Author Topic: Once again in 3-part (phase) harmony  (Read 19663 times)

0 Members and 4 Guests are viewing this topic.

kitestrings

  • Hero Member
  • *****
  • Posts: 1376
Once again in 3-part (phase) harmony
« on: November 07, 2014, 01:40:32 PM »
There's been some discussion in the past on the "growl" or "singing" of the axials.  Ours is still on this relatively low test tower, and the winds are back, so I've been watching and listening a bit more.

Ours has a distinct low-frequency growl just above cut-in.  It is always there to some extent, but seems more pronounced at cut-in.  We are using MPPT via the Charge Controllers, and I don't know if that has any influence.  I'd estimate that it is about 25VAC RMS, 60+ DVC and maybe 13Hz.  And, the sound maybe amplified through the hollow stub pole that it is mounted on and perhaps gong thru some resonant frequency at this particular speed.

My questions are:

   1) Is this a function of loading - causing vibration through the stator when it starts getting hit with, I guess, pulsating loading?

   2) Would it help to load it lighter - changing the pre-programmed load curve - in this range of speeds?

   3) Or, is it related to the rectification of power from 3-ph to DC?

   4) Does the number of poles have any effect?

I don't know that it is doing any harm, just trying to better understand what causes it, and what influences it.  Related discussion:

http://www.fieldlines.com/index.php/topic,144842.msg980835.html#msg980835

~ks

Flux

  • Super Hero Member Plus
  • *******
  • Posts: 6275
Re: Once again in 3-part (phase) harmony
« Reply #1 on: November 07, 2014, 04:47:00 PM »
Interesting.  Basically all electrical machines make noise but is often masked by fan noise in commercial alternators.

There is no doubt that machines feeding rectifiers make much more noise than alternators supplying linear loads. This I  suspect has always been true but probably not recognised. Alternators hum but the frequency is low and it is often masked by cooling fan noise, If engine driven the engine noise tends to totally swamp it.

Those of us old enough to have  had dealings with dynamos know that they whine quite audibly even  above engine and fan noise.Your question has made me suspect that this is exactly the same case, a dynamo is just an alternator with a mechanical rectifier, albeit a many phase one.

Single phase alternators are always noisy, there is constant pulsating which sets up electro mechanical forces and there is noise from magnetostriction, the forces are so high that noise is inevitable. Much of this noise goes away with poly phase machines and in theory a 3 phase alternator is vibration free. I doubt that this is entirely true, the forces all sum to zero over time, but there must still various modes of vibration exciting noises in the core.

Axials don't have iron cores and are a bit different but there have to be exciting forces moving parts of the magnet rotors and stator in various modes.

Immediately you introduce a rectifier things become much worse, with inductive loads, conduction periods are fairly long and things may not be much worse than a linear load, but once you introduce any form of capacitive loading on the rectifier the conduction becomes very rough, occurring in short intervals when the emf is above the capacitor mean voltage. A battery behaves just as a capacitor and is every bit as bad. The worst case is just at cut in when conduction is discontinuous, current just flowing when the peaks of the 6 pulse ripple are above battery volts. Once you get above the rectifier mean ripple things get better. (Cut in is at v peak i.e. 1.414 times v rms, the dc mean is near 1.4 v rms so there is a small band at cut in where this can be most pronounced.

It is not harmful, the forces involved are small compared with the pulsating load of a single phase alternator, but can be annoying. With variable speed machines it would be very lucky if you didn't hit a resonance somewhere in the alternator or some part of the turbine or tower, just bad luck if you hit a point where the worst exciting forces near cut in hit a critical resonance.

With battery charging you can include an inductor between the rectifier and battery and change the rectifier conducting mode, it is rarely worth bothering with 3 phase, but it does really take most of the damaging forces out of a single phase alternator. It also changes the load characteristics very considerably, possibly to advantage if stalled but if running well below stall it may not be so useful.

The change in loading with a 3 phase machine will be small and a much smaller inductor is needed as you are dealing with the 6 pulse ripple of the 3 phase bridge.

With mppt I assume that the input of the inverter is a bank of capacitors so the same thing happens as with a battery. How things will behave with inductors after the rectifier is difficult to predict, in moderation it may be beneficial but id you hit resonance between the inductor and input capacitors all hell may break loose, probably need a lot of inductance to cause this.

If the inverter takes in ac with its rectifier circuits inside you have no control over the rectifier characteristics.

Basically it is normal to have some growling at some speeds, you will struggle to eliminate it, but if it is bad then shifting something off resonance may make drastic changes.

The number of poles affects the ac frequency, it will not eliminate the noise. The main noise will be at 6 times the basic alternator frequency. A very small number of poles may bring the noise lower in frequency where the ear is less sensitive, high pole counts may change the growl to a much more frustrating whine, more like a dynamo.  . Load does change the noise but light loading is always needed near cut in for prop matching. There may be room for experiment with mppt by changing the cut in voltage and moving the curve but I suspect the gain may be minimal.

Flux

clockmanFRA

  • Hero Member
  • *****
  • Posts: 952
  • Country: fr
    • Renewable Energy creation
Re: Once again in 3-part (phase) harmony
« Reply #2 on: November 08, 2014, 04:39:40 AM »
Hiya KS,

As you know I have 3off 3.7m, (12footers) diameter Hugh Piggott design Turbines.

My No1 has a deep Hum (sort of deep moan as the Mrs say's) at cutin and the noise is constant no matter what the speed, air gap is about 0.6mm each side of my coils.
No2, has an air gap at  about 0.9mm and has a gentler quieter Hum.
No3, has an air gap at about 0.75mm and has a Hum noise between the No1 and No2 turbines.

No1, has logged me more power than No2 and No 3 more than No 2.

Note all windings are the same, all casting moulds are the same, all blade designs are the same, the only difference is the air gap.!
Everything is possible, just give me time.

OzInverter man. Normandy France.
http://www.bryanhorology.com/renewable-energy-creation.php

3 Hugh P's 3.7m Wind T's (12 years) .. 5kW PV on 3 Trackers, (8 yrs) .. 9kW PV AC coupled to OzInverter MINI Grid, back charging AC Coupling to 48v 1300ah battery

kitestrings

  • Hero Member
  • *****
  • Posts: 1376
Re: Once again in 3-part (phase) harmony
« Reply #3 on: November 08, 2014, 06:56:15 PM »
"Good, good, gooood, good vibrations....nah, nah, nah, nah, naaaah, nah, nah, nah... "  Oops, sorry, I got distracted there for a moment.

Well, I've got something new to research, because I confess I have no idea what "magnetostriction" is, though the word appears to a melding of magnetic forces and friction.

Much of the rest makes sense to me, in particular:
Quote
once you introduce any form of capacitive loading on the rectifier the conduction becomes very rough, occurring in short intervals when the emf is above the capacitor mean voltage. A battery behaves just as a capacitor and is every bit as bad. The worst case is just at cut in when conduction is discontinuous, current just flowing when the peaks of the 6 pulse ripple are above battery volts.

If I put my ear right against the Sched 40 tube, it really is a rough, drumming of sorts.  There are blade noises that can clearly be linked to a one per rev cycle.  If I had to guess I'd say this was more like a one per magnet pair pulsation.  It does seem less intense as the rpms increase.

Quote
With mppt I assume that the input of the inverter is a bank of capacitors

I'm not sure here if you are more broadly including grid-tie applications, Flux, or mean the converter?  I think you know that we're not grid-tied.  Our system has two, independent stud-mounted diode rectifiers.  Each set of six feeds one of two MS Classic controllers.  The controllers allow the input voltage to free range up to about ~160-170 VDC, and then buck the voltage down to the nominal 48V bank. 

I doubt that I'll take any pre-emptive measures...adding inductors and such, as this is likely closer to this turbine than we'll normally be by 200-300' (not to mention insulated walls and ambient noise of the wind itself).  Still, our power curve is pretty linear from cut-in to furling, and I'm thinking it may help to not be quite so steep out of the gate.  A bit lighter loading early might help both noise at cut-in and blade matching.  I may also try dampening the amplification through the stub tower/tube.

clockman, this is interesting, you've got the benefit or comparison to near identical machines.  Our gap is also fairly close, and the coils sectors are identical, so I'm not thinking there is any unnatural imbalance, or anything necessarily wrong.  This discussion has been helpful to understand better 'what' it is.

If you think about it we have many of the contributors - rectification, capacitance (batteries), current flow in six-pulse peaks at cut-in (as Flux explained), no white-noise cover of drives, fans, motors - and then, we have a hollow tube connecting this variable frequency, energy pounding device to the ground.  It may be then no surprise at all that it springs to life with a "grunt".

Thanks for the feedback.

~ks

joestue

  • Hero Member
  • *****
  • Posts: 1759
  • Country: 00
Re: Once again in 3-part (phase) harmony
« Reply #4 on: November 08, 2014, 10:44:31 PM »
All electrical motors except for the homopolar motor produce a torque ripple.

Hub motors, outrunners, and other small high torque motors are often designed to produce a trapezoidal waveform, and these motors should work well rectified with a single three phase diode block.-but they also have high iron losses and high cogging torque.

it should be possible to build an axial flux, trapezoidal waveform designed for low torque ripple when feeding the usual 6 diode block into a battery without an inductor in between.. but you would probably have to build it with trapezoidal magnets and trapezoidal coils, and that would be difficult and expensive.

A sine wave machine directly rectified into a battery will produce something like 100% torque ripple initially at "cut in"
some of you here may see a marginal efficiency improvement placing an inductor after the bridge rectifier for the worst of cases.
as the power draw increases the torque ripple will drop from 100% to 16%, combined with the frequency increase this is why the problem goes away.

For the exceptionally bad cases.. find yourself a three phase transformer and wind it for a 1:1 wye-delta ratio (which means the actual turns ratio is 1:1.73)
and connect that transformer to the turbine, and the battery or whatever, with its own separate diode block.
such a transformer need only be rated for 57% of the total power demand, and you can play the usual games such as series ac motor run capacitors in series with the primary to prevent startup issues. such a wye to delta transform shifts the phases 30 degrees and you will have a 12 pulse rectifier which delivers sufficiently flat DC with almost perfect sinewave input currents.. torque ripple should be reduced to a maximum of 5% assuming sine wave input.
« Last Edit: November 08, 2014, 10:49:07 PM by joestue »
My wife says I'm not just a different colored rubik's cube, i am a rubik's knot in a cage.

boB

  • Sr. Member
  • ****
  • Posts: 389
  • Country: us
    • boB
Re: Once again in 3-part (phase) harmony
« Reply #5 on: November 09, 2014, 11:16:00 PM »

This growling can happen with axial flux or laminated radial flux or anything.  I've heard it with the Bergey XL.1 too.
What Flux says makes perfect sense.  I've always thought it was at a certain tower resonance but maybe not ?

Either way, a couple ways it could be reduced or eliminated would be with a power factor corrected boost
converter that draws current proportional to AC output voltage.

An inductor as Joe mentioned might help...

An idea I had that ~might~ work somewhat is to just make the controller skip over the parts of the power
curve that cause it to growl. i.e., unload it until the voltage rises some. But then, as soon as the turbine is
loaded again it might slow down and growl again but it would hit that curve point and unload again.
It would be interesting to try !  The wind speed never sits still so it might actually be less irritating.

The Classic is definitely a capacitive input device so inverter or controller, pretty much the same thing
compared to a battery load.

boB

SparWeb

  • Global Moderator
  • Super Hero Member Plus
  • *****
  • Posts: 5452
  • Country: ca
    • Wind Turbine Project Field Notes
Re: Once again in 3-part (phase) harmony
« Reply #6 on: December 01, 2014, 11:04:00 PM »
You've just made me think of an interesting experiment to try.  I recently joined the 21st century, and got a real smaartphone.  I have downloaded a few interesting apps.  Among them are vibration loggers, some of which do an instantaneous FFT to extract the dominant noise frequencies.  Somehow you can get these for free (he says in total disbelief).  I've played with them a bit, mostly in the car (bumpy road vs. highway), but why not tape it to the WT tower for a minute and see what I get?

Something else you can get (also for free) are software that can turn your smartphone into a digital storage oscilloscope (with the use of a modified microphone, but don't forget the voltage divider resistors).  Anyway the sum of this is that you can measure both the waveform and the vibration using just a smartphone.

Knowing the dominant vibration frequencies, and comparing to the dominant frequencies in the output waveform, wouldn't that be informative about the relationship between the two?
Sure, there's damping and resonance to consider, but measurements are necessary before we wonder about that stuff.  I'm going to got try it...  next time it's windy.

I've had a bit of everything as far as vibration is concerned:

-Slight blade imbalance, visible as the tail waving once per rev.
-Generator cogging, because I'm running motor conversions with iron stators.
-Magnet come loose, that's really noisy.
-Free play between hub riding on tower stub.
-Bearing lubricant freezing, making a scrape sound once every multiple of rotations based on inner/outer bearing race velocity.
-Tower found a harmonic until I realized the middle guy wires were loose.

I'm sure if I think about it longer, I'll come up with more.  Currently the only annoying noises are blade tip swoosh, and the excessive number of guy wires I have on the tower.  They actually make more noise than the generator itself.
No one believes the theory except the one who developed it. Everyone believes the experiment except the one who ran it.
System spec: 135w BP multicrystalline panels, Xantrex C40, DIY 10ft (3m) diameter wind turbine, Tri-Star TS60, 800AH x 24V AGM Battery, Xantrex SW4024
www.sparweb.ca

DamonHD

  • Administrator
  • Super Hero Member Plus
  • *****
  • Posts: 4125
  • Country: gb
    • Earth Notes
Re: Once again in 3-part (phase) harmony
« Reply #7 on: December 02, 2014, 02:46:10 AM »
I've stood right under the blades of a 2MW and the blade swoosh was much quieter than the whine from the electronics in the base of the tower (inverter from no-gearbox mechanics?)...  No guy ropes there!

Rgds

Damon
Podcast: https://www.earth.org.uk/SECTION_podcast.html

@DamonHD@mastodon.social

Ungrounded Lightning Rod

  • SuperHero Member
  • ******
  • Posts: 2865
Re: Once again in 3-part (phase) harmony
« Reply #8 on: December 02, 2014, 07:34:27 PM »
IMHO the source of noise that starts at cutin, on a rectified permanent magnet alternator feeding a battery load, is obvious:  Clipping.

As the permanent magnet spins up from stopped, the waveform voltage rises - but there's no current yet.  As it reaches and passes cutin the voltage rises enough that current goes through the diodes, but only at the peaks of the waveform.  Current corresponds to torque resisting the turning of the blades.  So these short pulses of current become short pulses of torque.  They shake the whole mill - by intermittently placing a retarding force on the blades and an equal-but-opposite reaction force on the tower via the frame of the alternator.  Short spikes with long spaces between them correspond to lots of harmonic content, so while the fundamental frequency is near six times the cyclic rate, there's lots of sound energy in frequencies many times that.

As the mill speeds up further the current peaks increase - but also spread out, until they join and the valleys also start coming up.  It never goes away, but the variation becomes less severe.  As the pulses spread out the higher frequency components become less prominent as well - more than making up for the overall increase in frequency of the fundamental.  The basic growl rises in pitch but becomes less brassy.

There are other ways that current and magnetic variations transfer into sound, but torque variations are the biggie for a noise that appears at cutin.  (Among the others are vibrations of the coils from varying mag fields due to varying current applying varying forces to the wires, magnetostriction in the material of magnetic flux paths with cyclic flux variations, and torque variations from pole structures moving past each other.)

= = = =

"Magnetostriction":  When you magnetize metals it distorts their crystal structure and makes them change size slightly.  The effect is small, but very strong.

(I recall a fiction story, half a century ago, where a runaway giant gyroscope - due to loss of field excitation in a wound-field DC motor {lowering the torque but raising the speed where the reverse EMF balances the drive voltage} making it speed up drastically and dangerously - was brought under control by throwing a few turns of wire around a mounting beam and magnetizing it, causing it to expand enough to squash a bearing.  Struck me as somewhat unrealistic and a good way to weld the bearing {or bend the support and fail}, but hey, it was fiction - and a welded bearing would have caused substantially less damage than a disintegrating flywheel in the fictional situation.  It also struck me as a transparent ploy to let the author to explain magnetostriction to his readership.  B-)  )

electronbaby

  • Sr. Member
  • ****
  • Posts: 407
  • Country: us
    • Windsine.org
Re: Once again in 3-part (phase) harmony
« Reply #9 on: December 10, 2014, 11:01:40 PM »
From what Ive seen, a higher pole count seems to make this issue less noticeable. It is definitely caused by clipping with regard to the rectifier at cut in. An active rectifier may help, but probably not worth the trouble and expense. An inductor in the DC line would be an easy place to experiment. I notice more noise resonating from tail vanes on these machines (caused by the alternator) more than any audible tower issues, so thats where some attention could be spent maybe. 

RoyR
Have Fun!!!  RoyR KB2UHF

Mary B

  • Administrator
  • SuperHero Member
  • *****
  • Posts: 3169
Re: Once again in 3-part (phase) harmony
« Reply #10 on: December 11, 2014, 06:01:10 PM »
A trick ham radio operators use to stop antenna elements from vibrating in the wind and making a racket is to run a piece of nylon rope through them to damp the vibration. Might be worth a try to stick some in the tail boom. doesn't have to be thick, 1/2 inch would be fine

kitestrings

  • Hero Member
  • *****
  • Posts: 1376
Re: Once again in 3-part (phase) harmony
« Reply #11 on: December 12, 2014, 06:45:17 PM »
I'd been noticing a element of it that may well be highlighted by the tail boom.  Ours is heavy-walled AL tube, but if you ever compared aluminum snowshoes to wooden ones you get a sense of the sound difference.  So perhaps foam-fill, rope or some such would help.

ULR summaries a pretty compelling, collective sentiment here that the rectifier clipping is a primary contributor.  I don't want to overstate the problem, with ambient noise levels in any sort of wind it is not much from any sort of distance.  It is definitely more noticeable at cut-in as discussed.  Now since I started this thread I can report back one thing.  I programmed a new power curve for the thing as outlined here:
http://www.fieldlines.com/index.php/topic,148468.0.html

It seemed to have a positive effect, vis a vie boB’s suggestion.  The change (lightening the loading at lower rpm) seems to allow the thing to more smoothly transition through cut-in.  Still some noise above as before.  It may just be that it spends less time hovering above – then below cut-in in light winds, now letting it run a bit more if that makes sense.

Regarding the potential for a natural/resonant frequency, I think that is also a possible lesser contributor.  Years ago we had an engineer 'presenter' talking on the subject with respect to guyed structures and turbines.  He left me with this little illustrative tool: three small dowels.  One maybe 6" long with another 1" dowel on a free-spinning pin at one end.  On the underside of the longer piece are some evenly spaced kerfs, cut about a third of the way through.  You pinch the thing on the end of a table (cantilevered with the spinner hanging out).  With the third dowel you rub back and forth on the side, just touching into the kerfs.  At the right frequency the spinner spins and can be made to accelerate.  Change sides and it spins the opposite direction.  Fascinating to watch.

~ks

kitestrings

  • Hero Member
  • *****
  • Posts: 1376
Re: Once again in 3-part (phase) harmony
« Reply #12 on: December 05, 2016, 02:02:40 PM »
While this is an older post, it was very good discussion, and I've got some renewed interest in it now that our axial is back up.  We got a few low-cost/no cost strategies to attempt first to reduce noise at cut-in:

1) There are some fasteners and fittings to tighten (pal-nuts, wire wraps, etc.). It seems anything that can vibrate will.  Our rotor balance is much closer now and this has helped.

2) We have the benefit of flexing our power curve with the Classic CC, so it is conceivable that we could "play-thru" to a better spot as boB suggests.  Noise is most pronounced at cut-in.  In light winds it lingers at and just above this point.

3) I thought we might try some rope in the tail boom, or perhaps foaming the core (?)

To go further I'd like to understand better what type of inductor might work.  If it is on the DC side of the rectifier, I assume it like a "choke"; something to some this electrical/torque ripple, is that correct.  In our case we have two Classic's, each with their own isolated rectifier, so we could probably experiment on just the one that handles cut-in.  I assume it must be rated to handle the max current (50A), but I'm not sure how to spec this...

Something like this (?):
http://www.alliedelec.com/hammond-manufacturing-transformers-195e50/70230050/

~ks



« Last Edit: December 05, 2016, 02:42:39 PM by kitestrings »

Adriaan Kragten

  • Hero Member
  • *****
  • Posts: 1154
  • Country: nl
Re: Once again in 3-part (phase) harmony
« Reply #13 on: December 05, 2016, 03:54:30 PM »
If a 3-phase current with for which the voltage varies according to a pure sine wave is rectified in series, only two of the three phases are guiding a current. Therefore, the current in each coil starts and stops suddenly and this may cause vibration. The variation of the current in each phase is shown in figure 7 of my free report KD 340. A pure sine wave is created if a coil is rotating in a constant magnetic field.

If a rectangular magnet is drawn along a trapezium shaped coil, like it is done for most axial flux generators, the variation of the magnetic flux in the coil may differ a lot from a sine wave and so the voltage will also differ a lot from a sine wave. This may cause a much stronger vibration than for a pure sine wave.

For my VIRYA-generators derived from asynchronous motors, I have used magnet grooves which make a certain angle with the generator axis. These generators have laminated iron in the stator coils but the inclined magnet grooves prevent pulsation of the sticking torque. I have tested these generators for rotational speeds in between 0 and 1500 rpm but never heard any vibration. This type of generator is described in my free public report KD 341. Measurements of a generator with frame size 90 are described in my free public report KD 78.

joestue

  • Hero Member
  • *****
  • Posts: 1759
  • Country: 00
Re: Once again in 3-part (phase) harmony
« Reply #14 on: December 05, 2016, 07:52:04 PM »
Seems to me the ideal trapezoidal waveform should rectify to dc with 6 diodes, producing no ripple current and no torque ripple.


If you have a sine wave machine, the 16% ripple you get out of a 6 diode rectifier can be reduced to 4% if you use a secondary delta to wye transformer to feed a second rectifier, but that's expensive.

The 16% ripple from a 6 diode rectifier can be reduced to nearly zero percent current ripple with an arbitrarily large inductor, however you will still have torque ripple at the shaft, but electrically the current will be smoothed out and the initial clipping described by ULR should be mitigated.
My wife says I'm not just a different colored rubik's cube, i am a rubik's knot in a cage.

kitestrings

  • Hero Member
  • *****
  • Posts: 1376
Re: Once again in 3-part (phase) harmony
« Reply #15 on: December 06, 2016, 01:35:00 PM »
The turbine so far is working rather well, so I'm looking at strategies to improve the situation.  My thought was to try the lower costs things first; some I've mentioned.  Adding an inductor would seem pretty straight forward, but I was hoping for suggestions on where to start (size, type, mH, single- coil/choke/reactor?, configuration).  Am I correct that the suggestion is to add it in series between the rectifier and the battery bank on the DC output?

joestue, I'm not sure as I fully understand the delta to wye approach - and it sounds expensive - when you say "on it's own separate diode block".

For background, our alternator is a 16-pole dual axial PM design.  Magnets are rectangular (3" x 1.5") N42 NdFeB.  Coils are two in hand #14, 53 turns.  I put the resistance at .072 ohms/coil; .580 ohms L-L.

Thanks for any suggestions/clarification.  My education continues here...

joestue

  • Hero Member
  • *****
  • Posts: 1759
  • Country: 00
Re: Once again in 3-part (phase) harmony
« Reply #16 on: December 06, 2016, 05:22:29 PM »
Diy wind turbines are usually low frequency design, the inductor needs to be very large to do any good.
Which unfortunately means it comes at a high resistance and power loss.

What do you have for iron cores available?

A typical microwave oven transformer with a physical airgap of .5 to 1mm, will provide you with about 1 joule of magnetic energy storage for a constant power loss of about 6 watts. for 100 turns that's on the order of 30mH and current of about 8 amps. this will only cost you 1 whole kilogram of copper. these numbers don't scale linearly. to get 1.5Joules of energy storage from the same core requires you to waste 15 watts of heat in the copper.

Inductors scale the same as transformers, they get more efficient and higher power density with size. i have an inductor wound with something like 8 turns in my welder, the core is made from 4 microwave oven E cores fitted together. to wind the 4 cores separately would require nearly double the volume of copper for the same energy storage. (the I sections are discarded when two E cores face each other)

my inductor design spreadsheet is here (use sheet three for reclaimed transformer cores)
http://johansense.com/bulk/spreadsheets/

note that the air gap part of the math is not accurate, because it makes no regard for fringing flux. for low frequencies this doesn't really matter. you just adjust the air gap in real life to get the relationship between number of turns and inductance where you want it to be. so if the spreadsheet says X number of turns at Y inductance at Z energy stored at A flux density. then the air gap B in real life will be somewhat more than half of the number in the spreadsheet (for a transformer, there are two air gaps in series. for a gapped toroid there is only one air gap) 


anyhow, google 12 pulse rectifier. you use a separate transformer to shift the phase rotation 30 degrees. you have a choice. you can run the second rectifier in series or in parallel with the first rectifier. if in series it shares current better (but you get twice the voltage), and if in parallel you need inductors to share the current. the transformer needs to pass 57% of the total KVA.
« Last Edit: December 06, 2016, 05:40:11 PM by joestue »
My wife says I'm not just a different colored rubik's cube, i am a rubik's knot in a cage.

Mary B

  • Administrator
  • SuperHero Member
  • *****
  • Posts: 3169
Re: Once again in 3-part (phase) harmony
« Reply #17 on: December 06, 2016, 06:25:53 PM »
The rope would likely be better than foam. It vibrates up and down and can set up a canceling motion to the vibration.

You should stand under my 144mhz antenna array in high winds. The rattles are obnoxious!

While this is an older post, it was very good discussion, and I've got some renewed interest in it now that our axial is back up.  We got a few low-cost/no cost strategies to attempt first to reduce noise at cut-in:

1) There are some fasteners and fittings to tighten (pal-nuts, wire wraps, etc.). It seems anything that can vibrate will.  Our rotor balance is much closer now and this has helped.

2) We have the benefit of flexing our power curve with the Classic CC, so it is conceivable that we could "play-thru" to a better spot as boB suggests.  Noise is most pronounced at cut-in.  In light winds it lingers at and just above this point.

3) I thought we might try some rope in the tail boom, or perhaps foaming the core (?)

To go further I'd like to understand better what type of inductor might work.  If it is on the DC side of the rectifier, I assume it like a "choke"; something to some this electrical/torque ripple, is that correct.  In our case we have two Classic's, each with their own isolated rectifier, so we could probably experiment on just the one that handles cut-in.  I assume it must be rated to handle the max current (50A), but I'm not sure how to spec this...

Something like this (?):
http://www.alliedelec.com/hammond-manufacturing-transformers-195e50/70230050/

~ks

kitestrings

  • Hero Member
  • *****
  • Posts: 1376
Re: Once again in 3-part (phase) harmony
« Reply #18 on: December 07, 2016, 09:50:09 AM »
I'm confused by these two statements which seem to be odds with one another.  We are definitely low frequency - part of the appeal of the axial's IMO is they are inherently direct-drive, low speed - I believe we are at ~13-18 Hz where this occurs.  Is it a question of the magnitude of inductance added, or am I missing something more?

Quote
some of you here may see a marginal efficiency improvement placing an inductor after the bridge rectifier

Quote
Diy wind turbines are usually low frequency design, the inductor needs to be very large to do any good.
Which unfortunately means it comes at a high resistance and power loss.

Are there stock units that would test the theory? Is it possible to get a variable inductor?

Mary B, I think rope in the tailboom would be relatively easy to try, though more vibration seems to transmit down the tower legs.  I've got a couple points where neoprene/rubber isolation might help.  I'm wondering what type of rope.  Hemp might be best, but I'd prefer polyprop. or dyneema; something that wouldn't absorb water.  Is it left slack or need to be under tension?

Thanks, ~ks

joestue

  • Hero Member
  • *****
  • Posts: 1759
  • Country: 00
Re: Once again in 3-part (phase) harmony
« Reply #19 on: December 07, 2016, 07:05:57 PM »
so what i was talking about there is the poor power factor caused by non sinusoidal waveforms under very low load. once the current becomes continuous then the inductor's effect on efficiency may be negligible, particularly so with low efficiency, high impedance, long transmission line wind turbines, where the transmission line itself may be enough inductance once the frequency climbs, and the resistance of the generator is so high that the initial poor power factor at initial cut in is of no concern.


anyhow it doesn't matter if the inductor saturates under heavy load, so you don't need a variable inductor.. all that matters is that it doesn't burn out. so you can use an ac current transformer on one or more phases, rectify it and drive a relay to short out the inductor when the current starts to climb.


as far as variable inductor, you can change the air gap, that's easy, or just wind it with say, 12 wires in hand. the inductance could be 1,4,9,16,36,or 144mH if each coil is 1mH. (without losing any copper and thus increasing losses, just reconnect it for a multiple of 1,2,3,4,6 or 12 turns)

My wife says I'm not just a different colored rubik's cube, i am a rubik's knot in a cage.

Mary B

  • Administrator
  • SuperHero Member
  • *****
  • Posts: 3169
Re: Once again in 3-part (phase) harmony
« Reply #20 on: December 07, 2016, 08:54:11 PM »
Any good grade UV resistant poly rope is any will be exposed, otherwise any good poly rope. The light weight lets it move to cancel vibration. I have it in the elements of my 6 meter antenna and without the entire tower gave off a low hum...

Adriaan Kragten

  • Hero Member
  • *****
  • Posts: 1154
  • Country: nl
Re: Once again in 3-part (phase) harmony
« Reply #21 on: December 10, 2016, 03:33:35 AM »
It might be that the position of the rectifier also has an influence on the torque vibration in the generator. There are three options for this position. 1) directly after the generator. 2) At the tower foot. 3) Near the batteries.
Placing directly after the generator has as disadvantage that the rotor can only be stopped by making schort-circuit in the DC-line (if you don't climb in the tower) and the short-circuit current then flows through the rectifier which may ruin the diodes. The short-circuit is also not as strong as for short-circuit before the rectifier because you have a voltage drop of about 1.4 V over the rectifier diodes.
If the rectifier is placed at the tower foot you still have the opportunity to make short-circuit in the AC current. I prefere this option for my bigger windmills.
If the rectifier is placed near the batteries you have long lines in which the current is flowing only during 2/3 of the time. This starting and stopping of the current in that long lines may cause extra torque vibration in the generator. The heat dissipation for an AC current in three lines with a certain copper area is the same as in two lines with the same copper area for DC current, so you also safe a lot of copper if the rectifier is positioned at the tower foot.

Ungrounded Lightning Rod

  • SuperHero Member
  • ******
  • Posts: 2865
Re: Once again in 3-part (phase) harmony
« Reply #22 on: December 10, 2016, 10:44:46 PM »
NASA had a hack for killing vibrations.  It was an adhesive tape with rubber in the middle and lead foil on the side opposite the adhesive.  Stick a piece on some part (like a piece of sheet metal) that had a nasty resonance, near a "loop" (most moving spot) of the standing wave if possible.  The rubber absorbs the energy while the mass of the lead provides an out-of-phase "skyhook" to apply the forces against.

If your tower has a bad resonance at some frequency (like one of the near-cutin harmonics) you could do something similar by finding a place where the vibration is large (listen to the tower and find a spot where it's loud) and mounting something like an energy-absorbing rubber part from an automotive front end at that point, with a weight on the other side of it to provide the skyhook mass, so the rubber gets squeezed and stretched by the vibrations, absorbing the energy.  (Don't forget to shade it from the sun to keep it from rotting.)

= = =

The adjust-the-power-curve hack lets the mill speed up beyond cutin when you start to load it, so you get past the "short clipppy spikes" stage before you start to pull power.  The controller amounts to a switching regulator so you get to load it as you want.  But I bet this will also raise the effective cutin and cost you power when the wind is low - when there's little to be had but you need it the most.

If you're going all the way to switching regulators, I can imagine a purpose-built buck-boost converter that would track the waveform and apply a constant or near-constant (versus shaft position) torque to the shaft, eliminating the torque-variation vibration source entirely.

Edit:  Now that I think about it, since torque is proportional to current in a PM alternator, the regulator wouldn't be all that complex:
 - Determine the current rotational rate.  (Capacitor, diodes, low-pass filter = tachometer circuit.)
 - Determine a desired torque for that rotational rate (a function of the low-passed output voltage).
 - Make the buck-boost try to regulate it its INPUT CURRENT to match that function.
bingo:  You get vibration free operation, max-power-point loading, and anti-runaway / anti-burnout functionality, all from the same gadget, by appropriately designing the desired-current function.

It it's a far cry from the "cheap and simple" approach of alternator, rectifier, battery, dump load with
controller, and gives you an extra, complicated, piece of electronics to fail.  On the other hand it could replace the "no power below cutin" effect with a gradual loading, starting with providing a trickle of power at very low wind speeds and smoothly tracking the max power point.
« Last Edit: December 10, 2016, 10:59:46 PM by Ungrounded Lightning Rod »

SparWeb

  • Global Moderator
  • Super Hero Member Plus
  • *****
  • Posts: 5452
  • Country: ca
    • Wind Turbine Project Field Notes
Re: Once again in 3-part (phase) harmony
« Reply #23 on: December 11, 2016, 04:40:25 PM »
KS,
Lots of good ideas (and some far-out ideas!) but I can't tell if they are appropriate to your situation.  Two more if you want a different approach:

Have you any interest in posting a waveform from your WT at the noisy speed?  This sounds complicated and expensive, but a pocket o'scope is less than a hundred bucks and handy for other things once you get your feet wet in electronics.  With a plot of the waveform, you can figure out where it differs from a pure sine wave.  These differences are "harmonic distortions" which are the REAL targets to stamp out with devices like inductors on the line.  Some harmonics can never be completely eliminated because both PWM and MPPT control schemes are based on switching, not linear changes.  You have to know which harmonics are causing the noise, however, before coming up with a strategy to damp them.

When it comes to mechanical harmonics, the principle is surprisingly similar.  If you can measure the intensity of the vibration in the tower and its frequency, you can focus your attention on the most likely sources of noise.  This can be as simple as downloading an app to your smartphone and then taping it to a tower leg for a few minutes on a windy day.  Armed with this information, you can tell whether the noise comes from the wind through the tower legs, or the turbine itself.
No one believes the theory except the one who developed it. Everyone believes the experiment except the one who ran it.
System spec: 135w BP multicrystalline panels, Xantrex C40, DIY 10ft (3m) diameter wind turbine, Tri-Star TS60, 800AH x 24V AGM Battery, Xantrex SW4024
www.sparweb.ca

kitestrings

  • Hero Member
  • *****
  • Posts: 1376
Re: Once again in 3-part (phase) harmony
« Reply #24 on: December 15, 2016, 11:26:24 AM »
Thanks for all the suggestions.  Much to consider here.

As always, I shouldn't rule out the little things, as cumulatively they have a way of adding up.  Since I posted I've tightened all of the tower connections.  In particular, as we raised the new tower we didn't take the time to torques the leg bolts - there's a set of six at each 10' section - in the interest of time; we we're up against winter weather near the end of this project.  This effort made a substantial improvement.  There is also some guy noise stemming from vibration of the (caution yellow) guy guards.

I've also changed the power curve to load the alternator earlier (cut-in at 58V/90 rpm), but with light current levels from cut-in through modest level winds.  This seems to work better.  It may be less productive at the  lowest winds, but it seems to avoid a harsher transition from unloaded to loaded conditions.

Quote
If the rectifier is placed near the batteries you have long lines in which the current is flowing only during 2/3 of the time. This starting and stopping of the current in that long lines may cause extra torque vibration in the generator.

Yes, this is what we have by choice.  I recall years ago working on an early Dunlite/Quirks turbine that had to be taken down in order to access the diodes - a 600# machine.  They later moved them to a mounting ring at the back of the generator housing, but you still had be on the tower for the repair.

In our case the rectifier and all electronic components are in the house near the batteries.  IIRC we have about 185' of 2/0# AL underground to the tower which terminates up about 20' up the tower.  From there it another 65' of #6 CU SO cord.  In the house there is a lightning termination panel, which as I understand it has spiral wound inductors with MOV's on each phase lead.  The 2/0# transitions to #6 CU for about 10-12' to the rectifiers.

I was hoping there might be a stock inductor or inductor/capacitor filter that we might try.  It's been a challenge building a turbine.  I'm not sure I want to design & build an inductor here, as my first attempt.

Spar, I have messed with sound & vibration recording from an iphone app.  I'll give that another look and try to report back.  I probably can access a o-scope from work.

Thanks again, and happy holidays, ~ks

joestue

  • Hero Member
  • *****
  • Posts: 1759
  • Country: 00
Re: Once again in 3-part (phase) harmony
« Reply #25 on: December 16, 2016, 03:19:39 AM »
Thanks for all the suggestions.  Much to consider here.

As always, I shouldn't rule out the little things, as cumulatively they have a way of adding up.  Since I posted I've tightened all of the tower connections.  In particular, as we raised the new tower we didn't take the time to torques the leg bolts - there's a set of six at each 10' section - in the interest of time; we we're up against winter weather near the end of this project.  This effort made a substantial improvement.  [.....]
I was hoping there might be a stock inductor or inductor/capacitor filter that we might try.  It's been a challenge building a turbine.  I'm not sure I want to design & build an inductor here, as my first attempt.

inductors are trivial for low frequency.

take a steel core on the order of 1 to 10 kilograms. measure the magnetic path length. make the air gap 1/100'th of that distance. calculate the number of turns such that the current you need is on the order of reasonable power loss. maybe 10-100 watts of copper loss per kilo gram of copper.

but that's not what i'm wondering about.

if torquing the bolts fixed certain vibration problems, then those bolts weren't doing a damn thing.....
My wife says I'm not just a different colored rubik's cube, i am a rubik's knot in a cage.

kitestrings

  • Hero Member
  • *****
  • Posts: 1376
Re: Once again in 3-part (phase) harmony
« Reply #26 on: December 16, 2016, 09:32:05 AM »
Quote
if torquing the bolts fixed certain vibration problems, then those bolts weren't doing a damn thing.....

That's mostly true joe.  The tower leg bolts are in shear between the overlapping leg sections, and the guys keep near constant downward pressure on them.  Oncd in place there's little damage of them falling out.  We had intentionally left them loose - with nuts started on hand-tight - in the interest of assembly time.  Rohn uses deformed thread galvanized locking nuts where they used to have regular hex-nuts with pal-nut retainers, so they are a bit time consuming to run in (all while the ground-crew is waiting on you).

I'm not suggesting that they were the source of vibration, but lose, they vibrate and also allow the over-sized leg bells to vibrate against the next lower section.  The worst offender was actually the first section where it bolted to the base-plate (which in turn rests on the concrete center pad/footing).  Torquing the bolts also slightly stiffens the sections between guy points.

Quote
inductors are trivial for low frequency.

? Now I'm confused, because I thought this was one of the suggested strategies early on here...

Isn't it fairly common on power supplies for there to be a filtering sequence after the rectifier to smooth out the unwanted AC components of the waveform?

~ks

joestue

  • Hero Member
  • *****
  • Posts: 1759
  • Country: 00
Re: Once again in 3-part (phase) harmony
« Reply #27 on: December 16, 2016, 03:31:56 PM »
I meant they are extremely easy to design and build.
My wife says I'm not just a different colored rubik's cube, i am a rubik's knot in a cage.

kitestrings

  • Hero Member
  • *****
  • Posts: 1376
Re: Once again in 3-part (phase) harmony
« Reply #28 on: December 16, 2016, 08:08:14 PM »
Ahhh...I got it.  Too funny.  I was reading inconsequential, ineffective (at low frequencies).  Hard to read expression, intonation on these darned things....

Jason Wilkinson

  • Full Member
  • ***
  • Posts: 181
  • Country: bb
Re: Once again in 3-part (phase) harmony
« Reply #29 on: December 17, 2016, 07:33:54 PM »
 After some of magnets corroded and lifted off the  fiber  glass potting , i remade the rotors with out the potting . Same 18''  rotors 16 mag/ 12 coils  and now there is this deep growl, wasn't there when the rotors were potted.  I checked and the growl starts at about 10 amps   and disappear as the amps gets higher.  Surprisingly, when the smaller turbine  10 footer  12/9  and not potted was on the same tower, that growl /hum was only heard if you placed your ear on the tower. When your tower is only 10 to 12 feet from the house  the growl can be annoying especially to those who don't know what it is.
 Jason

SparWeb

  • Global Moderator
  • Super Hero Member Plus
  • *****
  • Posts: 5452
  • Country: ca
    • Wind Turbine Project Field Notes
Re: Once again in 3-part (phase) harmony
« Reply #30 on: December 18, 2016, 11:20:10 AM »
Hi KS,
Some reading material on the subject:
http://literature.rockwellautomation.com/idc/groups/literature/documents/wp/mvb-wp011_-en-p.pdf

The white paper isn't too technical so even I can follow it.  The current waveform at the beginning is the kind of thing our generators experience.  The reasons are different - the paper is written for VFD drives, after all - but the effect is the same.

In my case, I have a "parallel resonance" situation with my motor-conversion. That's why a bank of capacitors proved to be so effective in my system when I tried it.

In your case, you could have the "series resonance" but not necessarily.  The capacitance due to long AC lines from tower to rectifier could be the culprit.  There isn't a lot of inductance in Axial-flux genny windings (compared to an iron-core motor/generator, that is) making an inductor installed close to the rectifier effective at getting out of the resonance zone.  In series, however, the two will increase the total impedance so don't go overboard or you will reduce the current that can flow.

Do you still have the low-speed noise that you heard earlier?  Or has the tower-tightening solved the most of it?

Jason,
Sorry to hear you have similar trouble.  Are you considering similar solutions?
No one believes the theory except the one who developed it. Everyone believes the experiment except the one who ran it.
System spec: 135w BP multicrystalline panels, Xantrex C40, DIY 10ft (3m) diameter wind turbine, Tri-Star TS60, 800AH x 24V AGM Battery, Xantrex SW4024
www.sparweb.ca

kitestrings

  • Hero Member
  • *****
  • Posts: 1376
Re: Once again in 3-part (phase) harmony
« Reply #31 on: September 29, 2017, 11:36:08 AM »
I've been tracking this and taking some vibration recordings and output screenshots for a while.  I think I have it narrowed down a bit more.  It seems where this is most pronounced is actually not right at, but slightly above cut-in, likely at a resonant frequency that is exasperated by the tower and or turbine itself.  The rough point is when the turbine is putting out ~300 watts; the DC voltage is in the low to mid 70V range, and I'm estimating that the AC frequency is 15 Hz.

I'd like to experiment with an inductor.  After researching inductive filtering for rectifiers a bit, I found this formula:

RF= 0.4714/√[1+(4πfiL/R)2]

ref:
http://www.irjes.com/Papers/vol2-issue6/Version-1/E02064249.pdf

where fi = input frequency, L = inductor value (mH), and R = resistance (ohms)
I thought I might try an off-the-shelf unit in series with the first controller, which is current limited at 40A.  If my math is correct, a 50A, 2.5 mH choke with a resistance of .008 should result in a ripple factor of less than 1%.

What am I missing?  Is this worth a try?  How do I calculate the loss that it introduces?
I think there may be some other mitigation measures that could be employed – rope in tube, vibration damping tape on stator bracket, tower damping – but if I can reduce the source of the problem, that'd seem logical to me first.

SparWeb

  • Global Moderator
  • Super Hero Member Plus
  • *****
  • Posts: 5452
  • Country: ca
    • Wind Turbine Project Field Notes
Re: Once again in 3-part (phase) harmony
« Reply #32 on: September 29, 2017, 02:51:44 PM »
The article you found barely mentions impedance.
That would be the reactive resistance, out of phase with the voltage, but still there, working against the current. 
Read the author's Table 3.  Look at the column Irms, which goes from 0.6A to 0.15A as he selects values that correct the ripple.
The author's goal was to reduce ripple in the output current, not to optimize power transfer from the source to the load.

I'd try some of your mechanical ideas first.

Also, if you want to take a peek at my old postings (ca 2012?), you can see the vibration isolators in my motor-conversion mount which wouldn't be too hard to adapt to a pancake alternator.
No one believes the theory except the one who developed it. Everyone believes the experiment except the one who ran it.
System spec: 135w BP multicrystalline panels, Xantrex C40, DIY 10ft (3m) diameter wind turbine, Tri-Star TS60, 800AH x 24V AGM Battery, Xantrex SW4024
www.sparweb.ca