My experiences with a 2kw turbine generator

[Edgar Paternina]

Hello Everyone,

As promised I want to share with you my experiences during more than 6 month with my prototype eolic system, you can find this experience at:

http://www.geneolica.com/Data/EolicExperience.pdf

I would really appreciate any critical comment about it, and my apologies if some times I seem so personal. Internet offers us a unique opportunity of that saying that goes: a picture is worth more than thousand words.

My best regards

Edgar

[Rob Littiken]

Nice write up Edgar.

Thanks

[Edgar Paternina]

Nice write up Edgar.

Thanks

Thank you Rob! I have already made some corrections Hugh Piggott kindly suggested, one regarding my english, and another a typo regarding a formula. On the other hand the main problem is regarding the altitude; my altitude is 2,570meters high, and I just made a correction of about 3.3% in the WindLog to obtain the correct results; according to the theory it would have to be about 25%, I really don't know what's going on here. Again I would appreciate any comments about this!

My best regards

Edgar

[Edgar Paternina]

Thank you Rob! I have already made some corrections Hugh Piggott kindly suggested, one regarding my english, and another a typo regarding a formula. On the other hand the main problem is regarding the altitude; my altitude is 2,570meters high, and I just made a correction of about 3.3% in the WindLog to obtain the correct results; according to the theory it would have to be about 25%, I really don't know what's going on here. Again I would appreciate any comments about this!

My best regards

Edgar

Hi Rob and Everyone reading along,

After a long conversion with my great friend Hugh Piggott, I have already a better version of my work and experiences which I thought you might be interested:

http://www.geneolica.com/Data/EolicExperience.pdf

Regards

Edgar

[Rob Beckers]

Edgar, I finally had some time to read through your manuscript. Nicely done!

A few notes: The term "eolic" is French/Spanish, in English it's "wind". Your comment regarding the manufacturer's power curve not quite being reliable is very common; I never use manufacturer's power curves when creating MPPT tables. Instead I simply go by diameter and reasonable numbers for blade/alternator efficiencies. The only info from the manufacturer that I use is loaded voltage (AC or DC) and power at that voltage (usually rated power) so I can link voltage and power together. When using 3-phase theory to get to values of DC voltage and current be careful that the power factor of a regular bridge rectifier (as used for all small wind turbines) is very poor. It causes much higher AC currents, and bad wave forms, than theory would predict.

I liked your comment that with Chinese turbines you shouldn't expect warranty for the low price. The money saved in the purchase is what goes towards warranty issues later (paid by you). That is very well put. I hope more people realize that.

So you're living at about 2600 meters altitude. That's quite a height! No worries there with rising sea levels... :blink:

-RoB-

[Edgar Paternina]

Edgar, I finally had some time to read through your manuscript. Nicely done!

A few notes: The term "eolic" is French/Spanish, in English it's "wind". Your comment regarding the manufacturer's power curve not quite being reliable is very common; I never use manufacturer's power curves when creating MPPT tables. Instead I simply go by diameter and reasonable numbers for blade/alternator efficiencies. The only info from the manufacturer that I use is loaded voltage (AC or DC) and power at that voltage (usually rated power) so I can link voltage and power together. When using 3-phase theory to get to values of DC voltage and current be careful that the power factor of a regular bridge rectifier (as used for all small wind turbines) is very poor. It causes much higher AC currents, and bad wave forms, than theory would predict.

I liked your comment that with Chinese turbines you shouldn't expect warranty for the low price. The money saved in the purchase is what goes towards warranty issues later (paid by you). That is very well put. I hope more people realize that.

So you're living at about 2600 meters altitude. That's quite a height! No worries there with rising sea levels... :blink:

-RoB-

Thank you Rob, for your time in reading it! In fact, now I have, I hope, a more definite version, because of a long conversion with Hugh Piggott, specially regarding the need to introduce correction factors in the calculation formulas because of the density of air at my place, 2570 meters high.

The term oelic I think comes from the Greek god of wind, Eolos.

Yes, the problem I had from the very beginning for the MPPT curve, that in a certain sense has to do with the calculations formulas now I use to obtain correct data, I mean, the total energy generated by the system. It seems that the details about windspeed is not so important, after all, as I used to think at the beginning. It was a real huge problem, a reason why I bought two different wind loggers that gives each one a different windspeed for the same moment; what is the real windspeed in my case? I don't know!

I solved the problem by taking for granted the nominal data-not the windspeed vs power curve as you pointed out- of the turbine generator given by the manufacturer. Those nominal data are the RPM, the corresponding DCV and the nominal windspeed for the nominal power. This nominal windspeed of course will not be the real windspeed in my case because of the density of air, as for it the real wind power will be greater than the electrical power of the wind turbine generator. As I wrote, one thing is the electrical power and another is the wind power; they are not the same in value even at sea level, but for obtaining the correct energy generated for the electrical wind turbine generator, we must use its nominal data, independently of the air density correction. It seems confusing and not reliable using a variable that cannot be measured in the same way we can measure a DC voltage, isn't it?

With the Chinese it's quite difficult to expect a warranty, first because they are on the other side of the world, and on the other for the low price. For example, my dump load controller failed because it's a very simple one at a very low price.

Yes, I am living in a high place and my wind turbine generator has worked quite good in spite of it! From the very beginning I was afraid because of that I had read about the altitude, but in my case there hasn't been any problem. If we have a good average windspeed- even small as 3m/s- all the time, even at a high altitude wind power will be enough for generating good electrical energy!

My best regards

Edgar

[Rob Beckers]

Actually it is not all that hard to optimize your MPPT curve, if you wish to do that, for your specific wind turbine (though even working from diameter, TSR, and a few reasonable efficiency guesses will get you an MPPT curve that works quite well, without ever measuring a thing).

What you do is start with a 'best guess' MPPT curve based on theory, load that into the inverter, and let it run. Now log both the wind speed and output power of the turbine for some time, preferably through the entire range of working wind speeds (low to high winds). You don't need to know actual wind speed, so the anemometer doesn't need to be calibrated. As long as it is reproducible. A little binning and smoothing will give you the turbine's power curve (wind speed vs. power) from that data, for the current MPPT table that is loaded into the inverter. Now change the MPPT curve by moving the power values up 20% and repeat. Do this a few times, moving the MPPT curve both up and down in power a few steps (a range of -40% ... + 40% should do, so 5x in total), and then plot all those power curves together.

What you're after is the maximum power curve for your turbine, where you get the most power out of it, for all wind speeds. It should at this point be clear what MPPT values go with the best power curve; maybe for the low wind speeds you have to raise power vs. the theoretical MPPT curve, maybe at high winds speeds it needs to move the opposite. That is the MPPT curve you load into the inverter, and you live happily ever after (plus you just corrected for altitude as well by doing this).

By the way, if I recall from your document you gave your MPPT curve a smooth 'top', where the gradient gets less steep. Wind turbines do not work that way, what you need to see is a curve that is between 2nd and 3rd power of the voltage (closer to cube, 3rd power), all the way to the maximum power value. Only if you have to change TSR at the top end for some reason will the curve deviate from that. In your case, making the curve smooth at the top means you'll be running a higher TSR there than the rest of the curve, usually not desirable unless you have to to keep heat losses in the stator in check.

-RoB-

[Edgar Paternina]

Actually it is not all that hard to optimize your MPPT curve, if you wish to do that, for your specific wind turbine (though even working from diameter, TSR, and a few reasonable efficiency guesses will get you an MPPT curve that works quite well, without ever measuring a thing).

What you do is start with a 'best guess' MPPT curve based on theory, load that into the inverter, and let it run. Now log both the wind speed and output power of the turbine for some time, preferably through the entire range of working wind speeds (low to high winds). You don't need to know actual wind speed, so the anemometer doesn't need to be calibrated. As long as it is reproducible. A little binning and smoothing will give you the turbine's power curve (wind speed vs. power) from that data, for the current MPPT table that is loaded into the inverter. Now change the MPPT curve by moving the power values up 20% and repeat. Do this a few times, moving the MPPT curve both up and down in power a few steps (a range of -40% ... + 40% should do, so 5x in total), and then plot all those power curves together.

What you're after is the maximum power curve for your turbine, where you get the most power out of it, for all wind speeds. It should at this point be clear what MPPT values go with the best power curve; maybe for the low wind speeds you have to raise power vs. the theoretical MPPT curve, maybe at high winds speeds it needs to move the opposite. That is the MPPT curve you load into the inverter, and you live happily ever after (plus you just corrected for altitude as well by doing this).

By the way, if I recall from your document you gave your MPPT curve a smooth 'top', where the gradient gets less steep. Wind turbines do not work that way, what you need to see is a curve that is between 2nd and 3rd power of the voltage (closer to cube, 3rd power), all the way to the maximum power value. Only if you have to change TSR at the top end for some reason will the curve deviate from that. In your case, making the curve smooth at the top means you'll be running a higher TSR there than the rest of the curve, usually not desirable unless you have to to keep heat losses in the stator in check.

-RoB-

-RoB-

Thank you Rob for such a detailed explanation how to program the inverter with the MPPT curve. In my case I took for granted it was well programmed by Magnetek engineers. My main concern initially was to check the data given by the inverter, I mean, the energy generated according to it, with my kwh meter and my two Kill a Watt installed in each phase. This was specially in the month of January. When on the first of February failed the dump load controller, then I started thinking to install an additional anemometer that could log the information to know what could happen in such a moment of very high wind; the one already installed, the LaCrosse does not log the windspeed, you can just see it on the display.

You wrote:

By the way, if I recall from your document you gave your MPPT curve a smooth 'top', where the gradient gets less steep.

which curve are you refering to? Some of those curves are obtained calculating the windspeed in case of the inverter, and the others are obtained with the data given by the anemometer, in which case corrections factors must be introduced to obtain the right total kwhs. The point is that just the pattern of the curve seems to be the same, but not the detailed information. This is something I don't understand and that has taken me to think in the relativity of the windspeed concept from the point of view of the wind turbine behavior and of the wind loggers manufacturers. In the last analysis what is important after all is the energy generated(kwhs) in case of this grid-tie system.

My best regards

Edgar

[Steven Fahey]

Edgar,

I enjoyed reading your write-up, too. I know someone else who had an Exmork, and his experience was definitely not as good as yours.

Can you tell me why you plotted your Aurora power graph on page 36 the way you did?
At first I think the power curve is very variable, but then I realize that the horizontal axis is not in regular intervals.
Is it not possible to plot the graph with regular (eg 1 m/sec) intervals on the bottom?

I point this out because I have seen wind turbine sellers that DELIBERATELY hide problematic operation of their turbines by fudging the scales of their graph. The curve looks normal at first, but if you fix the intervals, the line goes all over the place.
(The opposite effect of what you're doing)

You would not want to be accused of doing the same thing. Here is an example:

http://www.wepower.us/images/falcon-5point5kw-powercurve.jpg

In this case, they hide the low-wind performance by stretching the curve upward. Spacing on the vertical axis is 500, 1200, 2800, etc., not 500, 1000, 1500, etc.
I consider it dishonest. :mad:

Obviously you aren't trying to do the same. I'm not offended at all! But I thought I should warn you that a graph with strange scales is very hard to read!

Again I enjoyed reading your write-up and I wish you the best of luck. Especially looking out on your beautiful landscape.:)

[Edgar Paternina]

Edgar,

I enjoyed reading your write-up, too. I know someone else who had an Exmork, and his experience was definitely not as good as yours.

Can you tell me why you plotted your Aurora power graph on page 36 the way you did?
At first I think the power curve is very variable, but then I realize that the horizontal axis is not in regular intervals.
Is it not possible to plot the graph with regular (eg 1 m/sec) intervals on the bottom?

I point this out because I have seen wind turbine sellers that DELIBERATELY hide problematic operation of their turbines by fudging the scales of their graph. The curve looks normal at first, but if you fix the intervals, the line goes all over the place.
(The opposite effect of what you're doing)

You would not want to be accused of doing the same thing. Here is an example:

http://www.wepower.us/images/falcon-5point5kw-powercurve.jpg

In this case, they hide the low-wind performance by stretching the curve upward. Spacing on the vertical axis is 500, 1200, 2800, etc., not 500, 1000, 1500, etc.
I consider it dishonest. :mad:

Obviously you aren't trying to do the same. I'm not offended at all! But I thought I should warn you that a graph with strange scales is very hard to read!

Again I enjoyed reading your write-up and I wish you the best of luck. Especially looking out on your beautiful landscape.:)

Hi Steven,

Thank you so much for your supporting and kind comments!

The main advantage with the chinese wind turbine at this moment is its price, just when engineers from this side of the world enter in a synergectic relationship with them as has been done with other chinese products, then we will have better experiences in this new field as they have specially the advantage of the rare-earth magnets.

Regarding the curve, the windspeed on the horizontal axis is calculated by making some assumptions not so well-accepted by some experts; I assumed in my formulas a sort of linear relationship between RPM and windspeed by using the same TSR in the whole range. The DCV value is the one measured by the Aurora as its corresponding power; the windspeed is calculated with the conventional formula without making no density air correction that should be done at my altitude according to the theory, and from this I calculate the DCVc.

The intervals of this horizontal axis were just selected by mere convenience, trying not to repeat the same value in a certain range, for example in that range lower than 200watts, and so on.

What I have found is that the DCV measured value differs from the DVCc calculated value by a percentage of 5% maximum in average, a reason why I assume it's ok to a certain extent as losses in the wind turbine and controller will make this difference. But the important thing is that I have an approximate value for the windspeed. Please note that my two anemometers, they do not coincide in their values for the same moment, a reason, why I say that the windspeed is a relative concept. What is important after all in a grid-tie system is the permanence of the windspeed and not a maximum windspeed as in off-grid systems with batteries, in this case there must be a minimum power to be feed, and it needs certain maximum windspeed to charge their batteries.

Another important conclusion of my experience is that wind turbines work well even at my altitude, 2570meters high, as there is written somewhere that they should not work so well at this altitude, but they certainly do work ok! The main problem that I have had is that this year the weather has been quite strange, it has rained almost all the time, so little wind. It was not so in seven years!

My best regards

Edgar

[Steven Fahey]

Hello Edgar,

The difficulty is making the data you collect make sense! You are not alone as you struggle with this. I have much the same trouble. My anemometer is 5 meters above ground, and the turbine is 14. The difference in wind at the turbine and the anemometer is ofter VERY different in speed and in direction. All I can do is try statistics to find averages, then sort them with the "bin" method.

My wind turbines do not operate at constant TSR. Even if you design the blades to work at a particular TSR, and then select a generator that will deliver peak power at a matching speed, at any other speed the system will probably run faster or slower. So inferring the wind speed from the RPM and the "design" TSR does not always work. If you do not trust your anemometer, then you may have no other choice.

I took a minute to see what your numbers look like on an even-scaled graph. Most of the zig-zag is taken out now. I attached a little copy of it.

The challenge now, is to make a stronger link between the output power data you collect and the wind speed at that moment. I am compelled to analyze and test every morsel of data. I think I do it too much in fact!

I live at 1000 meters altitude, air density affects my turbine, too. I only suffer a 90% penalty, not like yours at 75%.

[Edgar Paternina]

Hi Steven

Hello Edgar,

The difficulty is making the data you collect make sense! You are not alone as you struggle with this. I have much the same trouble. My anemometer is 5 meters above ground, and the turbine is 14. The difference in wind at the turbine and the anemometer is ofter VERY different in speed and in direction. All I can do is try statistics to find averages, then sort them with the "bin" method.


One of my anemometers is down about 3meters from the wind turbine and in normal wind they show the same behavior regarding direction, but this Rainwise anemometer, it gives a correct result regarding the kwhs predicted but the pattern of the windspeed is not reliable, a reason why my reference regarding that windspeed is not real.


My wind turbines do not operate at constant TSR. Even if you design the blades to work at a particular TSR, and then select a generator that will deliver peak power at a matching speed, at any other speed the system will probably run faster or slower. So inferring the wind speed from the RPM and the "design" TSR does not always work. If you do not trust your anemometer, then you may have no other choice.


Yes, my choice is to calculate based on some theoretical assumptions the windspeed and then a corresponding DCVc which is similar in about 5% to the measured DCV. This is my best guest. Why is it relatively acceptable, I wonder... I mean, if I assume a constant TSR for the whole range?


I took a minute to see what your numbers look like on an even-scaled graph. Most of the zig-zag is taken out now. I attached a little copy of it.

The challenge now, is to make a stronger link between the output power data you collect and the wind speed at that moment. I am compelled to analyze and test every morsel of data. I think I do it too much in fact!


Yes, thank you, the curve looks much pretty, but in my case, I always make the curve with minute by minute data, just like in the attachment, in this case, the windspeed is a calculated one.


I live at 1000 meters altitude, air density affects my turbine, too. I only suffer a 90% penalty, not like yours at 75%.
[/QUOTE]

So no problem with the altitude too, which was my real concern. Regarding the air correction, it really does not matter too much in my case, because it affects the theoretical windspeed, not the real data regarding the kwhs predicted, it problably is reflected in that 5% of the DCV measured and the calculated one. I really don't know how to associate this 5% with the 75% of my altitude.

My Best regards

Edgar