Planning a new turbine, Toshiba 7.5 kW Conversion

[SparWeb]

This one's been a long time coming.  The generator was built many years ago, and it has patiently waited to see the light of day ever since.



I really should write better notes about the construction process, but here they are:
http://www.sparweb.ca/3_Gen_MoCo/Toshi.html

Next are the blades.  I have a rough size in mind, and I would like to carve laminated cedar blades again, like I did for the Baldor conversion.



For this blade design, I'm using Onshape.com. I've really embraced this CAD system for a lot of projects.  Onshape is new, and it's not as rich with features as Inventor or Solidworks, but it's growing fast and it completely suits my needs as a hobbyist.  There are already a number of things I can do with Onshape that I struggle with in Solidworks or just take longer.

This rotor will have a diameter between 14 to 16 feet.  I'm still working it out exactly.
Thankfully, I did some run-up tests of the generator's power, and took a lot of notes.



There's something to be said for INefficiency.  The power curve of this thing is strange.  You can see how the output power flattens out, no matter how I connect it.  Even Delta flattens out, just at a higher level.  There could be a benefit to this if it lets me use bigger blades.  The cause is still not clear to me.  Flux and Zubbly understood it, but I never quite grasped... If I can figure it out, then I can show that the rotor can take lots of power but none of the power electronics on the ground melt down, because the current is limited.  If that's true, I think I can over-size the rotor, get a low-wind advantage, worry less about high winds... sweet spot.
Still sorting out if it's safe to use it this way.

I also haven't figured out what the best connection scheme on the ground would be.  My battery stack could take the juice with the DC hooked up directly, but I'd like to experiment with MPPT (finally; I'm like the last luddite hold-out).

[Adriaan Kragten]

Any PM-generator has a certain maximum torque level. The maximum torque level is about the same for any voltage but is a lot higher for rectification in delta than for rectification in star because higher harmonic currents can circulate in the winding for rectification in delta. Measured Q-n curves for different DC voltages, for star and delta rectification are given in figure 8 and figure 13 of my public report KD 78 which can be copied for free from my website: www.kdwindturbines.nl at the menu KD-reports. It can be seen that the torque Q is even decreasing at high rotational speeds.

As the mechanical power Pmech is proportional to the product of the torque Q times the rotational speed n (Pmech = Q * pi * n / 30),  Pmech will increase at increasing rotational speed even if the torque is about constant or slightly decreasing. The electrical power Pel depends on the generator efficiency which decreases at increasing current. The result is that Pel is also increasing at increasing rotational speed but not as strongly as Pmech (compare figure 9 of KD 78 for star rectification and figure 14 for delta rectification).

The Pel-n curves of your generator for different constant voltages are about in agreement with the effects which I have measured for my radial flux PM-generator as used in KD 78.

[joestue]

Any PM-generator has a certain maximum torque level. The maximum torque level is about the same for any voltage but is a lot higher for rectification in delta than for rectification in star because higher harmonic currents can circulate in the winding for rectification in delta. Measured Q-n curves for different DC voltages, for star and delta rectification are given in figure 8 and figure 13 of my public report KD 78 which can be copied for free from my website: www.kdwindturbines.nl at the menu KD-reports. It can be seen that the torque Q is even decreasing at high rotational speeds.

it took me a while to understand it but the maximum torque drops off as the rpm rises because the frequency is rising and the leakage inductance isn't changing, so as the rpm rises the leakage inductance becomes proportionally much greater at for example, 200hz than it is at 20.

regarding the higher torque in delta.. yes, i think i know what you meant to say but the improvement in performance as a generator is because the re-circulating current due to harmonics disappears as the generator's output voltage is clamped.. by a "dumb" 6 diode rectifier and 24 or 48 volt battery.

where as a star connection, the current though each winding is conserved and even though the generators voltage is clamped at much lower than open circuit voltage, the star connection is still removing the 3rd harmonic. (where as in delta it would circulate)

but the delta connection is free to push that third harmoinic into the rectifier if the rectifier clamps the generator volts lower than the difference.


so for example the 5kw singlephase  generator head i rewound, develops something like 140 volts line to neutral but only 230 volts line to line. the ~15 missing volts in wye means it should consume 6% more torque connected as a generator in delta.

so in the case of my generator, provided the generator is loaded by a "dumb" rectifier such that the terminal volts are less than 132 vac, then there is no circulating current in the delta connected winding. now, for a 5kw motor, full load is far more than a 6% voltage drop, or a wind turbine where efficiency might not be 60% you will see the benefit of a delta winding immediately.


obviously, my estimate of 6% is subject to a full study of the other harmonics and their proper addition and subtraction. they don't add directly but rather a bunch of math i never learned is involved.


it would not surprise me if Spar web re-runs the test with a 3phase resistor load and find less of a difference between star and delta.

[mbouwer]

Is it an option for you to apply blade pitching?

[SparWeb]

Hello Mbouwer.  About blade pitch:
I would, if I had access to the tools to make a good hub with pitch control.
I only have simple tools available to me now.  I would not trust it to be robust if I couldn't use industrial tools that maintain tight tolerances.
There was a time that I had a job where I could use lathes and milling machines on the weekends, but those days are over!
I want a machine that I can trust to operate reliably for a decade (like the one I currently have).
I would also demand the ability to make adjustments so that the speed at which the blades feather can be fine-tuned.
If I wanted a hub with pitching (feathering) blades, I would choose to make the bearings with a lathe and the pitch control with rod-end links. 
You can buy some of that hardware, and some of it has to be machined to make it all fit together.

These are the things that I see you working on on your turbine - I would want no less for mine.

[SparWeb]

Adriaan,
I'm continuing to work out the matching size of rotor using the test data I collected.  I measured mechanical power input during the tests, simultaneously with the output electric power.  This is proving to be a good resource of information, and making the match is so much easier when the generator power in & out has been measured directly.

I'm also working out if the 3rd harmonic in Delta can safely be shed as heat (since the strong wind that produced the output is also a powerful cooling effect on the motor housing).  I am weighing this against the trouble I once had with rectifiers breaking down while my turbine was connected in Delta.  I believe my system is much more robust now, so I shouldn't worry as much about switching at the rectifiers driving that harmonic.

Joe,
I fully agree with your last statement, about running the test with resistive loads.  I expect the behaviour in Delta would be much less extreme, and furthermore I expect the efficiency would be higher, too.

[midwoud1]

    Hi Sparweb .

  Pitch control with (45 deg. gear   on the shelf )  https://www.youtube.com/watch?v=MRVGYVrgsdc

 

[Adriaan Kragten]

Using the generator rectified in star has another advantage than the higher efficiency and that is that the unloaded torque is rising less fast at increasing rotational speed (see figure 2 KD 78). So this facilitates the starting behaviour of the rotor at low wind speeds as more torque is available to accelerate the rotor. But the higher maximum torque level in delta is favourable if the generator is used as a brake by making short-circuit in the winding. To prevent damage of the rectifier diodes and to prevent the voltage drop over the rectifier, this should be done before the rectifier. So the retifier and the short-circuit switch have to be placed at the foot of the tower. However, if the generator is normally used in star, an extra cable is required from the star point to the short-circuit switch.  Short-circuit in star including short-circuit of the star point is the same as short-circuit in delta as all six ends of the three phase coils are connected together.

[kitestrings]

So the retifier [sic] and the short-circuit switch have to be placed at the foot of the tower.

Or, extend the 3-phase power wires into the house/shed and have the rectifier there, correct?  There is no other reason that they have to be at the tower unless I've missed something.

I've seen turbines with rectifiers in the machine, and have never liked this arrangement for maintenance purposes.  The base of the tower is an improvement over that, but everything still has to be protected from the elements.  We put our rectifier in the house which has some other benefits.

How much more braking torque is gained by including the star point vs. not in a wye configuration?

[SparWeb]

I don't consider it necessary to place the rectifier at the base of the tower.  Or any specific location, for that matter.  It can be in the generator, on the tower, or at the end of the run by the batteries.
Overall performance of the system in response to electrical load can be made equal regardless of the rectifier location.  Obviously unfair comparisons can be made, eg, long small-diameter DC wires could be used to "show" that DC is bad in long runs, or attempting to "prove" the opposite by biasing the system design the other way.  A careful examination of the system design would quickly reveal that there any belief that long DC runs are good or bad just shows that someone has made a mistake somewhere else.  The best system design will locate the rectifier at the most economical position for acceptable system performance, and since its location is a very weak determinant of the performance, ultimately anything goes.

Location of the rectifiers is truly a non-issue, and it will be overridden by a dozen more important issues, such as optimizing the cost of the buried cable runs, and how well this will work when tied to a MPPT charge control system.

I do support putting the electric shorting switch as close to the generator as possible.  The base of the tower seems to be the closest spot practical.  That switch doesn't have to be anywhere near the rectifiers, just upstream of them.

[bigrockcandymountain]

Agreed that the rectifier can go anywhere.  When i bought underground wire i was told 2awg aluminum was the cheapest because it is very common.  I have 400' 3 conductors and it cost about $400 cad it is direct buried.  Now with mppt i could go smaller, but the losses are mininum this way.

Your turbine should be higher output than mine so size accordingly.   I dont have a stop switch.  If i did it would be at the tower base. 

I am excited to see the tower plan.  60' free standing is no small feat of engineering. 

I would go 4 post with a similar taper to old water pumpers.  I think about 12' square at the base.  Maybe 16'  My gut tells me the 4 main posts would be 3" angle 3/16 thick.  Could be pretty easy and cheap to weld up your own tower.  My dad built one 30 years ago 30' high and 8' diameter fan.  It has held perfectly.  You have real engineering knowledge, so you'll optimize way better than i could.  I am just rambling throwing out ideas. 

[joestue]

How much more braking torque is gained by including the star point vs. not in a wye configuration?

about as much as the difference i mentioned.

you can measure the difference yourself by applying a constant torque load (such as a rope wrapped around a shaft with a weight) short circuit the generator and measure the rpm.  then short circuit the star point as well and it might be say, 10% slower.

[Adriaan Kragten]

There are two reasons why the tower foot is the optimum place for the rectifier and the short-circuit switch.
 
1) The cable losses for DC current are much smaller than for a star rectified AC current if the same amount of copper is used for the cables. This is because the AC current is flowing during only 2/3 of the time (see figure 7 public report KD 340). For the same cable losses one therefore needs three cables of a certain size for AC current and two cables of the same size for DC current. So the cable costs for AC current are 50 % higher than for DC current for the same cable losses. Therefore the rectifier has to be positioned as close as possible to the generator. This fact is especially important for big wind turbines positioned at a large distance from the batteries.

2) The short-circuit switch must be positioned in the AC part of the cable circuit to protect the rectifier diodes and to prevent the voltage drop over the rectifier. But for stopping of the rotor by making short-circuit you don't want to climb in the tower as stopping of the rotor may be needed at high wind speeds. So the short-circuit switch must be positioned at ground level and so at the tower foot. As the rectifier is positioned after the short-circuit switch, it must be positioned at the tower foot too.

For some of my small VIRYA wind turbines, the short-circuit switch and the rectifier were mounted in the box which contained the voltage controller and the dump load because it is easy that all electronic components are in one box. This box is positioned near the batteries. So I accepted that a more expensive 3-wire cable was needed and that there is some voltage drop over this cable when short-circuit is made. This voltage drop makes that the rotational speed where the braking torque is maximal is somewhat higher than for short-circuit directly at the generator terminal.

The difference in maximum braking torque for short-circuit in delta and in star can be read in figure 4 of KD 78 by comparing both Q-n curves. The peak torque for short-circuit in star is about 30.2 Nm. The peak tor for short-circuit in delta is about 37.8 Nm. So the peak torque for short-circuit in delta is about a factor 1.25 larger than for short-circuit in star for the measured PM-generator. The ratio may be different for other PM-generators.

[SparWeb]

Agreed that the rectifier can go anywhere.  When i bought underground wire i was told 2awg aluminum was the cheapest because it is very common.  I have 400' 3 conductors and it cost about $400 cad it is direct buried. 

That's much like what I want.  I had to bury a new line from my transformer pole to the house about 10 years ago, when the existing buried line "broke" for some reason.  Underground I didn't dig it up to find the fault, I just buried a new line.  Expected to find the old line while I was digging but never struck it.  They must have gone the long way around the house...  Anyway, I paid about 300 dollars for armored "Teck" cable at 6 gauge (or maybe it was 4Gga) and a day's rental for the ditch digger.  This is the same scale of job with heavier cable but probably no need for armoring, so that's a job I know I can take care of.

I would go 4 post with a similar taper to old water pumpers.  I think about 12' square at the base.  Maybe 16'  My gut tells me the 4 main posts would be 3" angle 3/16 thick.  Could be pretty easy and cheap to weld up your own tower.  My dad built one 30 years ago 30' high and 8' diameter fan. 

That's impressive.  I am not really thinking of such a wide base.  Certainly larger than your typical TV tower, though.  If you peek at my other posting about this you can see the kind of self-supporting tower I already have for my internet antenna, which I want to scale up to suit a big turbine.

I've worked my way through a lot of stress calc's and this seems to be doable because the stresses aren't really "scary" yet.  Your typical hot-rolled steel channel and angle stock is good for about 50, maybe 60 ksi (400 MPa for those reading in metric).  I can get away with a fairly slender tower base (4 feet) provided that I use a lot of cross-members.  This proves to be effective at preventing buckling problems.  It is the condition of "column buckling" that limits the strength of these truss towers.  I don't seem to need steel any stronger than HRS because the buckling strength depends on geometry (thickness of the cross-section) not strength.

In other words, when the wind pushes on the tower sideways, and one leg gets stressed in tension to 40 ksi, then the opposite leg will be in compression to 40 ksi.  It doesn't matter if the tensile strength of steel is stronger than 40 ksi if the column collapse of the compression leg would happen at 30 ksi.  These are the stress levels that I'm finding in my analysis so far, and what it's telling me is that adding plenty of cross-braces raises the column strength to match the tension strength.  This makes it resist the column collapse condition with a nice margin of safety.

One of the way's I've been able to self-teach the analysis is to look carefully at some other towers.  I looked at a Trylon tower a few years ago, took measurements, and kept the datasheets for it so that I could do a sort of reverse-engineering analysis of it as an example.  It's been very helpful, now that I've really put in the effort to work it out.  FYI, Trylon is a Canadian company with an office in Calgary not far from where I work.  https://www.trylon.com/products/index.aspx  Doing the calculations on one of their towers has given me the surprising result that I could just buy one of their towers and it would work just fine.  It might not cost much more than building it myself.

[Adriaan Kragten]

To my opinion it is not a good idea to use a wide base tower which is normally used for slow running water pumping windmills for a fast running electricity generating wind turbine. This is because a wide base tower is very stiff and therefore has a high natural frequency. The maximum rotational speed in rev/s of the rotor of a slow running water pumping windmill is always lower than the first harmonic natural frequency of such a tower and therefore the tower will not be pushed to vibrate in its natural frequency. But the tip speed ratio of a fast running rotor for generation of electricity is about a factor five higher and therefore at a certain rather high wind speed, the rotational speed will become the same as the natural frequency of the tower and this will give strong oscillations of the tower if the rotor has even the smallest imbalance.

A slender free standing tower has a much lower natural frequency and now the rotational speed will become the same as the natural frequency for low wind speeds for which there is only little energy in the wind and so the oscillation of the tower is not a big problem. For moderate and high wind speeds, the rotational speed of the rotor will be much higher than the natural frequency of the tower and then resonance will not occur.

[bigrockcandymountain]

Adriaan,
That is very good to know.  I know my 13' turbine on a guyed tower has resonance issues at low rpm.  It goes away just after cut in.  Probably about 200 rpm.
Spar,
Interesting that you use 50ksi for hot rolled steel.  I always use 30ksi but don't really know how to do proper load analysis, so it becomes very approximate anyway. 
What does everyone use for a thrust load on the disc?
 I figure worst case, furling stuck full facing the wind at 30psf.  I am curious to know what others think.

[SparWeb]

BRCM,
You are probably using 30 ksi for yield while I was using 50 ksi for failure.  I was writing with numbers off the top of my head yesterday.  Looking at my actual calculations right now... yeah I use Yield as the limit, which is indeed 32 ksi.  So I didn't need to refer to 50 ksi earlier.

[kitestrings]

I would agree that the location of the rectifier is mostly a matter of choice, and that there are myriad of other design considerations that are likely more important.  I do think a portion of my comment may have been missed.  Years ago I worked on a few early Dulite/Quirks 2 kW generators where they had the rectifier built into a ring mount in the back of the generator.  A lightning strike can easily pop one of them.  In this case you had to take down a roughly 600# machine to get to, and replace a ~$2 diode.  Basically anything that can fail in time will; the less of those things on the tower, or out in the elements for that matter, the better.

You have to properly account for losses in your selection of wire, and the braking, if employed, needs to go on the AC side of the rectifier.  I'd agree with those two thoughts, but that hardly justifies the base of the tower as "the optimum location" IMO.

Thanks Joe.  I'd considered bringing down the common star point for braking, but for us it is mostly a parking brake for service.

[SparWeb]

I just learned something interesting about vibration.  Actually, about damping, where towers and trusses are concerned.
It turns out that two similar structures can have different vibration damping simply from how they're assembled.
If I were to rivet a tower together, or if I welded the same structure together, I would get twice as much damping from the riveted structure.
The rivets permit some rotation at each joint, so vibration is not as efficiently transmitted from one member to the next, making the riveted tower less prone to carrying as much vibration than if it were all welded together.

It is also fun to see that a stronger tower is a stiffer tower, which raises the natural frequency, which makes resonance more probable.
Go for soggy wet-noodle trusses held together with pop-rivets. 

Working this stuff out really does reveal, every time, that there ain't no such thing as a free lunch (TANSTAAFL) and every design is a combination of trade-offs.

[Mary B]

If I was doing a 4 legged Windcharger type tower I would bolt it together with stainless hardware and nylock nuts... I have seen rivets pop under high wind conditions on Rohn BX towers with just a small TV antenna on them!

[joestue]

If a bolted or riveted joint is absorbing vibration there is a good chance it will fail, or the bolts will loosen up, etc.

Remember, the vibration dampening is because of friction between two moving surfaces clamped together...

[mbouwer]

That is why I like to make models and test setups.
As this tilt mast.

[kitestrings]

Nice tower mbwr.

When I worked on small wind systems years ago, we had a vibration consultant help with taming some turbine/tower vibration and noise.  In some case if the natural frequency and the normal operation of the turbine overlap you can get a 'whole lot of shakin' going on'; no surprise to most.  I was surprised that often the fix was not necessarily a heavier, or tighter guy wire.  At times it was simply moving a set up or down a few feet; in others it might be a slightly smaller diameter cable.

I had this prop from a training - still in a drawer somewhere - it was a notched wooden dowel with a pin on one end, and a smaller spinner of the same material.  If you held it over the edge of the desk and rubbed a second stick up and down the side of notches it spun in one direction, and the speed accelerated if you matched the natural frequency - like sending pulses thru the material.  If you moved it to the opposite side, you could cause it reverse directions.

~ks

[SparWeb]

Here's something I did for my current Zubbly machine, and I intend to re-do for this Toshiba.
Where the mounting "feet" of the motor housing are attached to the tower head, I have vibration isolators that have been very effective.

The tower head supports the motor on a frame, and the holes through the frame for the mounting bolts are made oversized, so that a rubber tube fits around the bolt going through the hole.  Also sandwiched under the feet and the other end of the bolt are flat rubber "washers".  When all clamped up, the assembly can absorb vibration in all directions.  The rubber is chosen to the rather hard, so the generator doesn't actually move very much.

[SparWeb]

Might be helpful to show a top view of the turbine mount.
This drawing is for the 3HP Baldor.
The 7.5 HP Toshiba motor housing is MUCH bigger, of course.

[MagnetJuice]

I used rubber and silicone rubber bushings to isolate a motor from a table. It was very effective in eliminating the transfer of the motor vibrations to the table.



At first I bolted it straight to the table and the noise was terrible. After using the rubber isolation at 2 places, all I can hear is the motor running.

I think that using a good isolation setup on the base of your motor will keep most of the vibrations from transferring to the tower.

How about building the tower in three sections and using isolation between the three pieces when bolting them together? The nuts on the bolts can be tack welded to keep them from loosening up. And if you ever move, it would be easy to disassemble it and take it with you.

If the three sections are not the same length, there is a very good chance that they will have different resonant frequencies, very possibly cancelling out each other's vibrations.

Ed

[SparWeb]

Yeah, that set-up looks very similar.  The silicone and the tire rubber will have different damping ratios, so whether you intended to or not, the combination of 2 materials probably helped even more.  I'll file that one away in the mental toolbox, too.

Not keen on building a tower with rubber isolators, though.  Small deflections at the base make for big displacements at the top.  I'm thinking of a metronome now!

Three sections of different length leads in an interesting direction.  The lower section of the tower is wider than the upper sections, so it already has a different resonant frequency.  But that doesn't mean they can't be spread farther apart with changes to the section length.  Hmmm wait, not really section length has to change.  I can change the spacing of the cross-members.  That would be most effective.  Ok, now the coin drops.  The cross-members can be spaced at varying intervals along the length of each section, then in the next section down, the cross-brace spacing could change the other way.  That could make too much variation in the structure for the vibrations to set up a big wave.

And it could look cool.

[DamonHD]

@SW

I friend of mine is an engineer on the HS2 train project here in England.

I suggested that he/they somewhat randomise placement of support pillars in tunnels and even trees outside to help avoid resonances.

When building distributed software systems, likewise I try to build in small amounts of randomness to reduce the risk of unexpected 'deadly embraces'.

Rgds

Damon

[Mary B]

Yeah, that set-up looks very similar.  The silicone and the tire rubber will have different damping ratios, so whether you intended to or not, the combination of 2 materials probably helped even more.  I'll file that one away in the mental toolbox, too.

Not keen on building a tower with rubber isolators, though.  Small deflections at the base make for big displacements at the top.  I'm thinking of a metronome now!

Three sections of different length leads in an interesting direction.  The lower section of the tower is wider than the upper sections, so it already has a different resonant frequency.  But that doesn't mean they can't be spread farther apart with changes to the section length.  Hmmm wait, not really section length has to change.  I can change the spacing of the cross-members.  That would be most effective.  Ok, now the coin drops.  The cross-members can be spaced at varying intervals along the length of each section, then in the next section down, the cross-brace spacing could change the other way.  That could make too much variation in the structure for the vibrations to set up a big wave.

And it could look cool.

Look close at commercial lattice type towers. You will see exactly what you mention. At the base the braces are closer and as you go up they get farther apart in some designs. Less weight is a partial reason, and as you go up the tower the loading forces decrease so the braces can be further apart.

You want harmonic resonance you should hear the Rohn 25 I have house bracketed with 2 antennas at the top. As wind speed changes the pitch of the hum off the tower shifts! Since it is on the wall of the living room it can be quite annoying during high winds LOL

[SparWeb]

Mary, I never thought of it that way before.
Damned if I can google me up an image of this...  anyway...
I always thought that a variable angle in the cross-braces simply came from an economy of parts.  By using the same cross-brace, going up the tapering tower, they would space themselves out farther apart.  From a manufacturing point of view, it's one part used 16 times, rather than 16 parts used once.
Maybe the economy just happens to synchronize with resisting vibration, in this case.

[MagnetJuice]

SparWeb, I was thinking about the rubber in the isolators being compressed really tight. But you are right when you said “Small deflections at the base make for big displacements at the top”.

In your case, isolators on the tower might not be necessary because you can have the isolators at the base of your motor.

I have seen others here mention unwanted vibrations in their towers. So I think that people that have regular axial flux alternators on their tower, can benefit from having a short section at the top of the tower isolated from the rest of the tower. That would be the section that the yaw bearing slides over. It would be so high on the tower that any deflection would not be a problem.

Mary, if your tower was isolated from your house with rubber mounts, that Rohn wouldn't make your living room sing. 

Ed

[MagnetJuice]

Mary, after I uploaded that post I thought; “there is no humor in having to put up with any kind of noise”.

If that noise bothers you, there could be an easy and inexpensive fix for that. If you can get someone to take a picture of the bracket that is attached to the house and uploaded here, we might be able to come up with a fix.

SparWeb, I don't want to hijack your thread, but if there is any interest in an isolator to eliminate vibrations from a tower, I have been working on a design.

I can post the drawings here when I'm done or I can start a new thread.

Ed

[SparWeb]

Thanks for the suggestion Ed.  Vibrations are a problem in many things, so any idea can potentially have use in 1001 places. 
Also... bitter reality... without looking I bet there are 1001 solutions already out there.
I have a few "go-to" products that work well enough for me when the problem needs to be solved at work, and from there I pick up ideas that help on the hobby projects.
It wouldn't be a thread hijack if you posted tower ideas on the thread I started for the tower.