Use of an asynchronous motor as generator for a wind turbine

[joestue]

if it weren't for leakage inductance due to geometrical effects, pole count doesn't matter. The limit is air gap flux density and thus torque limit is simply a function of rotor surface area multiplied by the radius.

electrically you get the best efficiency from core cross sectional areas that are round.. its the shortest path for a given area, and this is how the largest transformers are built up of dozens of different sizes of laminations so that the core through the middle of the coil (which might be 1 meter in diameter, is round). obviously for motors this doesn't really work very well and so you have rectangles. most motors have a limit of a 2:1 ratio for length to width of the coil.

typically 8 pole induction motors have a 3:1 ratio or even 4:1. you will notice that most hydro dams operating at 120 rpm max out at a 4:1 ratio of length to width of the coil. at some point its cheaper to make the motor larger in diameter than it is to make it longer.

[SparWeb]

Step by step I've been preparing to do some experiments to understand better what will happen.

I have spare motors, unconverted and suitable for this kind of test.

This one barely fits on the lathe.



There’s enough space as long as I don’t let the baseplate whip around and strike the bedways of the lathe.
There are boards on the lathe ways to protect them until I start to run it.

This mounting is premature because I also need to re-connect the wires to parallel-Wye so that there’s no chance of generating 480V! This is a dual-voltage motor rated at 230V/460V and it’s currently set up for 460V.

[SparWeb]

I have a plan for several tests.

All of them will have a trio of capacitors to excite the field in the motor.  Hopefully it will start to self-excite and generate power.  I've marked on the diagram that I have 45uF capacitors but I also have 270uF capacitors that may work better.   By calculation I think I need even more capacitance than that.  A friend has dug around his shop and found plenty more, so he's bringing a few boxes of them over to try.

The simplest is a 3-phase load bank.  I don't actually have the resistors to do that but I do have a schwack of nichrome wire that will do nicely.  If this setup generates anything at all - puts any current through a resistor bank, I will consider my tests a success, and take whatever measurements I can, so that I can understand better what happened.  [edit: taking a second look at those resistor specs, I realize I'd better use 200 ohms, not 20.]

If I want to go further, I can rectify the output to DC and try battery charging.  Maybe 48V will be better if the generator isn't self-exciting at 24V (because that corresponds to a very low frequency).  Maybe I'll need the resistors to start the self-exciting process and then can switch them out for the rectifiers.

And if all that goes well (who knows?) then I've come up with an even crazier idea using a 3-phase transformer.  Actually it's going to be 3 big 500VA transformers connected in a Delta.  The advantage of doing this is that it allows the V/Hz ratio of the generator to be very close to the same ratio that it was as a motor.  I think this could be necessary for it to work efficiently.  But I'm getting too far ahead now.  Gotta sort out the basic assumptions first.

[kitestrings]

I'm holding out to see the associated truck testing.  That's much more exciting ;>].

Adriaan, slightly off topic, but I wondered if you might share photos, and or information from turbines you've built or contributed toward?  And, to share things you feel have worked well, or to anticipated theory, and perhaps where you got different or surprising results from what you expected.

~ks

[MagnetJuice]

I have a plan for several tests.

That’s great Spar!

Thank you for taking the time to do some tests. Some good and useful data should come out of that.

Ed

[SparWeb]

I will be posting more, later.  Still more testing but the early results are interesting and encouraging.  I am better understanding the "why" behind the asynchronous generator schematics that Adriaan posted earlier.

Right now I just want to include a link to a book that is very useful now.

https://wiki.duke.edu/download/attachments/13373206/Johnson,_Wind_Energy_Systems.pdf

The author, (the late) Dr. Gary Johnson at the University of Kansas published this book not long after I started taking an interest in WTs and it seems to be popular enough that it still can be found on the internet.
His sections about electrical generator integration with the turbine is why I kept a copy - too many books about WTs are focused on mostly the aerodynamics or mostly the generator.  Few on the machine as a system.  Johnson's book has tips and experience about the kind of generator I am building.  I got stuck a few times and this book got me unstuck.

[joestue]

https://wiki.duke.edu/download/attachments/13373206/Johnson,_Wind_Energy_Systems.pdf

I believe i started reading his tesla coil books about 10 years ago.
didn't know he passed.

https://ece.k-state.edu/people/faculty/gjohnson/

[Adriaan Kragten]

I know that motors with a 220/440 V winding are used in countries which have a 3-phase grid connected in delta. So star-delta switching for loads with a large inertia, like it is done in Europe, isn't possible. This type of motor has a 2-layers winding which can be connected in series or in parallel. The terminal box has nine terminals and the winding is connected in star for 440 V and in double star for 220 V using brass strips at the correct positions. I think that you have such a motor. I have visited the Philippine Islands in 2007 where one uses such motors for a windmill project and wrote report KD 356 about it (not public). Figure 2 and 3 out of this report give the wire diagrams for both voltages and I have added these wire diagrams as attachment.

[SparWeb]

This was a lot of fun, and I have a lot of interesting results.  I'm not done either.  I've really only tested one capacitor size (270 uF) and one load (30 ohm) combination.  I can come up with many more combinations and see what happens.

Here's the circuit I actually tested.  Previously I drew the resistors in delta, but later realized I should use star as the starting point.  It turns out these circuits are very sensitive to phase imbalance, so using a load in Delta makes collecting information confusing, for reasons I'll explain later.

Wiring in Star also made it convenient to switch the load on and off with a single switch.  This turned out to be very important, too.

[SparWeb]

Here's my set-up.
The lathe drives the shaft of the generator.  The generator is "upside down" with its feet clamped to a long board.  The board keeps the generator body from spinning around as the lathe drives the shaft.  Voltage is being measured at the terminals of the generator where they connect to the capacitors.  You can see some extra capacitors on he right - I've got a variety of these and this set-up has two kinds in parallel.  I can add more like this for more variations of the capacitance.  The capacitors are always in Delta, but I'm questioning that it wouldn't be better to have them in Star, like the load.

I have a tachometer on the lathe with is VERY helpful right now.  The lathe is belt-driven, and when heavy loads are put on it the belts slip.  The nominal speed at this setting is 287 RPM but as you can see in the photo the slippage has dropped the speed to 269 RPM.  This little lathe is workin'!

[SparWeb]

Different capacitance values give a definite "cut-in" speed.

45 uF    404 RPM
270 uF  230 RPM
315 uF  224 RPM

From this I am guessing that to cut-in at 150 RPM this generator would need about 650 uF capacitors.
This would be suitable for a 8-10 foot (2.5-3 meter) diameter turbine.

I won't be able to confirm this until I actually HAVE enough capacitors to assemble this much capacitance, and verify this by measurement.

[joestue]

changing from star to delta or vis versa changes the voltage across the cap from 1 to 1.73, which increases the energy stored from 1, to 3.

the value of the capacitor for self excition is going to follow the inverse square of the rpm.

so at 200 rpm you will need 360uF compared to 45 at 400.

[SparWeb]

Yes, indeed, Joe.  Slowly working through all the combinations over the next few days.

[Adriaan Kragten]

I'm holding out to see the associated truck testing.  That's much more exciting ;>].

Adriaan, slightly off topic, but I wondered if you might share photos, and or information from turbines you've built or contributed toward?  And, to share things you feel have worked well, or to anticipated theory, and perhaps where you got different or surprising results from what you expected.

~ks

I have built six different horizontal axis wind turbines with the hinged side vane safety system being: the VIRYA-2.2, the VIRYA-3.3 and the VIRYA 4.2 with wooden blades and the VIRYA-1.25, the VIRYA-1.8 and the VIRYA-2.2S with cambered stainless steel blades. I have photos of all of them but as I am no longer commercially active, I think it is better to no longer stimulate interest in these wind turbines. There is a photo of the VIRYA-2.2 on the front page of report KD 35.

I have several surprising results from tests which I have performed in the wind tunnel or on the test field.
One is the discovery of the so called self orienting moment. So a fast running rotor has a tendency to turn itself in the wind. This phenomenon is described in report KD 213 and KD 223. The first measurements are described in the (non public) Dutch report R344D of the TU-Eindhoven of July 1978.
Another surprising result was that a rotor with tapered cambered blades can have a very high peak efficiency if the boundary layer becomes laminar. The main wind tunnel measurements of such a rotor with a diameter of 1.8 m and a design tip speed ratio of 6 are given in public report KD 616 but report KD 617 and KD 656 give calculations of bigger rotors.
Another surprising result was found for a wind servo for which the rotor runs backwards at high yaw angles if it had large blade angles. The main results of measurements performed on a scale model of this rotor are given in report KD 671.
It is also surprising that if a wind turbine rotor is designed according to the theory as given in KD 35, and if a scale model is tested in the wind tunnel, that the designed geometry gives a maximum Cp which is in very good agreement with the theory and that the Cp-lambda and Cq-lambda cures can be predicted rather accurately (see method as given in chapter 6 of report KD 35).
My experiments with safety systems shown that one gets a good impression of the functioning if a small scale model is made which can be hold in one hand while driving a bike or a moped.

[kitestrings]

That's unfortunate, I for one would be interested in seeing them.  Are any of the six still in operation?

I very much favor using three dimensional study models.  You always see something differently (one of mine below).  The last part is good.  I have a vivid mental image ;>]

[Adriaan Kragten]

That's unfortunate, I for one would be interested in seeing them.  Are any of the six still in operation?

(Attachment Link)

Okay then, one photo of the former VIRYA-3.3. The same 12 m tower is also used for the VIRYA-4.2 with two wooden blades. None of the six ones is in operation but the 12 m tower has been lengthened by a 2 m pipe and at this moment a small windmill is running on top of this tower which uses a small axial flux generator of Hefei Top Grand (see report KD 595).

[kitestrings]

Nice.  Thanks for sharing.  3.3 is the rotor diameter I assume?

[Adriaan Kragten]

Nice.  Thanks for sharing.  3.3 is the rotor diameter I assume?

Yes, for all the VIRYA wind turbines, the rotor diameter in m is a part of the name.

[SparWeb]

Since I've started to do more tests, inspired by this starting point, I've created a 2nd thread to focus on that.

https://www.fieldlines.com/index.php/topic,150544.0.html

Feel free to continue discussing Adriaan's experiments here on this thread, everyone.