Wind Sensor

Introduction

When riding a bike, energy is added by pedaling. We, unfortunately, live in a less-than-ideal world, and not all of this energy gets put directly into propelling you forward! While I was at SRAM it became necessary to begin quantifying these losses in order to find the most efficient race setup. My portion of this project was solving for the aerodynamic losses experienced by the rider because of wind.

Requirements

Measure wind speed and direction OR
Measure wind speed opposing direction of travel
Limit moving parts

Possible Solutions

Ultrasonic Windspeed Sensor

Multi-hole Pitot Tube

Weathervane + Anemometer

Multi-hole Pitot Tubes

Given the requirements for this project a multi-hole pitot tube became the solution to pursue. The bumps and vibrations experienced on a bicycle have the potential to damage or affect the data of many other sensors.

Pitot-static probes take two measurements: total pressure and static pressure. Total pressure is measured from the front of the tube. Total pressure increases as wind speed increases as more air is "pushed" into the tube. Static pressure is measured from the side of the tube, and is equivalent to the pressure that would be experienced if there were no wind.

Pitot tubes can be found anywhere wind speed, or speed relative to wind, is important. This includes everything from airplanes to Formula 1 cars!

The two pressure measurements, total and static pressure, can be combined in an equation with dynamic pressure, q. Dynamic pressure can then be related to windspeed, allowing us to calculate windspeed from our two pressure measurements!

So pitot- static probes give us a way to measure airspeed, but how can we extract direction information? By adding more total pressure holes! We can then measure the pressure at each hole and calculate both speed and direction.

Design

After researching existing multi-hole pitots, I created three different designs to test:

Round Multi-hole Pitot - The simplest of the three designs, this one is just a standard pitot with two additional holes. These are both oriented at 45° from the center.

Flat Multi-hole Pitot (Preston Probe) - These are often used for skin friction measurements and wind tunnel characterization. I thought it would be interesting to test if a lower profile, wider probe would work better or worse at bicycle speeds. This design also has two additional holes oriented at 45° from the center.

Modified Kiel Probe - A Kiel Probe is a shielded static-pitot probe. This is the make it more cable of accurate readings at higher yaw angles. This does not mean it can measure wind angle, but that it can measure speed accurately even when wind is not in the direction of travel. This design began with a NACA defined Kiel Probe. The modification is primarily the addition of two additional total pressure ports on the inside of the shield. When wind is off-center, the pressure reading at the inside of the shield at the opposite side of incoming wind will increase. After running flow simulations another SRAM engineer advised that I add windows in front of the shield ports. This allows air to stay attached and pass to the opposite port even at higher yaw angles. This brought the usable angle for the sensor from about 10° to about 45°. The most effort and optimization was put in to this design, the hope is that it gives us more reliable results!

The Kiel Probe was tested at multiple angles and wind speeds, it was iterated upon to reduce turbulence coming off of the leading edge. As can be seen above, some early versions performed better than later, more extreme designs. The ideal design sat in between V4 and V8, which eliminated the small bubble of low pressure that can be seen behind the leading edge in V4, but also elimated the dirty air into the primary sensor seen in V8.

Testing

For initial testing and setup we put together the GTT Wind Tunnel! It was built from a fan with 1 speed and some 4" tubing. Less than ideal but it allowed us to set up the necessary software and verify all designs could take measurements.

Pressure at center sensor at 0°: 13.2696 Pa

So, the GTT Wind tunnel runs at about 5.2 m/s or 12.6 mph

Not quite bike speeds...

Luckily, the lab I work in at CU has a much better wind tunnel! We were able to do all of our testing for all three designs here.

Results

The data collected at CU is plotted below:

The plot above is for the ports on the left side of the probes. The reason for the time put into the Kiel Probe is immediately obvious. At yaw angles opposite from the left port it almost acts as a static port. This makes our calculations a little easier! We can also see what we were trying to avoid on the other probes: negative pressure readings. Now, surprisingly enough, these negative values remain very linear. Maybe there is still hope for the more standard probes!

Next are the readings from the central pressure sensor. The more conventional probes behave as expected, with pressure slowly dropping at higher yaw angles. Unfortunately, though, the Kiel probes see an increase in pressure as yaw angles increase, exactly what this design is supposed to prevent! It would be worth exploring if this can be solved with a slightly different design.

Finally, the right pressure sensors. This looks very similar to the left and isn't interesting until...

... the outer pressure sensors are plotted together. This shows that the readings are very symmetrical, which is good news as we begin to try to calculate wind speed and direction from recorded data.

Next Steps

This data was all collected in my final week at SRAM. This unfortunately means my involvement in the project stopped here. Seeing all of the initial results made me very excited for where this project would go. I had the awesome opportunity to learn more about fluid dynamics, CFD, and sensor design. All of these will be super helpful as I progress at CU.

Page updated

Report abuse