Three-phase bridge SCR control

[jonpowers]

I want to use SCR's to act as a buck converter. 60 amp SCR's are reasonable in price and isolated versions can just mount to a common heat sink.

My idea is to have the wind turbine alternator cut in at low wind and buck the output to keep tip speed up and not stall.  I could make the efficiency of the alternator very high such that it would stall in direct battery connection. I would trigger the SCR bridge between 15 per cent conduction at start up to 80 per cent conduction or so at top end.    Shockley diode characteristics of SCR will lower rectification losses. 

Has anyone tried this?

Some parts I haven't figured out (I am a B S in mechanical for a reason).  The triggering criteria to make the power vs. RPM come out right could be voltage based (zener diode, resistor, capacitor) as when the rpm rises the trigger would get sooner in the cycle.  I have seen some FET based trigger circuits to get the trigger phase correct or trigger transformers.

Anyone have some ideas how this could be done?

Jon Powers

[joestue]

in theory this works.
there's a few problems

1) the load isn't resistive, or inductive.
2) the ac supply is non ideal, rather high resistance.
3) the current drawn increases as the as the firing angle decreases, because the current increases with the wind speed squared.
4) wind turbines are typically low frequency.
5) rms current in = rms current out.

due to all these factors you have to insert a rather large dc choke in series with the battery so that the current is continuous, because well, if its non continuous then the rms current is that much higher than average. the incoming ac line will have to be filtered with a rather large LC filter. otherwise the harmonic current drawn by such a controled rectifier will burn the alternator up.
said LC filters will have to be rather large.. large enough that due to the high resistance of said alternator, it won't start up in low wind.

much easer and cheaper just to boost the ac to a high voltage dc and send that into a dc-dc converter.
end result is controled rectifiers would be sutable for a situation where the current decreases as the firing angle decreases.. at least linearly.

[jonpowers]

joestue:

Please help me understand.   It seems to me that the batteries would not be poorly affected if the current comes in pulses (ripple is always present in a rectifier circuit).  An electron will make a chemical reaction when it arrives at the battery at any time.

How can the alternator heat up if the on time of the scr very small.  Will not the power delivered be small and current squared times the coil resistance also be small?

It would seem to me that the condition where the scr fires at the beginning of the sine wave is analogous to a standard diode and we know that works so I am confused why firing later would do any thing but reduce the load on the alternator and achieve my goal.


You say, "much easier and cheaper just to boost the ac to a high voltage dc and send that into a dc-dc converter. "

That sounds to me like building a full inverter system like a modern utility scale turbine uses and it seems much more complex then triggering some scr's,  but if it is "much easier and cheaper" please line me out on some schematics. Thank You
Jon Powers
 

[ghurd]

Might find this interesting, Flux's load matching posts,
http://fieldlines.com/board/index.php/topic,143631.0.html
G-

[joestue]

joestue:
Please help me understand.   It seems to me that the batteries would not be poorly affected if the current comes in pulses (ripple is always present in a rectifier circuit).  An electron will make a chemical reaction when it arrives at the battery at any time.

true, the batteries don't care.

How can the alternator heat up if the on time of the scr very small.  Will not the power delivered be small and current squared times the coil resistance also be small?

no, the power won't be small.
the firing angle has to decrease as the wind picks up and the voltage climbs with the rpm.
but the current climbs with rpm^2 and as the rpm climbs you have to reduce the firing angle.
SCRs based controllers do not increase the rms current flowing out of them unless you switched to forced commutation based methods.

It would seem to me that the condition where the scr fires at the beginning of the sine wave is analogous to a standard diode and we know that works so I am confused why firing later would do any thing but reduce the load on the alternator and achieve my goal.

it will reduce the *load* compared to a simple rectifier feeding a battery, but this is a wind turbine, you have to look at the torque-rpm curve and ask yourself what is the current at said operating point based on the performance data on said turbine.

You say, "much easier and cheaper just to boost the ac to a high voltage dc and send that into a dc-dc converter. "
That sounds to me like building a full inverter system like a modern utility scale turbine uses and it seems much more complex then triggering some scr's,  but if it is "much easier and cheaper" please line me out on some schematics. Thank You

i wouldn't say its easier.. its doable. as i said earlier, the LC filters needed to mitigate the very high harmonic current drawn by a phase controlled rectifier may prevent the wind turbine from starting up at all in low wind. its quite likely that you will have to switch in the capacitors as the load increases. that will require triac based switches and some way to switch them on at either the zero crossing or have a resistor network to keep the caps charged to ac peak and switch them on at the peaks.  the rms current flowing through those caps isn't trivial.

I don't think IGBTs are that much more expensive than SCRs now, so i see no reason to even think about using them.
3 relatively small inductors and capacitors would let you pwm a fully controlled rectifier at 20-60khz and get near zero no-load losses and a fully controlled, 100% power factor system.
cheaper still is to just rectify to dc, send that to a buck converter, adding interleaved phases to reduce the output capacitor requirements.

the reason i prefer boosting the voltage is because as the voltage increases and the power increases with the cube of the windspeed, the power requirements to boost said voltage decrease, not in real terms but in per unit terms. it ends up being cheaper compared to a buck regulator and a low voltage system.
doesn't change the fact that you still have to convert the high voltage bus to something usable, but you would have to do that with a low voltage system anyway.

[jonpowers]

Posted by: joestue said:
i wouldn't say its easier.. its doable.

I don't have the electrical design ability to do as you suggest.  Fluxes load matching posts cut to the issue I am addressing but he too is beyond my ability to visualize and build.
 
To repeat what my idea is :  Cut in alternator at a very low wind speed.  As wind picks up open circuit the A.C. connection in a Commutated modulation style and let the rpm rise to stay out of stall.  This will take out of the blades just what is available to stay at a 7 to 1 (or other value selected) tip speed ratio. My perspective is that the typical rectifiers used by direct connected battery charging machines are acting the same as a scr triggered when the phase voltage reaches the battery voltage plus the diode drop.  I will trigger the SCR later in the cycle to "buck" the output to match the blade output.

As i have only convinced one person (my self) that this is a way to run a battery charging direct connected alternator,  I will open a new topic as a follow up to tell others what happens when I try it.

Jon Powers

[oztules]

Jon,

I wouldn't try and be sensible about this, as there are too numerous a number of things to take into account.

Instinct tells me that being as it is a wind driven system, we get things happening that we weren't expecting, and we expect some things, and others happen instead.

I feel it is worth looking at in practice. If we chop the input with a chopper... be it scr/triac/fet/light switch with your hand etc etc, then at all those moments where in normal circumstances/electronics  you would lose some of your input power (eg light dimmer.... you get dimmer lights!), in this case, all the time the switch is off, you actually accumulate momentum and acceleration in your blades.....and maybe get into a better power band as a bonus.

It may well be that by storing the power in inertia in the off pulses, you may get a double bonus..... the prop will increase speed, the momentum stored can be accessed later, and the tsr can be run closer to ideal.

.........control of the chopping may be the sticky bit.


The other fly in the ointment may be runaway, but other than that, I think it is worth a dabble on the dark side, and an overspeed "turn them all on" included in the control circuits may solve this too.

The only serious negative I see thus far, is that it is not done as a matter of course at the moment, and usually there are good reasons for this. I have pondered just such a system previously, but never built it as I did not need too..... too much power as it is.


Happy to see your trials and their results, as I'm not unknown for doing things a bit different.


Best of luck with it.



................oztules

[Ungrounded Lightning Rod]

The mechanical load goes with the average of the torque.
Torque is proportional to current (given a fixed field, i.e. permanent magnets).
Resistive losses and the related heating goes with the SQUARE of the current.

So by attempting to control your torque by switching the current you're
turning a low current over a long period to higher current pulses over
short times.  This increases the stator heating and resistive losses for a
given RPM and horsepower level.

So you decrease efficiency and increase heating and burnout risk.

= = = =

If you want to do switching regulation of load to control the torque, track
max power point, or whatever, do it AFTER the rectifier.
Hang a big capacitor there and follow it with a classic driven-by-wild-DC
regulator / buck converter.  That way the generator/rectifier will see
something close to a constant voltage load.  You can run the switcher
at some high frequency (something you can't do with SCRs as controled
rectifiers unless you use gate-turn-off devices and a controller that can
use them) so you can use reasonably small components
(especially the input filter caps and inductor) and have low resistive
losses in the added inductor.

If you want to improve things for the genny even more you could use
small filter caps, let the input ripple, and compensate with the switcher
changing duty cycle over the ripple humps.  That will reduce the spikiness
of the current waveform and reduce stator heating a bit further - though
probably not enough to be worth the related complications.

[oztules]

Yes,

and it's all true, and what Joestue alluded to as well....... but.....

It is in the context of a windmill, and there is a lot  lot more at play than idealized generators to consider.

We know that with some windmills, the addition of a resistor will improve output into the batteries markedly.... even though the efficiency may go to hell. On a different machine, resistors will not help at all, or even degrade performance significantly. One thing I have learned thus far, is that with windmills, you get to write your own ending. There is a fair chance that what works for your bucket of bolts, won't work with other peoples of similar design even.

Consider a machine with a fair amount of stator, and early furling....It is a sedate machine, well inside it's limits because we don't want headline power, as we have realized it is a curse not a blessing............ it's all different now.

We care not about the heat from the higher current pulses of less duty cycle, because we can easily handle them thermally..... so not interested in this side of the problem in this instance.

We can get out of stall from duty cycle change, we can get into stall from varying the same thing. We can turn it on to 100% duty cycle when ever we wish, we may only use light duty cycle until there is enough wind to get well away from where hard stall would have been, and then we go to 100% duty cycle when we start to get nearer furling.

We may use the higher duty cycle to instigate earlier furling when we need to , and delay it when we don't. It all depends on the requirements we place on the design in question.

If we just want to make a poor mans buck to mimic the Classic, then winding for higher cut in (half the turns used normally) and boosting the low end would be preferable for the novice.... but if we are just modifying what we have already...... anything is worth a go.

I have come to the conclusion there is no right or wrong way with windmills, as just about anything will work to some extent, then finessing it further may take you into different scenarios for different machines. I still love using  transformers.... but most here don't, or tell me why they won't work very well if at all...... and yet they work brilliantly in practice  our HV machines ......despite those opinions.

Make no mistake, I do agree with your theory though.





................oztules

[RP]

This is really intriguing.  I'm struggling to get my head wrapped around one of the outcomes ULR was talking about though.

"So by attempting to control your torque by switching the current you're turning a low current over a long period to higher current pulses over short times.  This increases the stator heating and resistive losses for a given RPM and horsepower level."

On the one hand twice the heat for half the duty cycle would probably be okay at the frequencies involved but twice the CURRENT means 4X the heat because the losses are I^2 R.  At 50% duty cycle then you'd have 4X the heat for half the time or 2X total heat

Add to that, since the only thing this technique can do is LOWER the resulting voltage and current from the stator it seems like you'd tend to wire the stator with a higher voltage winding which implies more turns of smaller gauge wire which further increases losses in the stator.

Someone mentioned a light dimmer and that might be an easy way to test this on a low power single phase alternator (before or after the rectifier)
 
I hope you try this out because I'm really curious.   

[oztules]

It's worse than you think really......

At cut in, we are down to a very small duty cycle anyway........ ie we are only using the tops of the cycles, as the rest is below battV. So effectively, we are disregarding all the rest of the wave form below those points below BattV ... so we are clipping the waveform from the get go. The current will be (emf-battv)/R of the stator (axial flux now).... and then only while the wave form EMF is above battV (plus diode loss etc).  With the three phase this will be happening until peak ac is at least 130% over BattV..... then it will still be a mess, but at least conduction of some description will be happening all the time from here on.

So if your fiddling the start up area, the bogeyman is not as real as you think. It's a shambles anyway, and the stator will handle it without problem as we are miles below the current required to cause problems. So this is the interesting area, and where we are most likely to be switching low duty to get out of stall, and where the overhead design can handle it best anyway. In the midrange we still have plenty of room to play with, and if we achieve whatever we were after in this region, then all will remain fine from a heating perspective.

It is only when we start to get up into the higher range that the problems may appear, and I suspect we will go for full conduction anyway, not because of the heat problem, but because we probably want to stall regulate (forgive me Sparweb) and then furl.... so we may well not be clipping the waveform when it would be heat problematic anyway.

Like I said, you get to write your own ending to some extent with battery charging with a wind generator,  one change here or there designed to do a particular thing may end up changing a dozen parameters you didn't give consideration to.

Throwing a dimmer on any old genny, may not tell you much at all, just as tossing resistance on any alternator may give you results, not gotten by others. You need to be very careful before you write something in or off with these things. Jerry is good at designing gennies that confound everyone, and can get results that even baffle Flux sometimes.

With wind it is easy to do an experiment, and draw the incorrect conclusion too. (Jerry and his wind tunnel experiments??)

I too would like some other guinea pig to do this properly, but in the end, I fear the worst. It is generally not done for a reason, but you can bet there will be an application somewhere with a certain windmill combo and duty cycle..... that  will make a crappy mill work like a star..... just to annoy the rest of us.




.................oztules

[joestue]

We care not about the heat from the higher current pulses of less duty cycle, because we can easily handle them thermally..... so not interested in this side of the problem in this instance.

its still too easy to burn up the alternator with a controled rectifier, reguardless how de-rated it is.

hook up a dimmer switch to the 120vac line, put an 8 amp fuse inline with the 120vac.
connect the dimmer to a full wave bridge, and insert a 1.5 ohm resistor. this represents a 1 kilowatt alternator that is 90% efficient, rectified to dc. Connect that to a 48 volt battery.
now try and get 8 amps into the battery and let me know how much heat that resistor is dumping.
do the same with a 3 ohm resistor for a modest 80% efficient machine.
drop the battery voltage to 36 volts, or increase it to 72 and try it again.

you'll quickly find out just how much this really bad current distortion is actually costing you.

[SparWeb]

It is only when we start to get up into the higher range that the problems may appear, and I suspect we will go for full conduction anyway, not because of the heat problem, but because we probably want to stall regulate (forgive me Sparweb) and then furl.... so we may well not be clipping the waveform when it would be heat problematic anyway.
.................oztules

Ahh if only my motor-conversions were "stiff enough" to stall-regulate in strong winds!

Anyway I'm chiming in because I remember asking the same question on the forum several years ago.  Tried to look for the old thread, but no luck.  As I recall, all the same concerns raised here were raised back then, too. 

Jon,
I'm not big on electrickery either.  As I was trying to figure it out, I drew for myself drawings of sine-wave after sine-wave, and then modifed the drawings for the cut-in situation and again for the higher power situation too, to see the difference.  Then drew a corresponding curve for current to match each scenario.  Maybe that's when the "lights turn on" in one's head.  It worked for me.  Then go back and consider the instantaneous responses of some components, which produce spikes, and other effects like inductance which cause even more complicated things to happen.  Go as far with the subject as you can, then come back to it someday later and maybe more of it will make sense.

If you do give it a try, we're very interested in hearing about what happens.

[ghurd]

Oz and Spar,
I agree 110%.

What I think will happen,
what should happen,
what I am told will happen,
and what actually happens,
are often totally different.

The similarity to a Haiku is purely coincidental,
G-

[oztules]

Joestue,

Until you build a "wind transducer" you won't appreciate what is going on.
Efficiency is of little interest, (don't choke just yet) Most of us are trying to match the commonest winds to our batteries. This is a different approach than what an engineer will do. Every part is a compromise, all aimed at getting the most from a specific range. Now you may think "efficiency is the key" as a good engineer, but it's not.  If you need low wind performance, big blades, big alternator, and furl early..... throw the power away, you dont have to pay for it, and you don't want it. You don't need efficiency, you need watts into the batteries when YOU need it..... this is not when the wind blows hard as a general rule.  So efficiency and what we want is not always related as strongly as you may think.

My stator has phase resistance of .39 ohms in star..... no quite the 1.5R you allude to. It drives a 48v bank. It has 628 sq inches to radiate heat.... a bit more than the 1.5R resistor I suspect that you were thinking about. When it is working harder, it has wind chill to boot..... so I can afford a bit of loss, and not give a hoot in the low to midwind range. I need to introduce inefficiencies, or the thing won't work... it will stall. It runs near cold at 1kw, but I don't need more than that..... I want less....... more frequently/all the time. I don't want or care about efficiency. I want the impossible.... power all the time.... not lots, just 700 watts is fine..... yes I built 2 x 4 meter ones. I want to see 300- 400 watts most of the time out of each.

If I could have  achieved  that by scr control, I would have.... but due to the design, and wind here I easily achieve that without it, but I have oodles of room for heat if I so choose.

So I overbuild to hell, and then hobble it severely, doing whatever it takes to take the most advantage of low winds to mild to mid winds. I don't want to see big winds, or of I do see them, I want the mill to ignore them.... not take advantage of them...... not what a decent engineer would do.

An alternator of 0R resistance will not work with  a prop.... cut in = stall..... finished. So we need to compromise .......... we need losses in the system, just to work. Bench testing is only to get an idea of the alternator power curve.... simple stuff...... Then you need to match it to the wind...... variable impedance driving force to  a fixed impedance (battery) not so simple..... the blades and the alt must provide the slop to match the wind to the battery..... whatever it takes.

Like I said, I agree with the theory, but I won't write anything off with windmills, until it is thoroughly tried. Because there is no formula for mill matching, there will be mills that this may help more than hinder.... and thats all it takes to be useful. There will be plenty of mills it wont help at all.... but thats normal for mills.

Only those who don't "get it" with their furling will burn up a stator.

If you think that is all bollocks, then go build a serious mill and see for yourself... all is not as it seems.



................oztules

EDIT: If I had seen Ghurds post first, I would have not bothered to write this drivel... he said it better than me.

[joestue]

I'm not sure what compromises OP was looking to make.. cost-control-complexity tradeoff wise.
But is anyone seriously thinking that 6 scrs are going to be easier to control than a boost or buck converter?

Oztules, the resistor values are nominal.
The point of such an excursion is to show that in an average situation with efficiencies at nominal values (such as 90%), with low cut in and an immediate need to cut the voltage in half and then in thirds destroy any gains quicker than you might think.
Might as well built a set of relays to short out .1, .2, .4, .8, 1.6 ohm resistors in series, because this phase control isn't any better without massive lc filters in front and a dc link reactor between the rectifiers and the battery.

[oztules]

Quote from me "Make no mistake, I do agree with your theory though."
So no point me arguing that part.

Where we differ is what will it do to the mill. I don't quite know. I would like to, but I don't know for sure. Unlike the resistor scenario, we get some inertia build up in the rotor which may just be the ticket at certain rpms that we want. If that turns out to be a plus at the rpm we wanted/needed, then thats all that counts, and if it only costs efficiency and heat, but we get a gain where we want..... worth the experiment. I would expect we will be in the 70-100% range, rather than the 33-50% range. We should only need to finesse it, not beat it to death.

I have easily controlled kilowatts with triacs and opto drivers.... I have only built forward,  boost and halfbridge designs up to about 800 watts. I find them non-trivial to keep alive above these levels. I did do a 1.5kw pushpull, but I copied someone else's design. Wind mills are notorious at smashing folks choppers into smoldering ruins. If I could (and if I had need) to be able to better match the mill at the rpm I needed most, and I could do it with robust bits of silicon, rather than fets, I'd look seriously at that route.

I think better to design the alt for what you need in the first place, but I have no qualms in experimenting for experimentation's sake.... it is a hobby after all. Sensible folks would use solar cells. ( build them for 60-70 cents or so per watt.... cheaper than a mill ).

"Might as well built a set of relays to short out .1, .2, .4, .8, 1.6 ohm resistors in series, because this phase control isn't any better without massive lc filters in front and a dc link reactor between the rectifiers and the battery." ... we still get conservation of energy in the prop, not available with resistors.... could be the deal maker.

Better to switch taps on a transformer with a set of triacs not relays......, although we get excellent results  direct driving leaky transformers...... like a battery charger transformer. ( no I don't agree with scoraigwind using toroids for wind for this reason coupling is too tight, and leads to stall problems)

Interesting side note. I have repaired some large battery chargers (traction use) that used triacs to change from  24v to 36v  (via phase control) and fine tune. From memory, there was no LC in the front end worth noticing, and standard rectification to batteries after the transformer (using those crummy 35Amp blocks).... so it is not unknown.

We won't know for sure until someone tries it..... and even then,it will depend on the components and the control system, and what is trying to be achieved. Full control, or supplemental over part of the range I'm interested in.




..................oztules

[bob g]

why use a full set of scr's?

better to just use the top half diodes,and replace the bottom half with mosfets
then control the mosfets.

no need to control them with some phase angle scheme typical of scr control of three phase because we don't need three phase as a finished project, just control them as a group with some sort of pwm scheme

of course i am neglecting all the other concerns alluded to by others, just thought i would kick this out for consideration

bob g

[oztules]

Bob,

you have just described a boost converter...... as in Flux's treatise

http://fieldlines.com/board/index.php/topic,127288.18.html


..............oztules

[bob g]

my intention was to describe a sort of buck converter scheme, however
for it to work the alternator would either have to be iron cored or there would
have to be some sort of inductors following the stator to work against?

perhaps some large caps?

bob g

[BigBreaker]

It has been said before but this EE just has to say it again.

Low frequency, wild AC is a huge pain.  Too slow to voltage transform, filter or use directly.  Step one, rectify and do it early - top of the mill or base.  Big BEEFY rectifiers.  Now you have slow rippling DC which you can actually work with.  Trying to chop 3 Hz is a disaster.  Don't chop AC you wouldn't use a transformer.

I like high voltage mills bucked down to battery voltages in terms of the electrical engineering.  It's not as good for stall control b/c of a higher resistance stator but the diode losses and transmission losses are lower.  Higher voltage is usually easier to deal with than higher current but once you get out of MOSFET territory and into IGBTs where the gate driving gets complicated and power intensive.  Voltage goes up linearly; it's current that goes as the square of wind speed, so you have a bit more of a ceiling on voltage (but have a back up plan!).  There are practical considerations that argue for boost, especially if you have little use for infrequent high generation periods.

I like my boost or buck convertor near the batteries and after the transmission line.  Why?  Because you can use the transmission line as a shunt.  The voltage drop on the line can be sensed at very high input resistance IE low current.  That means a small cable next to your power cable is all that is needed to know your current.

[jonpowers]

Posted by: joestue "But is anyone seriously thinking that 6 scrs are going to be easier to control than a boost or buck converter? "

Yes that's exactly what I think!

Posted by: oztules
"We care not about the heat from the higher current pulses of less duty cycle, because we can easily handle them thermally..... so not interested in this side of the problem in this instance.

We can get out of stall from duty cycle change, we can get into stall from varying the same thing. We can turn it on to 100% duty cycle when ever we wish, we may only use light duty cycle until there is enough wind to get well away from where hard stall would have been, and then we go to 100% duty cycle when we start to get nearer furling. "  And  "I feel it is worth looking at in practice. If we chop the input with a chopper... be it scr/triac/fet/light switch with your hand etc etc, then at all those moments where in normal circumstances/electronics  you would lose some of your input power (eg light dimmer.... you get dimmer lights!), in this case, all the time the switch is off, you actually accumulate momentum and acceleration in your blades.....and maybe get into a better power band as a bonus.

It may well be that by storing the power in inertia in the off pulses, you may get a double bonus..... the prop will increase speed, the momentum stored can be accessed later, and the tsr can be run closer to ideal.

.........control of the chopping may be the sticky bit."

You hit the nail on the head, you understand where I am thinking.  I do agree that the trigger methodology is the hard part but several of the above posts bring out the results desired so it is just a matter of finding the criteria that produce that result.  I just received a bag O' silicon from Newark and will report back later.
Thanks for everones help!
Jon Powers 

[oztules]

"It has been said before but this EE just has to say it again.

Low frequency, wild AC is a huge pain.  Too slow to voltage transform, filter or use directly."


..............then you'd be wrong.


We use it here and have done so for many years, Hugh Piggott uses transformers as well, and finds then very good ( even though I think he needs more leakage)

If it works in the real world, and the most prominent man in small wind power in the world uses them personally and has done for years......

Simply size the thing for your required max output. It will work perfectly throughout the entire range.... with headroom to spare at the low frequencies. Think carefully about it as to why..... it's a windmill.   It needs to handle full power at F max, it needs to handle 1/8th power at F/2 and you should get the rest from here. As Flux has pointed out, if you design for low magnetizing currents, you need not even switch them in and out.

If it works well, it works well what can I say.

If F is very low, we use more iron, and microwave transformers are free.... It's a trifling matter to wind your own with these things.

I'm not qualified to carry an engineers suitcase for him..... but I know what works...... (and I'm still trying all the things that dont.... just in case..)


Bob, I think Flux covers this.

Jon,
We will all await with interest. Like I said, depends what your trying to achieve, as to how well it may help your best winds output. (not max winds output)

.................oztules

[joestue]

"Might as well built a set of relays to short out .1, .2, .4, .8, 1.6 ohm resistors in series, because this phase control isn't any better without massive lc filters in front and a dc link reactor between the rectifiers and the battery." ... we still get conservation of energy in the prop, not available with resistors.... could be the deal maker.

Perhaps i wasn't clear. The heat dissipated in those resistors is equal* to the heat dissipated in the resistance of the turbine when you use phase control with no other energy storage networks.
Its the same thing as putting a chopper after the diode block (a single switch running at low frequency, ZERO energy storage)

once you throw inductors and caps in there to buffer the turbine.. or the LC filter and diode of a buck converter.. then you get energy storage and then you can get more average current into the batteries than you are pulling rms current from the turbine.

I've covered transformers in other posts in the past. they are perfectly suitable, but they tend to be rather large.. if they are free though..
Its possible to make 95% efficient transformers from MOTs, just have to keep the flux density down below 1.2T and pack the whole thing full of copper.

*this may not be exactly true. the current increase that you get from letting the turbine run faster is because there is more torque available, not due to conservation of energy.. there is none. the turbine is a voltage source and a resistance, not an energy storage network. it has no capacitance.

If you switch to a buck converter, you'll get both the benifit of more torque (=current) and more voltage. multiply the two together to get a lot more power than from a phase controlled system.

[oztules]

"*this may not be exactly true. the current increase that you get from letting the turbine run faster is because there is more torque available, not due to conservation of energy.. there is none. the turbine is a voltage source and a resistance, not an energy storage network. it has no capacitance.

If you switch to a buck converter, you'll get both the benifit of more torque (=current) and more voltage. multiply the two together to get a lot more power than from a phase controlled system."

Not so sure about this.

If we let it run faster (<100% duty cycle), then EMF will by necessity be higher... so EMF-batV/R becomes bigger = stator current. For the same input power wind, the TSR can vary wildly depending on the loading, and how far away from design TSR we are...... even in a fixed wind, as we change the loading TSR will change, as will power either side of the design TSR. No load maybe 100% increase (2xEMF), overload=stall (not much EMF)..... Wind still the same.

So it does become a sort of energy storage network. If we don't use the energy. it  turns up in the TSR.... until drag messes it all up.

I agree with your Buck statement, but would never use one myself.... seems silly relying on electrickery when delivering full power and above when the furl re-adjusts. Not for me.... boost yes, buck no.

We've probably done this to death, and I still don't have a fix on the outcome. It needs to be tested. Too many variables for my head.




.........................oztules

[BigBreaker]

oztules, glad to hear it's working for Hugh.  Different strokes for different folks I guess.  You'd get a D from your professor if you used that much iron in your analog design lab project.  Ha!


I looked up what Hugh is doing: high voltage stator and transforming DOWN to feed 12v batteries.  That helps matching - I get it.  The transform ratio (down) would increase as the voltage and RPM increased.  That would keep the blades out of stall for longer.  In my own defense, I had 3Hz in mind while hugh's design settles in around 12Hz.  Big difference.

You could multi-tap the transformer for greater control.  Switching speeds would be seconds rather than thousandths of a second.

Ok consider me converted.

[Ungrounded Lightning Rod]

Efficiency is of little interest, (don't choke just yet) Most of us are trying to match the commonest winds to our batteries. This is a different approach than what an engineer will do.

Quite the contrary:  It's EXACTLY what a GOOD engineer does:  ignore the distractions and design for the REAL target.

Efficiency is usually a very high priority, because losses usually mean money thrown away.  But when the "fuel" is "free" (except for the cost of building and maintaining the hardware to collect it), efficiency takes a back seat to other factors.  Factors like capital cost, reliability (which implies simplicity), matching the available resources, and meeting the important targets (like not running out of power in times of light winds).

[Flux]

Just caught the tail end of this, I haven't read it in detail.

The SCR phase control bridge may help but it comes with a lot of problems that will be difficult to solve. I have never seen a variable frequency firing circuit, the normal type of firing circuit doesn't lend itself to easy modification to variable frequency. I have some ideas for doing it but never got round to trying.

SCR's don't commutate easily on trivial supplies, tiny reactive alternators will raise lots of issues. Cheap nasty scr's have other issues that will cause you additional problems with dv/dt . The conversion works fairly well in relation to dc mean current but rms values will be way higher than you expect. Your best chance would be with a mixed bridge ( 3 diodes 3 thyristors) but the mixed bridge does strange things at times.

I have had some thoughts over the years but I haven't done enough trials to see if the idea is really worth pursuing. For those not up to speed with modern switchmode techniques the idea of using thyristors at a frequency where a bit of wire is still that and not an inductor you may find layout and other things far less critical but don't underestimate the difficulties of dealing with a 3 phase controlled bridge on non stiff supplies.

Good luck if you try it, I think it has more hope for a heating controller than a battery charger.

Flux

[SparWeb]

Efficiency is of little interest, (don't choke just yet) Most of us are trying to match the commonest winds to our batteries. This is a different approach than what an engineer will do.

Quite the contrary:  It's EXACTLY what a GOOD engineer does:  ignore the distractions and design for the REAL target.

Oh, good, I thought I was going to have to stick up the the engineers all by myself!

...instead I'll just stand over here with the poet laureate.

[Ungrounded Lightning Rod]

"It has been said before but this EE just has to say it again.

Low frequency, wild AC is a huge pain.  Too slow to voltage transform, filter or use directly."

..............then you'd be wrong.

We use it here and have done so for many years, Hugh Piggott uses transformers as well, and finds then very good ( even though I think he needs more leakage)

The concern with transformers and low frequencies is core saturation.  Applied voltage is opposed by the inductance of the cored coil, where voltage causes current to ramp up and ramping up current in an inductor causes an opposing voltage which limits the the slope of the ramp-up.  Once the core saturates the inductance suddenly drops to the free-air inductance of the coil, causing the ramp to become very steep.  Current increases enormously, heating increases with the square of the current, and the transformer is likely to be damaged, even catching fire.  With a constant voltage a lower frequency means a higher current, so a sufficiently low frequency (with constant AC voltage) means saturation and runaway current.

However:

Transformers are a great match for both iron-cored generators and permanent-magnet coreless alternators.  This is because, as the frequency drops, the voltage also drops, resulting in essentially constant magnetization of the cores at the cycle peak.

Another way to look at it is that the generator is a "transformer" with the varying magnetic field through the "secondary" coils produced by spinning magnets rather than by varying current in a primary coil.  The flux cutting the current in the generator/alternator's output coils produces a current which is conducted into the primary of the transformer, forcing a corresponding flux through the coils of the primary and into or out of the transformer core.  Thus the magnetic field in the transformer core is proportional to the magnetic field through the coils of the generator (times the turns ratio of the transformer primary coil(s) over the generator coil(s)).  If the transformer core has enough material that it doesn't saturate at a field than corresponds to that in the generator at some normal operating frequency it will not saturate at ANY frequency.  The field in a permanent magnetic alternator is constant and the maximum excitation field in an excited alternator is limited by the saturation of the core through the excitation coils.

Or at least that's what would be the case if the coils and interconnecting wire was superconducting.  Since it does have resistance, at low frequencies the resistance causes the output current to decay - and the lower the frequency the longer it has to decay.  So the magnetization of the core in a transformer driven by a PMA actually falls off from the "proportional to the generator's field" level as the frequency gets very low, approaching zero as Hz approaches zero.  There's also a falloff at high frequencies, where reactance limiting causes the field to not fully penetrate the coils in the generators.  But the proportionality holds very well, resulting in an essentially frequency-independent peak magnetization of the transformer core, for a wide range of frequencies.  This typically includes the frequency for which the candidate transformers are rated.

So figure out the open-circuit voltage of your PMA at the RPM producing the frequency that the transformers you want to use are rated at, and pick one that is rated for at least that input voltage or a few percent higher.  That transformer won't saturate at any RPM and will work fine over essentially any RPM where the genny would produce usable power (and won't produce enough voltage to puncture the insulation of the transformer windings - which usually isn't an issue for a genny.)

[Ungrounded Lightning Rod]

... the current increase that you get from letting the turbine run faster is because there is more torque available, not due to conservation of energy.. there is none. the turbine is a voltage source and a resistance, not an energy storage network. it has no capacitance.

Now that is just not true.  The turbine is a massive flywheel coupled to a generator.  The mass of the turbine is a giant energy storage device.  If the generator were DC the mass would map exactly to a hysterically large capacitor.  It does the moral equivalent with an AC device analyzed over multiple cycles - and for a partial-cycle analysis it is still close to the enormous capacitor equivalent.

If you only allow current for part of the cycle you're only pulling energy out of the flywheel for part of the cycle - while the wind is pumping energy into it for the full cycle.  To resist the wind and hold the same RPM you must apply a proportionally higher torque during the shortened active portion of the cycle, which corresponds to a proportionally higher current in the coils.

(Current and torque are proportional so average current remains the same regardless of duty cycle.  But resistive losses and heating go with current squared and we're back to the "shorter duty cycle is less efficient due to higher RMS current" business.)

= = = =

By the way:  Another reason to keep the switching regulation after a rectifier and filter capacitor:  Radio interference.  You CAN do switching regulator tricks with the equivalent inductance of the generator.  But doing so puts current and voltage waveforms with sharp edges into tens or hundreds of feet of elevated wire that is NOT RF shielded.  Good luck listening to radio or watching off-the-air TV.  Also, good luck avoiding reprisals from authorities if you have any neighbors who also want to do this.

[oztules]

............ at the risk of being volatile.......


"If you only allow current for part of the cycle you're only pulling energy out of the flywheel for part of the cycle - while the wind is pumping energy into it for the full cycle.  To resist the wind and hold the same RPM you must apply a proportionally higher torque during the shortened active portion of the cycle, which corresponds to a proportionally higher current in the coils.

(Current and torque are proportional so average current remains the same regardless of duty cycle.  But resistive losses and heating go with current squared and we're back to the "shorter duty cycle is less efficient due to higher RMS current" business.)"

There are some problems with this as I see it........ ( I agree that it does not work like Joestru says either as I explained above )

I think it is instructive at this stage to look at what is really happening....

"If you only allow current for part of the cycle you're only pulling energy out of the flywheel for part of the cycle - while the wind is pumping energy into it for the full cycle."

I agree with it up to here..... but now we differ. Because we have a fixed R, we can only see EMF-BattV/R for our current in the stator and so into the load.

Unless we mystically decrease R we CANT increase current (torque) unless we increase RPM or  EMF or TSR (same things for constant wind input). I fail to see how we can increase the torque without first raising the TSR, as R will hold it to the max value we started with with full conduction at the original TSR or RPM or EMF.

If I'm right, we can't even increase instantaneous torque/current without first increasing TSR.

In this case if rpm does not increase (and it WILL with less than 100% duty), current can't increase. Until (and while) the TSR catches up with the lesser load current, the new current figure cant increase magically for a shorter period in order to have the same output. Only when TSR increases to the new "level" reflecting the lower duty cycle can we close the gap and have the same output as we started with, and still use a smaller duty cycle.... so it will have hysterisis too.

"Current and torque are proportional so average current remains the same regardless of duty cycle."...... only if the TSR rises so the EMF is high enough to compensate for the shorter duty cycle.


So:

We can clip, ..... so we relieve the load on the mill  (decrease continuous torque ) until it climbs to a higher RPM..... then with a higher EMF, we can get higher continuous current  through the stator.... but not when we started the clipping.

Now can we benefit from this?

In fact as has been said, we lose efficiency in the stator, ....but.... we may now get into a better position for the blades to do their work with far more efficiency  than the losses we sustain in the stator from the higher current spurts with the higher EMF. (similar to  the old resistor in line trick)

Thats the only reason I can see for doing this in the first place..... to fine tune the mill, to move it's operating specs up or down the wind scale, and how it behaves in those winds.

It will only be useful if there is mismatching in the first place. Thats why I said a single test will not tell the tale, you need the right mill to do it to to get an improvement.

For the above reasons, it will never be a buck substitute, only a tuning and control mechanism perhaps.

I still think windmills still have a lot to teach peasants like  me.



I'll probably get a beating..... but it seems right to me.



..............oztules

[joestue]

whether or not the wind turbine stores any energy... it certainly does...
BUT ELECTRICALLY IT DOES NOT.. it is a voltage source and a resistance... there is no capacitance, inductance is trivial at 10hz line.

you can pull 3 times as much current out of it at 33% duty cycle.
if when pulling nominal current out at 100% duty cycle you drop 10% in the resistance,
then switching to 3x for 33% you are now losing 30% of your available power.
drop the former to 80% efficiency and now you have a 40% efficient machine!


FLUX's comments that the scrs won't commutate.. yes, they will. electrically the turbine and the transmission line look like inductors.
what will happen is that you'll have overlap, and the time required for phase 2 to increase in current and let phase 1's SCRs turn off will be longer than optimal.
but once you put some very nice big delta caps on it this is gone. I would not use junkyard variety 1980's SCRs.. but you don't need inverter grade.

Ii think you'll need a DSP to find the zero crossing, or you'll have an unstable system.

I am not concerned about using phase control to get your turbine from tsr 5 to tsr 6 to get a little more current.
as i said before in another thread, to get peak power you have to run the turbine above optimal blade TSR. how much faster depends on the efficiency of the generator.

What i am concerned about, and indeed, just about saying its impossible, is to bring your turbine out of stall (tsr 2-3) and get it up to 8 where it belongs.
you can't do this with phase control, but your alternator could certaily handle it with a resistive load.