.

• Insight gained from testing varying factors:

  • Testing varying factors allowed me to understand how different angles of attack/lengths/shapes can contribute to how well it performs at different angles, as well as its influence how high or low its average voltage is. I ultimately found that longer and more uniform blades produced better results in terms of speed, whereas the unevenly curved blades (such as prototype 3) were better at catching wind consistently at different angles.

  • Particularly applying to prototypes 2 and 3, I discovered that a blade that was too thin failed to catch any wind, while a blade that was thicker had too much drag and couldn't spin as fast. The optimal design would lie somewhere in-between.

  • The most power efficient model ended up being prototype 4, mainly because its voltage output at 0º out-performed the other prototypes. However, if I were to select the most consistent model, I would choose prototype 3, as it was more flexible with different wind angles.

• Changes in the future:

  • In the future, I would try to be more thoughtful in adjusting the dimensions and thicknesses of the blades; my main sources of errors stemmed from failing to analyze the thickness of the base design. If I could create a new design, I would use a thickness that's between prototype 2 and prototype 3.