Designing, and Building My Own Wind Turbine
Field Notes from Irricana, Alberta, Canada
by Steven Fahey

24 September
"Lights ON"
24 October
14 November
"Battery Test"
15 January
15 February
"Solar Panels"
5 June
"Water Pump"
31 December


Boosting the Turbine Output in Light Winds

Last week, I tried using my computer to simulate the behaviour of capacitor-based voltage doubler or tripling circuits, in an effort to increase performance of the wind turbine in light wind. Some notes about the results of the sims are here.

Having no success with computer simulations, I chose to try them on my wind turbine anyway. I would use my little data logger and compare the performance with the capacitors to without.

A "Doubler" circuit, like Gordon's.

A "Tripler" circuit, based on the doubler.

The circuit as tested. Note that I used a bit of a mish-mash of 2200 and 4700 uF capacitors. It's what I had handy. When these polarized electrolytic capacitors are connected "back-to-back" to make them accept AC, the effective capacitance is divided in half. Therefore I used "1100uF" and "2200 uF" on the diagram, and those caps are not shown as polarized.

With these as a starting point, it was time to scrounge up the necessary parts. I could only find 3 full-wave rectifier bridges, so I chose the tripler to begin. I have numerous capacitors that have been pulled out of various appliances and bought for projects over the years. The only had voltage ratings of 35 or 50 Volts, so I have to keep them out of harm's way. Being polarised electrolytic capacitors, they needed to be wired "back to back" to accept alternating current.

Here is the kit assembled...

... and plugged in.

Later I removed the big relay that you can see in the bottom of the box. It was just making things complicated and wasting power keeping the coil energized. Shutting off the Positive lead from the capacitor bank is enough to disable them. I wired that into a switch on the side of the control box. You can see the three bridges in the photo, and the bank of capacitors wired together.

I was ready to run the tests on January 3rd. I'd tried a few times before, but the wind didn't actually blow for long enough to collect enough data. One thing I wanted to have was at least 1000 data points with the circuit turned ON and OFF. Testing under the same conditions would give me more confidence in the data. This idea proved to be a good one, when you see the power curves below.

January 3rd was not a very windy day, here, but a breeze blew fairly consistently all day. This was the ideal condition to test a circuit that promises extra energy collected in light wind. The temperature was -15C at noon and -20C in the evening when the test was done. I started at 9AM and finished at 10PM, collected a total of 4000 data points (after discarding the ones where the wind had dropped to zero).

Here are the power curves...

... and RPM variations.

The black line on the power curve is data collected before January 3rd, about 30,000 data points. Obviously the power is much higher. Under normal conditions, this is the power collected at the wind speed that my anemometer can measure. Since the anemometer is not on the same tower, and much lower, you can have some doubt about its accuracy. I correct it as best I can. On the day I carried out my test, the wind was from the south. It was relatively unimpeded at the anemometer, unlike when it comes from the west, when trees affect it drastically.

Later on, I should add a SCATTER PLOT here - to show what is happening.

The bins are not the usual "1 mph" each. I could start doing that now that I've collected enough data, but when I started out I found only a few data points in each bin that way. Instead the bins are 5kph wide, or about 3 mph. This will invariably cause some exaggeration of the facts. Forgive me. I'm using the curves for comparisons, not absolute claims.

Getting back to the Capacitors, the two lines plotted in blue and purple show the test runs with and without the caps ON. There is a definite improvement with them ON. Not only is there a 30% - 50% increase in power in wind below 10 kph, but the advantage can be measured up to 20 kph. Above that there is no obvious difference.

Another question is "how do the capacitors affect the loading of the rotor?" This question is mostly answered by the second graph. The two curves are roughly identical. In light winds, the blades turn at the same speed whether the capacitors are ON or not. When the wind is stronger, the blades seem to turn more slowly than usual. If the caps added to the power needed to turn the generator, we would see the rotor slow down. This didn't happen in light wind, so my guess is that the caps increase the efficiency of the generator for any given speed. At faster speeds, however, the capacitors do load the rotor more. The gains evaporate and the generator puts a larger load on the rotor. I call this a good thing - anything the slows the generator down as thew wind gets stronger just protects it from strong wind. ON the other hand, it could be a trick of the eye and my unreliable data. I would be very appreciative if anyone would be willing to provide data that can be validated better than mine.

Advantages observed:

  • Increased power collection in light winds
  • Increased load on rotor in stronger winds
  • Reduced scatter in the datalogger's reading of RPM
  • Inexpensive compared to other power optimization schemes (MPPT, buck/boost, etc.)

This is, finally, a satisfying result. Computer simulations failed to show a difference, nor can they illustrate how the capacitors are affecting my system. This makes me think that the effect of the capacitors is beyond that scope of the electronics. I have read some feedback on the forum discussions regarding these voltage doubler circuits where the writers expect the effect to be in the magnetic fields in the permanent-magnet generators.

What that means is that I should go back for an education on how the back EMF forms in the generator, how much happens as current flows, and how the MMF in the stator laminations and the rotor respond in time. Very complicated.

I'm satisfied to leave the capacitors in place for now. They may have problems in time due to things like voltage spikes, however I'm questioning everything I found in the simulations, including that. They will stay inside the metal box, and it's cold out these days. If anything does go wrong, it's easy to shut them off.