Dorothy - Hurricane Measurement Project

Written: February 16, 2019



In 2016 when Hurricane Matthew was expected to hit our coast, I thought it would be interesting to rig up a little wind measurement device. I opened up my desktop computer and pulled out the rear fan. I think the original intent was to treat the fan as a generator (like a windmill), and measure the voltage it was producing to gauge the speed. However, I was delighted to find that there was a third wire attached to the fan, and this wire would supply an analog reading (a voltage between 0-5v, or 0-255 as the Arduino would see it) that was proportional to however fast the fan was spinning. I'm pretty sure it was just that wire and GND that was connected to the Arduino. I remember it behaving oddly. The fan would periodically tick as though it were trying to start. It was a crude set-up but it did work, partially. The main issue was that the fan had a lot of resistance to overcome before it would start turning, so it would only turn during strong gusts. This made the readings very sporadic. Another shortcoming was that the fan had to be oriented in the direction of the wind. The readings were stored in the EEPROM and later retrieved through the serial monitor. It was far from perfect, but still a neat little project.


Fast forward to 2017. Now Hurricane Irma was on her way. I wanted to do better this time. I ordered a BMP085 sensor module from Adafruit that would record barometric pressure and temperature. Once again, I ripped out that poor rear fan out of my desktop (It had survived Matthew). But this time I had rigged up a little anemometer out of grill skewers and lefterover beer-pong balls cut in half and glued the base to the fan. This time it would have more torque to overcome the fan's resistance.

To conserve storage space in the Arduino, barometric pressure is measured as an offset from 900 millibars.

I named this new rig Dorothy after the weather recording machine from the movie Twister. She worked fairly well. There was still an area of resistance that the wind would have to overcome before the fan would turn, but it was much more responsive than the first edition. It recorded well into the night when the storm was at its worst, but the winds eventually tore the wind cups from the fan. Fortunately I was able to get some interesting data before that. You can see how the pressure and temperature drops as the hurricane approaches. Its also interesting to note how the wind has periods of high and slow speeds. If you've been through such a storm, you know that you don't get constant wind speed throughout, but rather you get hammered with the rotating bands as the storm approaches and exits. These bands are the violent areas of the storm.

I do not enjoy the damage that these storms bring, but I would love the opportunity to try and improve on this project again.

You can download the code as an Arduino sketch file here. I'll try to go back and comment/explain it in the future.