PMG design considerations for MPPT

[bobfandango]

...and I want to know if anyone has considered how using such a thing could or should effect the design of Piggot style turbines.

As I understand things (and please tell me where I don't), a more "efficient" and/or "powerful" alternator is only possible up to a certain point when the thing is tied directly to a bank of batteries.  That is, for example, more powerful magnets and/or reducing the airgap and/or using extremely low resistance runs of wire all can put the alternator into stall in higher wind.  Adding line resistance allows the turbine to see a higher voltage and stay out of stall a bit longer, but costs something at lower wind speeds.  

Now imagine that you have a device capable of dynamically varying the load the turbine sees.  An MPPT controller like Midnite's Classic is just such a device.  With a controller like that, what should be done differently?  

My first guess is to use more powerful magnets (assuming they are available).  Another would be to use larger stator coil wire (i.e. winding the coils "heavy duty" like the stator the Dans sell online).  The heavier duty winding would help a lot with heat and the stall problem should go away once the MPPT controller learns the optimum operating point of the turbine in the higher wind range.  Reducing air gap within safe mechanical limits might be another way to go.  

The state of the PMG turbine art seems fairly mature to me now having taken into account all the various tradeoffs and arriving at a design that "just works" across a pretty wide operating range.  But it seems to me a controller like this is going to force a reconsideration of the tradeoffs and open up new possibilities.  What are your thoughts?

Thanks!

[electronbaby]

The only thing MPPT will do is allow you to run a higher voltage alternator and make it work well with your lower voltage battery bank over a wider range of operating RPM. It will allow you to squeeze out a little more efficiency at higher RPM's by allowing the stator voltage to climb in a controlled manner.

It will probably allow easier configurations for grid tied systems as well.

As far as the Axial Flux design goes, Im not sure much will change. The operating parameters might change such as the nominal voltage of the stator, but this is nothing different than designing around a 48v system or a 12v system.

So to answer your question, im not sure much will change, others on this board have made MPPT work well before. MidNite just made it easy for you to go buy one. boB is a pretty talented guy, so Im sure his controller will be a hit.  :-)

[halfcrazy]

I would think by allowing the turbine to run freeer if you will by varying the mppt voltage based on windspeed would allow the stator to run a touch cooler as well as a touch more efficient.

Plus with the clipper we would be able to clamp the max voltage of the turbine so there may be a huge market to go grid tied if the clipper can clamp the max voltage and being able to lock the turbine down we would be able to go straight into a grid tied inverter like possibly a windy boy? i think this may be an exciting move for the market.

[Ungrounded Lightning Rod]

One thing we'll be able to do with a suitable MPPT controller is go to higher voltages out of the alternator.  This does nothing special at the alternator.  But it drastically improves things on the wiring from the alternator to the MPPT controller - wihch is probably in the shed with the batteries.

Higher voltages probably help with the MPPT controller design itself:  Some of the voltages have to do with voltage drops in semiconductors, which are essentially constant.  So running higher voltages there means a lower percentage of loss.  Buck converters are easier than combo buck/boost converters.  So running a higher voltage gives you a lower cutin with a pure buck converter, while the MPPT keeps you from paying for that with mismatches at high winds and early stall.  And so on.

I'm hoping this converter can switch modes to current limit (or slightly ramped current limit to take added cooling from higher wind speeds into account) when the wind is high, allowing the genny to collect its maximum safe rating of power from a range of high winds - then perhaps at REALLY high winds (where the RPM and yaw nodding are becoming a hazard) raise the current again to ramp up the drag, trigger mechanical furling, and end up with a current drop (after a momentary spike) from the furled and maybe stalled mill.

[oztules]

Before I had actually built a reasonable sized mill, I was filled with enthusiasm for this style of development.

Now I have a 4m animal to play with, I'm not sure anything with magic smoke enclosed should be attatched to the output to try to control it.

Once some serious wind turns up, all bets are off. The furling (which I have fiddled with endlessly) is never as smooth as I had imagined. Being close to the ground (3m test stand) means it cops considerably more abuse than it would when higher, but nevertheless it shows that their are times in the furling cycle, when things go awol for very short periods, when wind re-aligns just as it decides to unfurl and over correct. In these periods, even with a fairly stall limited load, it can put out scary current for very short periods.

In these circumstances, I don't hold out a lot of hope for a logic controller to guess at what is happening, and it may be increasing the impedance on the alt (because of the furl), just as it goes ballistic, (because of the wind shift and accompanying gust that seems to come with it).

I have seen it wipe out a 275amp curtis controller because of a poor battery connection (allowed the voltage to rise to the hundreds of volts for a split second) in a large gust, and it would have to handle at least 6kw bursts if it was to allow the mill to follow it's curve (because we are not stall limited now).... all for a modest 1kw rating when things are going fine.

Last I saw, the midnight thinggy was running on a 180v or so mill, I never found out the results, as I haven't rechecked their site, but I assume they were good.... but long term, I can see the mill just waiting for a chance or mistake to send out the smoke signals.

It would seem more sensible to me now to just build a slightly bigger mill, to get the same power with a full nights sleep.

I am happy to have a dump load (some electronics) across the batteries, as the batteries should (except for poor connections!) keep things under control, but to put an expensive magic smoke box in series with the mill looks a whole lot less attractive to me now.

Just how I see it now I have seen the fury/power that can be unleashed by a decent sized mill in a decent sized blow. Furling seems only to save you 99.99% of the time... that .01% will turn up.... if only for a second or two... without the electronics in series you should never know, but with it , you probably will find out.

If I really wanted to have HV transmission, I would build a high pole count (at least 24 pole ) and use transformers, and electronics to switch that like the AWP over here... very successful.

........oztules

[Flux]

If you can do mppt properly you should be able to keep the alternator efficiency very high and you should by definition work at the prop peak power point.

With a standard alternator if you let the volts rise with wind speed the efficiency comes out a lot higher than loading into a battery and the stator heating will be reduced a lot. The prop matching will be considerably improved so you gain all round.

Once you choose this option it does make sense to increase the alternator efficiency as high as economically sensible and larger magnets make sense.

If you can keep alternator efficiency over 80% and keep the prop on peak power then things really start to fly in higher winds. The improvement in low winds is less marked but there is always some prop loss from stall almost from cut in.

My experience is that for the same low wind performance near cut in you can probably expect a 3 fold increase in power in the 25mph region over direct stall loading and stator heating is no longer an issue.

My method is a cube law alternator controller not a true tracking mppt device. I am not sure how well you can do with a tracker unless it is a learning type device in which case it would combine my cube law control with the ability to self adjust to changing conditions such as changes in battery volts or prop efficiency from tip damage.

Flux

[halfcrazy]

Flux Thats what i was thinking so if thats the case what do you think one could expect from the 10 ft design turbine that is all ready wound for 48 volts? i will be willing to rewind my stator and most likely will so it can run at a higher voltage

[s4w2099]

Still not an easy deal. If I were to do it I would beef up the MPPT controller at least 4x of the max expected windmill output. Also would take some safety considerations in case the controller still blows up like adding a IGBT or big SCR crowbar circuit BEFORE the controller.

I guess someone will have to risk a mill to test this first. I dont think I would have the guts to do this with a 4m monster.

[Flux]

If this is for battery charging into a 48v battery then the existing winding should be fine. If it is a lower voltage battery then you may still be able to do it depending on the converter design.

All the grid tie inverters seem to incorporate some form of matching from programmed points. I suspect they are not tracking schemes. I share Oztules concern for a continuously tracking device and would prefer a pre programmed scheme or a learning scheme that starts with a basic programme and adapts. I have serious doubts about continuous tracking as is done for solar.

Starting at 48v into a 48v battery the converter is going to have to handle a minimum of 150V and more than likely it ought to handle 200V.

My tests have been at 24v nominal with converter input going up to about 80v.

Except for the line loss issue I still think the boost in low wind and match directly in high wind is the safest approach for self build but the commercial manufacturers should be able to produce a satisfactory converter to do it the other way. I think the grid tie inverters have been around long enough to be basically reliable although most of the early stuff was pitch controlled and far easier to deal with.

Flux

[boB]

My take on this would be to choose approximately the same (nominal) voltage turbine as the battery voltage, except for say, a 12V or 24V battery system in which case the higher nominal voltage turbine may be better... Especially for longer lines.  The Classic will be a higher input voltage unit anyway (>200V) so the higher voltage from the turbine won't matter too much.  (not counting hurricanes.)

Also, I find that the cutin voltage should be, (when nom Vbatt = nom Vturbine), when the turbine voltage becomes somewhat higher than the battery voltage (after a lull in the wind) because, depending on alternator impedance, there isn't any charging power when those voltages are equal, (and not even much if you boost V)... otherwise you are wasting energy just energizing things.

Clippers are another story and should know what the charger is doing to operate correctly.

boB

[halfcrazy]

well i will gladyly use my 10 footer to test the Midnite Classic/Clipper combo as i always say "Ya cant be scared" if she blowes up we will build her bigger and better

[halfcrazy]

After more discussion with Robin at Midnite we will be one of the first couple beta sights for the classic we will stack it up beside the FM60 and see how it compares and then n to the 10 ft turbine. The classic has the ability to learn the wind turbines characteristics and it will learn to run the turbine at its most efficient spot.

Also we will be set up to stream data live to Midnite solar's website also so everyone can see how things are stacking up we are excited and hope that this will bring us a little more efficiency and also controll over the wind turbine.

[rgudgel]

You all are pretty close to what we at MidNite are actually developing. I'm impressed! In a few weeks we should be starting to put some links on our website that show actual data from Hugh Piggot type designs working with our MPPT and Clipper. One of these sites will be two identical 2000 watt beefed up turbines on identical towers next to each other. One will be hooked direct to the battery bank and the other will be through our power electronics. This should be very interesting to see just how the two differ. According to our inhouse built wind turbine simulator, the difference should be impressive, but talk is cheap and our simulator may be inadequate. A third wind data logged beta site is planned for a VAT! That will also be fun. There will be a few 250 VDC Solar data logged sites too, but they are sort of boring.

Robin Gudgel

MidNite Solar

[rgudgel]

It looks like my first attempt at posting vanished, so here goes again.

You are all pretty close to right on the money. The product we are developing does act like some of these discussions. Look for some links on our website in the very near future. These links are to sites that will be data logged and monitored by APRS World for us. We plan on a few boring high voltage solar comparisons and three wind sites. One solar and then wind will be at HalfCrazies house. He has a Piggot machine that will be datalogged and available for all to see on our web. Another wind site is to be in Oregon. It will have a couple Beefed up 2000W Hugh Piggot style designs flying side by side on a hill top. These will have identical towers too so conditions should be close to equal. One system is hooked directly to the batteries and the other will be connected through our MPPT and Clipper. That will be very interesting to see if reality matches what we have seen with our home made wind turbine simulator. The Oregon site will use our 250V Classic MPPT. A third wind site is planned to monitor a VAT turbine. Yes, a VAT!

Should be fun.

Robin Gudgel

MidNite Solar

[boB]

When Robin types "VAT", read that as "VAWT",  actually.

   Close enough for wind work.

boB

[scoraigwind]

It's been a long time coming and I hope it works this time boB :-)

I believe that Ian is playing with the Outback version on a Bergey just now at Guemes island.  Moving on from the truck tests with DanB in 2007.  Slooow development there.

I can't seem to get much technical low-down on either of them - modus operandi etc.

I believe that a 'lookup table' approach to getting a cubic curve will be the best software approach, because 'learning' will be confounded by the effects of inertia, and by the rapid changes you get from gusts and other noise like that.  A couple of key variables for the user to tweak could be fun.

I can see that these controller will improve the efficiency of both the blades and the alternators to allow better energy capture in higher winds.  Probably won't make that much difference at a certain windspeed where the match is already good, but will greatly broaden the range of operating windspeeds and boost the output in stronger winds for sure.

The big question as Oztules has indicated is the ability to retain the smoke in the box.  The 'clipper' concept is key to this and I am sure this is child's play for you guys but a simple voltage clamp is the key item for both this, and for grid tie as has already been pointed out too.

and of course the question of the bottom line, because there is always the old fall back of building a bigger one cheap.

So what is the price and does it include the clipper circuit and when and where do we find these controllers on the market?  Soon I hope :-)

[Flux]

"and of course the question of the bottom line, because there is always the old fall back of building a bigger one cheap."

It always comes back to that in the end. For grid tie then some form of cube law tracking seems essential but I doubt that it needs to be more than 3 programmed steps.

With large powers to handle it would be crazy to load the alternator at constant voltage. Similarly for heating but the match is easier achieved by switching resistors and will be good enough.

Battery charging is a different issue. If you are restricted ( which I was on the buck converter machine)to a small prop diameter then the improvement is spectacular but without this restriction a larger prop would get the same energy capture more cheaply.

The major improvement is in high winds and if you adapt your living conditions to suit this then it is still likely to be a considerable benefit. If you just cook a dump load on windy days then it comes down to a waste of money.

Flux

[Ungrounded Lightning Rod]

I agree but have one nit:

The major improvement is in high winds ...

True.  But potentially distracting from a low wind advantage:  An MPPT controller also lets you tune your mill for low cutin, getting some power in low winds (when it's precious), without sacrificing performance in typical or high winds.

[hmccoy99]

Has anyone tried a mx60 with an input voltage clamp/diverter and a large capasitor > 1  

farad bank to smooth out the voltage surges the with a  small pma  400 watt machine?

henry