Axial Flux generator: maximum RPM?

[lifer]

Hello,

I'm using a transmission belt between the turbine (VAWT) shaft and the axial flux generator. My current transmission ratio it's about 2.5 but I want to rise it. I'm using a buck-boost converter between the generator and the battery string so I could handle any voltage between 0 and 200V or so.

I wonder if there's a speed limit (RPM) for the generator rotor disks. I'm not talking about mechanical limits (magnets flying off) but if the output voltage it's reaching a maximum value at a specific RPM (magnetic field limitations).

Beside, does anyone know if the voltage output (versus RPM) has a linear characteristic till that maximum point?

Any help is highly appreciated.

[Flux]

There will be a speed limit depending on the method of attaching the magnets. If you start looking at speeds above 1000 rpm you will need much better magnet attachment than the usual epoxy but for a VAWT you will need a lot of speed increase to exceed the speed of a fast HAWT with a 5ft prop.

The open circuit voltage is linear with speed but on load the effect of reactance becomes more significant at high speed but the power goes up accordingly so you can't loose out with more speed especially when optimally loaded with a converter.

Flux

[lifer]

Thanks for your answer, Flux!

I've been using UHU Plus two-part epoxy resin to fix the magnets in place (a pretty strong adhesive I believe).

I'm going to try with a transmission ratio of 6 and I'll check the rotors behaviour.

[joestue]

just drill a hole at the edge of the disk and insert a dowel pin.

air core machines have only eddy current to drag you down, for 15 gauge wire you might notice it around 200 hz and higher, and if you don't fill the gap between the magnets and have air holes in the disk then the centrifugal fan you just made will have likely at least an order of magnitude air drag over eddy current. you might even have to perform the test in a vacuum to get an accurate figure for eddy current below 1000 rpm.

[lifer]

It's a very handy solution but I don't have enough space to drill a hole (there's only 2-3 mm left on the outside).

The only possibility for further mechanical fixing is to mount an outer stainless steel ring (as used by the Otherpower guys on their AFPM generator for the 20" rotor).

Anyway, I'm very confident with that epoxy adhesive (and I don't think I'll go over 1000 RPM neither).

[lifer]

I've just changed the pulleys for a transmission ratio of 8 but now I have a big problem: the VAWT doesn't start spinning at low/medium winds at all. I even disconnected the load with no success.

But if I disconnect the generator (removing the belt), it starts spinning at no winds. I guess the PMA generator is oversized for my current turbine (I've planned a bigger one when I designed the generator). 

Still it's quite a strange behaviour as I could easily rotate the generator shaft. Is there any trick to make it work?

(In the mean time, I've just started to learn the transmission belt theory. Maybe a toothed belt could help, by reducing the belt tension?)

[DamonHD]

I'm no expert at all, but as there is very little energy in low winds all you've done I guess is add enough friction (and/or other losses, etc) to the system to prevent start-up until higher winds.

Rgds

Damon

[TDC]

(In the mean time, I've just started to learn the transmission belt theory. Maybe a toothed belt could help, by reducing the belt tension?)

Possibly, synchronous or toothed belts are very flexible. Most people would be quite surprised how thin they are and the amount of power they can handle.  I'm speculating you changed to a smaller generator pulley and that increased the bending resistance/drag of the belt. What size pulleys and type of belt are you using?  RPM?  BTW, I'd love to see some pictures of your project. 

[lifer]

I'm using an intermediate shaft between turbine and generator (the generator is located 2m bellow turbine bottom).

I'm using a 2:1 transmission ratio at the VAWT level (20cm/10cm diameter pulleys) and a 4:1 transmission rate at generator level (40cm/10cm diameter pulley). I'm using 13x8 mm belt (picture bellow) for both transmission lines.



If I use 1:1 transmission ratio at generator level it spins even in low winds but the output voltage is too low for direct battery charging (24 V).

Anyway, I'm just building a buck/boost converter to charge the batteries so the output voltage is not supposed to be an issue BUT.. there are some limitation dictated by the coils wire amp handling.

I've used 1.7 mm (AVG 14) wire for coil windings so the limiting current is about 25 A. For a 24 V direct charging that's about 1 kW maximum power. I need to get more than that (especially in high winds and for braking purpose) so I thought about 100-150 V (2-3 kW). 

I couldn't speak about RPM as I don't have a rpm meter right now. I'll try to quickly build one by connecting a hall sensor to an arduino (used as MCU for the buck/boost charger).

[TDC]

Without knowing more about your turbine, I would guess the v-belts are a large part of the problem.   A synchronous belt would be more efficient and turn a lot easier, but they need more precise tension and parallel shafts.  I think your best options are probably roller chain or single stage synchronous belt. The 8mm pitch belt is probably a reasonable size and quite common.  Search for belt and pulley names like GT2, HTD, polychain and powergrip.  A big diameter pulley won't be cheap, but I suspect a simple flat pulley might work on the turbine if large enough diameter.  For such low rpm, you can't get the drive pulley to large.

[lifer]

My turbine it's a Lenz2 model, 2m x 1.6m (H x W).

Thanks for your suggestions, I think I'll go with the synchronous belts. Though I thought that the flat v-belt "slip" feature could better protect the generator in high winds.

I'll try to make some measurements with various transmission ratios then I'll choose a synchronous belt/pulley accordingly.

[Vertical]

I have posted a new topic that has to do with RPM... you might want to take a look at it... I am sure it will help you obtaining more RPM's in low winds... here is a summary...

The benefit of a counter rotating vertical axis wind turbine has been demonstrated over a single rotating vertical impeller wind turbine when using slow wind.

The results were found to be as high as six times more output. No investigation has been done for motors or impellers rotating at optimal speeds.

Having one VAWT positioned above the other in the middle between the two VAWT's a PMG is positioned in such a way that one WAVT rotates the Rotors while at the same time having the other VAWT counter rotate the Stator…

[lifer]

Thanks for passing by!

I still prefere to keep it simple (avoiding the slip rings). Beside, I think there could be some interferences between two closer (counter rotating) turbines. I guess I'll manage to get more power from my generator by only using transmission belts.

Anyway, I have to finish the buck-boost converter to regulate the output voltage (I'm in the process of migrating from 24V to 48V, too).

I still have a weird problem though.. I've made some tests few days ago and even the open voltage was over 70V, when I've connected the load (24V battery string) there's was almost no charging current (just half an ampere or something). The turbine RPM has lowered a bit when I've connected the load but still.. just 12W of power (at 200 rpm)??

My gen has 2 x 12 N40 neo magnets (2" x 1" x 3/4") and 9 coils (180 turns x 1.7mm) and I was expecting around 2 kW output at 600 rpm (assuming the turbine will be able to supply that amount of power).       

[Flux]

Need more information.
What is alternator speed at 70v open circuit. What was alternator speed when connected to 24v battery and producing 12W.

I strongly suspect the speed is falling drastically. With 58v dropped and 1/2 A flowing we are looking at 116 ohms internal resistance, not possible with that size wire.

I did see your request in the other post but replying here was less of a hijack.

Flux

[lifer]

Thanks for your quick response, Flux.

Unfortunately, I don't have a rpm-meter (or anemometer) available (noob mistake). I presumably thought the rpm was about 200 (I counted 3 rotations per second).

I'm not familiar with a turbine/generator power curve (maybe I'm just experimenting that cube speed / power relation). I was affraid there was a problem with the generator (bigger airgap or something). Speaking of that, the airgap is about 22mm (for a magnet thickness of 15mm). As far as I read, it's an acceptable value (the stator thickness is 18-19mm).

[Flux]

I have run a few quick figures on the alternator and 70v ties up with about 600 rpm. Running at about 200 rpm also works for cut in about 24v.

The output is limited by the power the turbine is producing. It seems badly matched and would most likely work better at 48v battery but I know nothing about its performance or your wind speed so can't comment there, you may be stalling it.

I believe I have seen mention of a buck converter and that would match the alternator to the battery. For now can you manage a temporary 48v battery to get a better match.

Flux

[lifer]

I guess you're absolutely right. The generator is mounted in a shelter (underneath the turbine) so I had no visual contact with the turbine while I was taking those measurements.

But now that you mention I remember that I heard a big "knock" at the turbine level when I coupled the load (batteries). Maybe I just stalled it, as you have suggested (though I thought it just spun down a little).

Sorry for the lack of informations - I'll try to be more accurate next time.

Yes, I'm in the process of building a buck-boost converter so i could "play" with the input voltage (by allowing the turbine to catch some real power before starting to charge the batteries).

My target is to get a higher output voltage from the generator (as the thickness of coils wire is a current limiting factor). I could handle 150V input voltage (or even more, if I replace the MOSFETs with high voltage rated ones).

[TDC]

Unfortunately, I don't have a rpm-meter (or anemometer) available (noob mistake). I presumably thought the rpm was about 200 (I counted 3 rotations per second).

Most digital multi meters have a frequency function, that can be used for rpm.

[lifer]

Good point (actually I could have been used the oscilloscope, too). Anyway, I was thinking of a more portable solution or a more advanced one (a hall sensor connected to an arduino, for continuos logging). I guess the last one will be my homework for this weekend.

[lifer]

Without knowing more about your turbine, I would guess the v-belts are a large part of the problem.   A synchronous belt would be more efficient and turn a lot easier, but they need more precise tension and parallel shafts.  I think your best options are probably roller chain or single stage synchronous belt.

So roller chain it is!

@TDC, you were absolutely right! After further investigations, I did noticed that the v-belt was slipping hard on load. I've tried to find out an optimal belt tension with no success. And there was a lot of "rubber dust" around the pulley - it wouldn't last longer anyway. And, of course, a lot of friction/slipping noises.

Now I've changed every belt/pulley with a single 10B roller chain (2:1 transmission ratio) and everything is ok (including noise). Of course, I had to step up the voltage using a boost converter (at least I finally have something to convert!).

I wrote some kind of MPPT algorithm for the boost converter, by reading the rotors RPM (btw, I placed the Hall sensor along one rotor disk so it's activated by every magnet passing by, for increased accuracy), the generator ouput voltage and the battery charging current.

This way, I gave my VAWT some time to spin up before extracting some useful power and I keep it at resonable speed during normal wind condition. We had some strong gales lately and there was no problem at all (the generator is oversized).

Many thanks to you too, @Flux! You have also suggested roller chains on another thread and now I can confirm: that's the way to go. Sorry for I giving you wrong informations on previous posts - I just didn't noticed at the time that the belt was actually slipping.

[MattM]

You're killing my dream, lifer.  I've been drawing up plans on a hubless rotor using a belt.  My plan was to use a conical drive that could be adjusted to match to turbine speed like a constant velocity joint.  The idea was to gear down the rotational speed of the generator as wind speed picks up.  It wouldn't cure gusts, but it would keep a generator output in a sweet spot longer before furling or braking.

Maybe your belt issue is just in its execution?

[lifer]

It might have been a belt issue but the variable wind speed/power won't help it anyway.

If the belt are too tight, it would't start in low winds; a loose one will lose grip on increasing load.  Maybe a spring-loaded tension pulley could help - I haven't tried that.

[MattM]

You don't really want it spinning in useless wind anyhow, right?  I'm aiming for multiple blades in the two dozen range, no blade longer than a foot on a four foot diameter hub.  Nothing real big, no super electrical production.  I'm having trouble with the parts count is the real issue, and it might just end up with a three-spoke rotor to simplify parts.  But belt drive around the outer diameter of the hub is essential.  Adding a belt tension pulley is going to add to the parts count again.  And what works in miniature for the transmission isn't going to scale adding a second tensioner for the transmission. /facepalm

Some ideas look so good on paper

[lifer]

For a small-scale turbine, belt drive might be feasible (cheap, silent, custom machined pulleys).

I don't know, maybe the belt I was using for my application were undersized. Anyway, the chain drive is _much_ better (no slipping, easy starting).

For your project (conical belt drive) it seems a good choice.. so good luck! It's quite an interesting project.

[joestue]

how hard could it be to make a centrifugally loaded idler pulley?

[Windytech60]

Its been months since I last posted or replied. I have built several OtherPower style axials and made a few improvments to a great design. I created a magnet insert that is attached to the steel rotor disc that retains the magnets in position. I'm adjusting the composit materials to minimize cost to fabricate. The design also incorporates a passive cooling idea to provide greater air circulation on the stator.

Also fabricated several dual stator axials that I would be glad to collaborate with anyone interested in producing one for themselves. It was a reall trick to safely place 24 magnets on both sides of a seel plate.

[ChrisOlson]

I wonder if there's a speed limit (RPM) for the generator rotor disks. I'm not talking about mechanical limits (magnets flying off) but if the output voltage it's reaching a maximum value at a specific RPM (magnetic field limitations).

Beside, does anyone know if the voltage output (versus RPM) has a linear characteristic till that maximum point?

Kind of an old thread.  But no, it's not linear due to reactance in the generator core.  The faster you spin an axial, the more rpm it takes for every open volt almost right from the word "go".  I've spun some pretty good sized axials at 2,000 rpm and blew one up with big ferrite blocks on the rotors.  I had tabs welded on the outside of the rotors thinking I was only dealing with centrifugal forces.  But the reactance load on the magnets ended up exceeding the tensile strength of the epoxy adhesive and when the magnets came off the rotors the explosion was not anything you'd want to be anywhere near.  It cleanly blew a 3/16" thick blast shield right off the machine and put holes in my shop wall.  And then the resulting vibration from the remants of the machine still spinning at close to 2,000 rpm caused it to flop over on its side, and it finished self-destructing, spraying the shop with shrapnel.  I only got one piece of a magnet that went thru my leather boot and embedded itself in my leg.  But that was the end of my experiments with high-speed axial generators.

[lifer]

That's pretty scarry. You did it on purpose (experimenting) still you wasn't prepared. But just imagine it somewhere in your backyard - that puts the whole neighborhood in a real danger.

Anyway, I was thinking of 900 - 1000 rpm and my entire generator is enclosed in a strong case (made of thick perforated metal sheets), along with the chain & gears.

I always wonder how those unprotected HAWT generators  are still in place. I had recently witnessed a big storm and I watched my VAWT (from a close distance) being ready to take over. I was using multiple and strong guy wires so there wasn't a real danger but the feeling was terrible. 

[ChrisOlson]

Anyway, I was thinking of 900 - 1000 rpm and my entire generator is enclosed in a strong case (made of thick perforated metal sheets), along with the chain & gears.

I've run them on geared turbines at 1,000 rpm with no problem.  The one I blew up was being tested for a gas engine driven DC generator.  If I could've gotten ferrite blocks with counter-sunk holes to bolt the magnets to the rotors with stainless steel bolts it probably would've held together.  The weak point was gluing the magnets to the rotors.  Even with the tabs on the outside of the rotors, the epoxy didn't hold.

I ended up giving up on that idea and used a modified Leece-Neville 145A alternator on my DC generator, driven by a Honda GX-160.  The Leece-Neville isn't quite as efficient as an axial, but it puts out tremendous power up to the full hp capacity of the engine and has never failed.  I've tried Delco SI alternators in that application too, and they last about 1 day and it takes the stator out of them.

One of these days I'm going to invent a way to make a core for a home-brew radial that's efficient.  The radial flux units are much better than axials because they make much more efficient use of the magnetic material in the generator.  Axials are too big, too heavy, and too expensive.  The only saving grace for axials is that they're easy for homebrew builders to build.  But they have a lot of flux leakage, make poor use of the effective flux, they're not very efficient from a cost standpoint, and they tend to burn up because the cooling is poor and they don't have enough reactance limiting on current output at high speeds.  After working with axials for several years I became less and less impressed with their limitations and high failure rate.