Symmetry was also important. If it wasn't balanced, the turbine had difficulty experiencing enough torque to cause it to continue rotating. After having groups of students design their own wind turbines and testing them using the Vernier
Rotary sensor with the TI-Nspire we used the blades from our new LEGO 9688. The TI-Nspire Lab Cradle is needed in order to collect data with the Rotary Motion Sensor. (For more about he Lab Cradle see this entry.)
Student were asked to predict the number of blades to optimize the rotational speed. Here are the results of the Quick Poll from one class.
As you can see in the graph below, 6 blades were the slowest, then 2, and finally the magnitude of the 3 blades rotational speed was the greatest. The data from this experiment was gathered with a wind speed of around 5 m/s.
Update 3/5/2012: This number of blades hypothesis and experiment was done with several of my classes and at the T^3 International Conference. The results from the QuickPolls are interesting to see how the prediction between 2,3, and 6 blades are often evenly split. The data from the experiments is beautiful. For most wind speeds we've done, but not all, the results show that 3 blades result in the greatest angular velocity. The discussion about why this is also enjoyable: More blades will catch more wind, but they will have more mass.