Improving axial flux alternator reliability?

[gizmo]

It seams like every week or so there is a story about another burned out axial alternator. A alternator capable of 1000 to 2000 watts can only supply a fraction of that under continuous use. At full capacity it will destroy itself very quickly. For those who have bought the expensive magnets and spent the many hours making the stator it must be depressing to discover the thing burned out because the furling was not right.

So just thinking out loud.....

The magnets are acting against the copper coils as they pass by. So in a windmill under load, the coils are taking the full load of the turbine, right? Now this isn't a constant load on each coil, its actually pulsing as the magnets pass by. Are the copper coils held securely in place? Only if they are impregnated with the stator material ( epoxy or whatever ), which is not the case in the stators I've see. Only the outside layer of copper is secured with the epoxy, and as I see it, the copper coils in the middle of the coil are not held securely in place, and will be vibrating against each other. No matter how tightly they were wound, they would be moving around in there, wearing off the enamel coating, shorting against each other and eventually burning out that coil.

Under heavy load, as the stator heats up, the epoxy will get soft. The magnets are trying to pull the copper coils through the stator material, distorting the now soft ( compared to cold ) stator. This gets so bad the stator deforms and strikes the magnets, causing surface friction and pending doom.

So how to fix it. Dunno. But have a couple of idea's. First up, those coils need to be impregnated with epoxy, those windings need to be locked solid. Remember those wires not only carry current but act against the magnet flux, so they must be load bearing. I also think the stator needs to be made thicker outside and inside the magnets, see my picture below. The top design is the way the stators are made now, ( Yellow=stator epoxy, Red=Copper Winding, Blue=magnets ), the bottom design has a much thicker stator to give physical strength as the stator heats up and starts to soften. This should stop the stator distorting, and would still be easy to mold.





I also think the bolts used to hold the stator need to be beefed up, slipping a tube over the all-thread will make a much more rigid structure.

Like I said, thinking out loud, could be totally wrong here.

Glenn

[Flux]

Interesting thoughts but I don't share your concerns about wire vibration causing failure, it is purely a question of excess temperature. These things are running near class H at the surface in high winds, the centre of the coil will be way over this.

The simple truth is that if you want to dissipate over a kilowatt in the stator of a small alternator you need a fantastically good way of removing the heat from the coil.

Even slotted core alternators are not normally capable of being run at 40% efficiency continuously but they are far better at dissipating heat and with the intermittent rating and good cooling on a wind turbine they are more robust.

The only way of preventing burn out is to control the output to stay within the thermal rating of the stator.

All this coil rubbing is nothing more than the final stage of destruction of a winding that is already burnt out.

Either make it furl and not be greedy or make some attempt at matching the load so that the alternator efficiency can be kept up if you want crazy output in high winds.

The other simple solution ( but not cheap) is to design the alternator for 70% efficiency at full load and add resistance in the line to get it out of stall. The heat then goes somewhere other than into the stator and sometimes you may be able to use it to advantage.

Flux

[David HK]

Your not wrong by any means, but I have another point of view which will induce some thinking. Its now dark here in Hong Kong and too late to photograph what I wish to illustrate, so can I invite interested parties to view this entry again in about 24 hours time.

David HK

[scoraigwind]

I accept that some of my designs are prone to overheating, notably the 12 footer in the 2005 plans.  In my recipe book I have gone for larger diameter alternators with more copper and the heating per unit area of surface area is way down.  I do not anticipate any further problems with burned stators under these new standards.

David's cross sections seem to show increased coverage of the coils with resin which would be counter-productive.

I do not recommend using epoxy to encapsulate stators.  Vinyl Ester seems to be the best so far.

[tecker]

I think  it's a Question of balance of a three phase configuration ( classic problem with all three phase circuits ) . Not a problem when working below rating of the wire and connectors but when over as Glenn points out the stator could warp the wire might scuff off insulation .A mess of things happen . That's another reason to bring the star point out of the stator casting and into your connection can easy to maintain and easy delta or series or just monitor . I think all problems with over amperage could be controlled on the ground by manipulating the the stator configuration .

[elt]

Hi Glen,

I did that, partially by accident, partially by choice. At one point I was going to make my stator 5/8" thick. After making the mold, I decided to go thinner. When I cast the stator, I put a round piece of 3/32" plywood in on top of the coils; the diameter was smaller than the mold so that the outer inch was still 5/8" thick. The thicker rim is visible if you know it's there in the pictures I posted when I made my stator but it's not obvious... here's a enlargement, I hope it is more clear:





That was so easy that I can't imagine not doing it again.

I also wanted the vinyl ester infused through my coils. I made a "pre-preg" by soaking the coils in resin and then using a vacuum bag to draw the air out:

I think the best thing I did, though, was taking Flux's advice to match the the load to the steep and powerful part of the curve instead of chasing itsy-bitsy wind. Because of that, my alternator makes less heat in higher winds.

None of that saves you from a 50 mph wind; the mill has to limit it's power output by furling or it will be killed. Between cut-in and death, though, I think there's still room to get some heat out of the stator. I'm studying both the notion of a buck-converter to take extra power from the alternator and an idea to switching in a load ahead of the diodes to do the same thing.. I'm looking at both ideas but don't know enough get to know which direction to try. (Yet!)

- Ed.

[David HK]

First of all let me say that the following observations are in no way critical of Hugh Piggott. I merely state what is written in his 2005 book.

In Hugh's book he makes it quite clear that the constructor should choose either Imperial or Metric - I chose Metric.

Taking the coil wider first, the recommendation is a spacer width of 13mm and a length of 46mm and a breadth of 30mm. I followed exactly and wound ten coils as required. Page 25.

Next is the stator casting for which all the dimensions are given including the thickness which is specified as 13 mm. Page 29 Col midway down.

After all the resin was poured and dried I unveiled my stator and was very interested to see that most (not all) of the `start' wires on the coils showed a white glistening line where they passed over the coil turns.

Photo below.

My thought are:-  the coil winding is 13mm thick, but the start wire which runs over the over coils is 1.6mm thick plus microns for insulation which makes the overall stator thickness 14.6mm plus a fraction. The fibreglass chopped strand mat also has thickness, but this no doubt varies when it is wet or dry.

So I am left wondering how good the copper wire insulation is when the `start' wire passes over 6, 7 or 8 wires forming the coils.

When I cast my second stator I deliberately slipped some heat-shrink tube over the start wire where it crosses the coils and have not had a similar repeat of the top fibreglass mat showing glistening lines.

Flux, Hugh, and many more experienced minds will no doubt be able to offer more expert advice, but could this be a possible source of coil failure?

It seems as if I am being pedantic, but I merely raise the observation for opinion and comment.

David HK

<

[gizmo]

That vac bag is a good idea. It would be interresting to have two axial alternators side by side connected to an engine, one with a conventionally made stator and one with a "vac bag" made stator. Both stators run up under load and allowed to "cook" so see how well the modifications help, if they do at all.

Glenn

[Flux]

As someone brought up in the days of cotton covered wire I never do things like this but in reality the modern wires are so good that that you can get away with it on a low voltage machine. It makes a lot of sense to slip a piece of Nomex or Kapton under the wire just as you would have to do in a high voltage machine.

Sleeving is ok but I object to the space wasted by this crossing wire anyway, I think sleeving makes it thicker and wastes more space. ( I never bring wires out like this but it is good enough for everyone else).

I suspect that when the wire is severely burnt the breakdown here may be the final straw in the failure mechanism. At low voltage I am inclined to suspect that many machines are running burnt out and the voltage stress between turns is held off by burnt mess for months or years. This highly stressed point will have a full coil voltage and may fail first.

In the pictures we have seen recently the machine was still running with smoke coming out, After shorting to stop it it failed to run and I suspect the mechanical stress from the braking forced the turns apart and it rubbed the magnets and one coil disintegrated.

Flux

[Flux]

The idea of impregnating the coil is a good one but although it may slightly prolong the running of a drastically overloaded machine it will not cure the problem.

Running stators side by side with different attempts at cooling would be a better idea as I really don't think this area has been explored in real tests. There has been a lot of speculation about various filers and holes in coils and lots of other things but no real useful test results.

Wire we know will run at 200C so that is the limit at the centre of the coil. It would be useful basing things on this to see what surface temperatures this equates to with various cooling methods. If the limits can be got to the point where vinyl ester resin could survive and retain reasonable strength with the centre turns of the coil at 200C then that is your limit.

If the vinyl ester will not hold up at the point where the centre turns reach 200C then you need to work at a lower temperature.

Unfortunately these will be life tests not instantaneous ones and will take a lot of effort. Short term you may be able to exceed these limits and I am totally convinced that many machines exceed these limits at least for short times. It may not result in failure but it will shorten lifetime.

I really have no idea where the limit lies, I found with polyester the stuff is stinking at 60C and I stopped testing at a surface temperature of 60 directly on the area over the wire of the coil. I based my rating on this and assume that with air cooling it will not reach this figure with my loading.

With this limit I have never tried the expensive vinyl ester but perhaps pushing the surface temperature to perhaps 100C may not be too bad but you would need to check the temperature of the centre turns of the coil, they will be the limiting factor.

Measuring temperature rise by resistance will only give average temperature and a rtd or similar would be needed to check centre turn temperature.

As I said previously I leave this field of research to those who are struggling with it. I have had no problems with burn out on any of my machines with my methods of loading so I see no reason to do this research for others.

To be honest I see nothing wrong with the simple approach and I also see no need to have burn outs, it is nothing more difficult than keeping the furling point safe. If you want more output then there are ways to increase it but always work within the stator limitations and there will be no problem.

If you build things from plans then follow the plans, everyone seems to want to change something or improve something and this is fine if you know what you are doing but such people don't need plans anyway.

Apart from one machine that Hugh mentioned I think his designs are conservatively rated and if you follow the instructions exactly you will be ok.

When Dan publishes his book I suspect the same will be true. There may be a slight problem following all the stuff on Otherpower as it is evolving all the time and it would be wise to use the latest data, whatever you do don't mix and match bits of early and later designs. Changes in one thing affect another.

I have no experience to confirm this but using a 10ft machine using the 2 x 1 x 1/2 magnets I would not let the output exceed 600W.If you find it exceeding this do something very quickly or wait for the inevitable. That may be 10 years away on one site or 2 days on another.

This burn out nonsense is not necessary, the axial design is excellent and is the best way to make good use of low winds. If you want enormous power once or twice a year then you need a different approach. If you are in a consistently high wind area and you can use lots of kW during the windy period without boiling your batteries then the simple design optimised for very low winds is not the thing you should be using. It suits most people with normal loads in normal wind conditions who normally have excess power on very windy days. Don't spoil it by failing to furl at a safe limit.

Flux

[gizmo]

Maybe a different approach. Like you say, if the windmill is kept under control with furling, etc, then the stator will last may years. Its the little unexpected events that can cause problems.

What about a microtemp? These may have a different name in the USA, its basically a $2 temperature fuse. Often used in heating applicances, they come in a range of temperatures, can handle several amps, and are about 10mm long, 4mm wide. If a microtemp was embeded in the center of the coil, and the stator heat rose too high, the microtemp would go open circuit. This would save the stator, the microtemp could be dug out and replaces with a new one, but this also means an unloaded windmill that may self destruct in a different way, ie throw a blade.  

Again just thinking out loud.

[veewee77]

Plan B would be to embed one of those temp switches like used in appliances that is a NO one and when the temp gets high enough, it just shorts the phase. It'd take 3 of them on a 3 phase machine, but it would lock itself down and go into stall unless it was very poorly designed or if the blades were way too powerful for the stator.

JMHO - YMMV

Doug

[zeusmorg]

 There are several ways to improve the current axial flux designs, one of them being heat related. However, would they just add to design complexity and costs at little or no benefit?

 I could see several methods of ways to improve this. Even though vinyl ester is the potting component preferred, it's ability to dissipate heat is very low. So we could improve that by adding non-conducting elements that do dissipate heat at the possibility of weakening the structure, also the costs related may be prohibitive.

 It would also be possible to increase air flow to the stator by drilling angled holes through the rotor plates and having air passing through the stator. This would possibly reduce the heat up to the point additional power could be utilized at higher wind speeds. A lot of electric motors do use fan cooling to dissipate excessive heat.

 It would also be possible to install a thermistor and design a failsafe furling deign at the cost of more complication and more initial build costs..

 I suppose it comes down to either finding a simple cost effective way of cooling the stator with a good design, or just sticking with the current ones and furl early enough to keep the stator from self-destructing.

 I'm doubting increasing the stator thickness in the non-coil areas would improve anything dramatically, and may even hold in the heat better. Not allowing it to dissipate heat like the current design would, to me, it would just increase the costs.

[ghurd]

I agree with Flux.  Furl early.

What is the point of NOT furling early?

The wind here runs near nothing for 3 to 6 days, then too much for about 2.5 days.

I expect most fairly balanced battery systems would be full voltage before the wind gets to max, and certainly before it peters out at the end of the windy 2.5 days.

I expect my wind is like everyone else's wind. Either too much or not enough.

I believe it was Dan who said his windmills still make 70% max power while furled. My batteries would not be at a low voltage well before that, let alone need more than 70% of peak power at that time.

Later furling would mostly be an excersize in building larger capacity dump loads.

It is just theory for me.  I never flew one that furls.

G-

[DanB]

A couple comments...

of course I have to agree and learn a bit from flux on just about everything.  The 'burnout issue'... most ideas to improve things Ive seen here would make very little difference in my opinion.  These machines run pretty well/reliably down to about 50% efficiency and any alternator will die somewhere around there if  not sooner - so this is  not a problem with the design - the trouble is folks building alternators too small and expecting too much from them.  If you want more  power build a more powerful alternator with lower resistance - if you can live with out the high winds then furl early enough.  There are surely other solutions - perhaps limiting the current with inductors but that might get expensive and introduce other problems.  (its something Im playing with a bit)

One small comment about vinyl ester for flux - you say it's expensive but I actually find it to be quite cheap, between $30-$40/gallon  - it is only slighty higher in cost than the cheapest of polyester resins.  I think it holds up well to almost 200 deg C.  I think it's much more rigid and much less prone to warping than polyester.  All that said though, polyester works fine for stators - not my first choice though.

[Flux]

Dan, thanks for the comments on vinyl ester. In large quantity I don't think the price is too bad here. It most likely has been around longer in N America and your supply is better.

I can get it at a reasonable price to build about 10 machines but for a one off I would have most left over and I believe it has rather a limited shelf life. I would love to try it and may in fact do so on the next machine just for the experience but I shall not need the higher temperature.

For those building machines that are going to run hot then it seems logical to use it if cost is not prohibitive. I suppose when considering that it can be used for the magnet rotors as well then having to buy a larger quantity is not so bad, but the quantities here at present are too large for one machine for me to play with. If I was dependent on wind power then I would accept the extra cost just as I would accept the cost of more magnets and copper to build a more robust machine.

Well done for pointing out that this is not a design problem, I have rambled on for months implying that and complaining about people not furling at safe speeds and basically implied that it is not a design problem. It's time someone put it in print and dispelled this common notion that these machines are going to burn out.

If anyone builds one decently and runs it within its limits it will last as long as the magnets will survive without corrosion if you maintain bearings and other things.

Flux

[Beaufort]

Has anyone looked at the how much heat affects the length of the coil (thermal expansion)?  With encapsulated stators, there is nowhere for this expansion to go.  For example, a 24V machine with a 150 turn coil has about 34 meters of magnet wire.  Copper has a linear coefficient of thermal expansion (alpha L) of 17 x 10^-6/C at 20 deg C.  For a 15 deg C change in temperature:

    Delta L = Original L * alpha L * Delta T

            = (34)(17 x 10^-6)(15)

            = 8.67 mm (0.34")

Can someone check these numbers?  Seems very large, but many locations will see a 50 deg C swing in ambient temperature not to mention any heating effects from running.  So if these numbers are right, what's going on inside a stator with this change in length?  All the more reason to keep the temps down for these types of stators.