Mixing wind turbines with the tropics

Predictions from Previous Article

Based on curves and measurements of the wind turbine (hanging outside a car), and the average wind speed on Curaçao (15mph), I was expecting about 100mA into the charging battery or about 0.5W. I thought it would not generate enough power to run the Pi without the wind speed going to at least 30mph (power available from wind goes up with the square of the wind speed).

Actual Results

Project Curaçao also has a MySQL database that records current going into the battery charging system. When the system is set for solar, it records current into the charging system from the solar cells and, when set for wind, it records the actual wind power current into the charging system. While the Pi is up and running (after sunset), it is recording the actual current from the wind. When the solar cells are connected, I can still estimate the potential available current based on the unregulated wind turbine (right off the turbine) by looking at the unloaded 50W wind turbine curves given in the previous article.

In the same way, I can estimate the wind speed by looking at the unregulated wind turbine voltage (when it is unloaded – it is not connected from the charger during the day). Remember that only voltages up to about 18V are significant, because anything more will be clamped by the voltage protection circuitry. However, the lower part of the curve should be very close to the unloaded 50W wind turbine curves.

To do this analysis, for March 16th (two days before the big wind storm and the best solar power day of the week), I crunched the data in the MySQL database. The "Data Collected" box shows a summary of the information.

I also spot-checked wind power and current during the day using RasPiConnect. The recorded observations are as follows:

• 1. The curves from the previous article indicate almost zero current generation until the wind speed gets to about 7.5mph. My data more or less agrees with the curve.
• 2. I thought that it should be generating much more current at 15mph, and the data seems to support that. Note that I can't tie an exact wind speed to the 138.6mA value because the wind turbine is loaded at that point, so unloaded wind/voltage curves aren't applicable. This is one reason why, during my upgrade in September 2014, I will be adding a separate anemometer.
• 3. As expected, the 50W turbine does not even come close to powering the device, but it does provide a trickle charge. Eight hours at 22mA will result in 30 minutes of running the Raspberry Pi.

Figure 4 shows the RasPiConnect Pi power panel with the wind turbine manually turned on with the panel during the day. I stood at the foot of the tower and watched the wind turbine while refreshing my iPad and taking screenshots of the data. Note that although the wind charging current is accurate (under the label of Power from Solar), the wind voltage is delayed 15 minutes – making the wind voltage value pretty much worthless. This issue also will be addressed in the upgrade.

Figure 4: Monitoring power usage with the RasPiConnect power panel.