Accurate Power Measurement

[brokengun]

There has been a lot of discussion about people achieving different numbers for power output with similar or the same blade sets recently. It has been making me wonder how people have been measuring their power output and perhaps different methods are introducing error.

I know when I measured my alternator, I only had a drill to test it with and it was a tough job to monitor all the equipment and hold the drill. When I measured mine, I had one meter testing voltage, another testing amperage. I tried to document my results with and without load however they never seemed to be very consistent. It could be just me or it could be the equipment I used but I found that even at a relatively constant rpm my numbers varied (this could be partially due to using a drill and not a lathe).

 What do most of you do to measure your alternators and do you think it is possible that someone measuring a jumpy digital multimeter might get bum results?

Disclaimer: This post is not aimed at anyone on this board in particular, I was just wondering how I could improve the accuracy of my own test results.

[imsmooth]

If you go here http://www.mindchallenger.com/wind/axial12.html

you will see a few pictures down my alternator test rig.  I have a 1.5hp dc motor attached to a variac.  I connected my rectifier/cap bank to the alternator, and then I connected the output to a known load.  I then measured the voltage across the load and the current through it.  I have a RPM meter so I know the rotational speed.

[commanda]

Digital multimeters are notoriously unreliable in this application, unless using an (expensive) true rms meter, and also applying advanced techniques to filter out spikes and noise.

Once the whole thing is up a pole, accurately measuring true wind velocity, coupled with the cubic nature of power vs wind velocity, is a whole other can of worms.

Amanda

[pmurf1]

I have a watt's up meter hooked to my battery and it does it all. It was like $60 on Ebay.

[Flux]

Alternator power measurements on a bench test is not too difficult but as Commanda has pointed out, actual wind power measurements is going to be a different game altogether.

With a decent constant speed drive you should be able to measure the alternator power to within about 3% with reasonable meters. Using a drill you may still be able to get 5% or better. This may seem crude but it is good enough for normal purposes.

If you can measure it to better than 100% under wind conditions you are doing fairly well so a few % here and there with the alternator measurements doesn't matter much.

There may be more error in what you measure than in the way you measure it. You need to know what you are trying to measure and choose meters accordingly.

Normal alternator power measurements for other than wind purposes are ac measurements into a load that is resistive, or a combination of resistive and reactive (Leading or lagging power factor).

If you want accurate measurements for this you need 2 wattmeters for a 3 phase set up. If the load is resistive then rms reading voltmeters and ammeters will be near enough .

For a wind machine the thing is almost invariably used with a rectifier and a dc load so there is no point on taking any readings on the ac side, you will never be able to sort out the various rectifier issues.

For charging batteries you need the battery volts and current, both of which need a mean reading meter and a reasonable digital multimeter is ok but some of the cheap ones do strange things with ripple and noise and if your speed is not constant you have more troubles.

Unless you have a charge controller connected you also have fun with battery volts changing during the test and if you start with it near full charge it may change drastically.

For grid tie inverters I suspect the mean voltage will be near enough and again I suspect mean current will mean more than rms ad far as the power fed to the grid is concerned but you may need rms to sort out alternator efficiency . For accurate results you would need the inverter as load and have it programmed as for the wind operation. Your best results would be from measuring the ac power fed back to the grid with a wattmeter as this includes inverter efficiency.

Few will have facilities to do comprehensive tests and what you are doing with a drill will be good enough to prove your alternator is working. If you can measure speed rather than using the drill nominal figures you should be able to get something like 5% if you take plenty of readings. You will see which ones are rogues and you can disregard them.

If you really want to measure power under wind conditions there is some information on Hugh's site and I seem to remember that Paul Gipe has something similar on his.

Have a look at Michael Klemen's site as well. Comparing with a weather station anemometer somewhere near the turbine is little better than guessing.

No wonder you see various people giving different figures.

Flux

[Flux]

Perhaps I should have added that there is a large amount of information here about alternator power out but few people indeed make any effort to measure the power in to the thing.

The blades have to supply the power in so you need to know that if you want to do any blade matching. I think some would be surprised to know how inefficient some of these alternators become when heavily loaded. If you get 1kW out but need to put 3kW into the thing then it may become a bit more obvious why some of these things burn out.

I often wonder why so many people do tests that are of no use such as measuring open circuit voltage and short circuit current and things like volts into a resistor when the thing is going to be used as a battery charger. Once someone publishes something it seems to set the trend and everyone follows.

Flux

[FishbonzWV]

don't have time to comment right now, but here's what I rigged up.

http://www.fieldlines.com/story/2008/10/5/235214/518

Bonz

[taylorp035]

I used a drill press and three different hand drills.  I then compared the results with the rated rpm of each drill and found that all of the data lined up.  As flux said, I think the small amount of error in these tests are not really cause to get all worked up about.

[SparWeb]

It's a really good question, and there's a lot I could say.

I've done several run-up tests using a lathe at work.  You can see details at the backshed's forum Here

(there is a similar thread on fieldlines with a different discussion Here).

You can see my spreadsheet plots are curvy lines, but acually there was scatter in every single number I measured.  Torque load was +/- 1 or 2 pounds due to vibration (or more), current varied by an amp or two at high speed, and the voltage was constantly creeping upward, even with rotor speed running constant, due to high charging current and gassing electrolyte.  By using a lathe I had fairly constant RPM, but even so, I was checked it with a bicycle speedo, and at high load the slow-down was several percent (AC slip in the lathe's motor).  Relying solely on the speed setting of the lathe would have compounded my errors.  Note that the lathe has a 5HP motor, and at times I was loading it beyond 5hp!

By the time I was finished the tests, the generator was too hot to hold comfortably.  This temperature rise would affect resistivity of the wire.  It took 1/2 hour to cool down enough that I could carry it out of the lathe chuck onto the table (didn't matter because you can see all the other paraphernalia that I had to take apart before getting to take Toshi out of the lathe).  I don't have a rational way to deal with the increase of temperature, so its effect is just there, whatever it is.

A lesson learned early on in experimenting with PMA's was that the low frequency causes many multimeters to LIE.  I bought a true-RMS meter and I've been crossing my fingers ever since.

One last comment: after you've collected a bunch of numbers, you should sort them, plot them on graphs, look at the results, and then consider every "outlier".  Think about what happened at each one, and eliminate the ones you think are bad.  You are allowed to fix erroneous numbers in your calculations as long as you can justify the method of the fix.  I did a fix on my Toshi numbers to account for wildly rising voltages, so that they wouldn't exaggerate the output power.  Do this before posting to the forum because glitchy numbers can lead people astray while they are trying to learn about what you did.  Then if you want to post about the fix itself, the subsequent quetion is asked in a clearer context.

[DamonHD]

Old science joke: if you want your graph to be a straight line then only measure two points!

Rgds

Damon

[brokengun]

"I often wonder why so many people do tests that are of no use such as measuring open circuit voltage and short circuit current and things like volts into a resistor when the thing is going to be used as a battery charger. Once someone publishes something it seems to set the trend and everyone follows."

Sometimes you just don't know any better, I know it was curiosity for me. I just wanted to see how much power I could make on the bench.

[ghurd]

Bench Power can be deceiving for windmills.

The best example I can think of is a windblew PMA.

Their charts can easily be extrapolated to mean they make a lot of power in the wind.

Not many people seem able to break 45W with wind.

But they will make some decent power when connected to a gas motor.

Wind doesn't seem to be able to supply the input power and RPMs to make much output.

Works fine with gasoline.

Might get a bit of insight from here,

http://www.fieldlines.com/story/2009/6/26/223425/021

I hope you get your gun fixed.  

G-

[SparWeb]

Measuring input power is where all of this can get dodgy, for sure.  A lot of things are going on simultaneously and it takes a lot of preparation in advance to have a sensible test run (for me, at least).  I do my best to get clean results on output AND input, but when working with odd materials at hand, one isn't always going to have the best set-up.  Nobody's said so in the past, but there must be things that I've been missing, too, in my tests.  I've also noticed that few others go the extra mile, but when some do, the information is valuable to many many others.

There are numbers you obtain to satisfy your own curiosity, and then there are numbers that you can publish to back up claims of a product for sale.  Two entirely different things, and one wouldn't expect the level of detail used for proprietary (and reputable) purposes to show up on a web forum, nor should we forget that we're here because we're "amateurs".

I hope you're not feeling frustrated about that fact, Flux.  Science and technology doesn't come easily, even to educated people, because very few of us get beyond "electricity is magic".  Everyone uses the word "efficiency" but very very few could deliver a technically correct definition of the term.  Those that join this kind of hobby have to un-learn a lot of stuff before getting anywhere.

[bob g]

thought i would wait a bit to add my .02 and glad i did

it is refreshing in a twisted sort of way to see others struggle to get good

test results,

particularly the interesting thing is the effort that is expended.

i belong to other forums, just as most of all of us do, on in particular

is populated with folks the majority of which make all sorts of claims

but their eyes glass over if presented with even the simplest of scientific method.

i guess it is easier to make a claim than it is to back it up?

because i am not working on testing windgen's at this time, i can't add anything specific to this sort of testing, only thing i can do is assure each of you that

testing can be a real pain in the rear.

my interest is with diesel engine driven gensets, with emphasis on co and trigeneration.

it has taken many months to work out testing protocols, go through batches of test instruments, meters, scales, etc weeding out those that are unsuitable for my purposes.

only when i get repeatable results over and over again, on different days, and after i compare my results with commercial units published graphs/charts/specs

and then after i plug all my numbers into a spread sheet and reverse work the math

using standards that are known do i feel comfortable with my test results.

thats a lot of work to say the least.

and it can't be done using cheap meters from harbor freight, at least in my experience.  i have enough trouble with true rms meters from fluke agree'ing with one another.

some days everthing works as planned, you run a battery of tests, they deliver consistant results (step one is a go), you go back and crunch the numbers and plot the graphs, and find everything fits neatly in a curve (step two is a go), you then

reverse prove the math against known standards and it proves out (step three is a go) and then you compare your results against those published by the big boys

on units of similar size and find you are in the hunt (ok.. step four is a go)

thats a good day!

then other days not a single step pans out, everything goes nuts, your graph looks like a drunk had his way with a shotgun, nothing you can do to make the math work either forwards of backwards, and when compared to commercial units ,,, well there

is no comparison...

thats a bad day!

sometimes it is a simple thing like a bad connection to a meter, or any number of little details that drive you insane.

seems like setting up a test cell is much more work than it took to build the unit you are wanting to test!

ain't it fun?

bob g

[Flux]

"I hope you're not feeling frustrated about that fact, Flux.  Science and technology doesn't come easily, even to educated people, because very few of us get beyond "electricity is magic".  Everyone uses the word "efficiency" but very very few could deliver a technically correct definition of the term.  Those that join this kind of hobby have to un-learn a lot of stuff before getting anywhere.

Steven Fahey"

This is very true. With wind you have to start out with a clear idea of what is happening. Concentrating on the alternator alone is not the right thing unless you use some form of mppt converter to battery or grid. If you do use such a thing then alternator efficiency is important and you want to get it as high as possible.

For the common direct battery charging approach you have to be careful. If you could make your alternator 100% efficient the overall system performance would be dreadful with blades hard stalled.

Assuming Betz is right then 100% efficiency is the betz limit of 59%. We then have to keep the blades on top of the power curve to keep the prop efficiency as high as possible for a mppt machine. For conventional loading we have to compromise alternator efficiency to keep the prop reasonably well up on its curve.

It takes a long time to get your head round the fact that you need to introduce losses to increase power out. I keep telling people to add series resistance in the line to increase power out and I suspect that many never even try it because it seems a crazy idea.

For rectifier loads you also need to forget sine waves and most of the equations used for normal ac power measurements. Testing a machine into a resistive load using conventional alternator testing procedures will give you little indication of its performance when clamped to a fixed voltage via a rectifier.

By all means do all the alternator tests, at least you learn a lot about alternators that may be of use if you ever go down the mppt route but bench testing with no idea of input power will only tell you how much you can hope to get out within the thermal limits of the machine. If you find you can get 1kW you will get it as long as the blades can produce the input power to do it but get the matching wrong and you may need blades way bigger than needed for a matched system.Or they may have to be smaller to get the speed and you need 90mph to get to the power level.

Anything that encourages people to try to useful tests benefits all of us in the end as long as we know the exact test conditions and can decide how relevant they are to the matter in hand.

If building a common design that is known to work then checking cut in speed and shorting the thing and seeing that it is hard and smooth to turn are sufficient tests to confirm that you have done nothing wrong. Any other tests can be regarded as fun and I wouldn't change the design based on results without knowing the full implications of any changes.

Flux

[Ungrounded Lightning Rod]

Everyone uses the word "efficiency" but very very few could deliver a technically correct definition of the term.

Usually people think in terms of efficiency as power out / power in or power out / fuel in.  But when the fuel is essentially free (as with wind power), the correct way to measure efficiency is power out / cost in, where "cost" is money, labor, time, land, ...

Thus it may make sense to use a somewhat less efficient design (in the wind-to-watts sense) if it gives you a lot more watts/buck, watts/hours of labor (including maintenance).

Thus the simple, inexpensive, easy-to-build, no-critical-tolerances windcharger designs you see a lot of here - tuned to be reasonably power efficient at the low end of the common winds (when you need the power most) at the cost of falling off in power efficiency when the wind rises (when you have more wind power than you need if it keeps up) and finally deliberately turning themselves down or off (when you have more wind power than you can handle).

You could do better, power efficiency-wise.  But it would cost you money, time, and reliability.  So we "satisfice" (do well enough to meet a goal and don't bother to do better) rather than "maximize".

Now if you were designing a mill to sell power to the grid (which has an arbitrary capacity for usefully absorbing power), rather than just powering an offgrid site, you might find your cost/benefit peak at a higher efficiency.  You could sell more power that way.