Continuously variable "transmissions" and wind turbines. Has it been tried?

[teelo888]

Hello everyone! As you can see, I'm new here but have been reading a lot of content on this forum over the past couple weeks. This is by far the best resource for wind energy info on the entire internet.

To preface, I'm in the renewable energy policy field and have always had a strong interest in everything science. For this reason I felt that I should take the time to completely learn how wind turbines operate. During this pursuit of knowledge, I have discovered some of the drawbacks of modern wind turbines. Please correct me if I'm wrong, but it seems that turbines having a discrete gear ratio are forced to rectify the AC to DC, and then invert back to AC that is synchronized with the grid. If so, aren't these conversions responsible for power loss? And if so, wouldn't it be better to be able to produce AC current that is synchronized with the grid straight from the generator just like it is done in fossil fired plants?

This is the chain of thoughts I had that led me down the rabbit hole of the history of continuously variable transmissions and different designs. What I have discovered is that most designs have significant frictional losses and possible differential slippage, but if we assume that these were negligible, what benefits would we get from the CVT versus average designs? What would be the drawbacks?

Also, it seems to me that even if you employed a CVT and were able to maintain constant synchronized generator output, you still wouldn't be much better off unless you were able to take advantage of very high winds with your variable gearing because you would still only be spinning one generator the same speed (i.e., you have extra torque you aren't using, right?). So unless you had an additional way to create more power, the CVT doesn't mean a whole lot, right? Or am I wrong?

Thanks!

[joestue]

The electrical losses associated from converting dc into ac are significant, but in theory at least, they can approach zero at low load output and remain at 98 or 99% efficiency across a 10:1 range in load.

While you are probably aware that converting from ac into dc has some constant losses... at 600 volts, 2 volts of diode drop is close enough to zero that we often have to pay more attention to harmonic current (drawn by the rectifier) inducing losses in the generator and we can ignore the diode losses.. and so basically you have to optimize the generator to be purpose built for a rectifier load. There are a few ways to do this.. incorporate a 30 degree shift in the windings and rectify separately for a supposed 6 phase output.. or you could go to an odd number of phases.. 5, 7, 9, etc, or you can add phase shifting transformers after the generator to develop more phases and present a more perfect sinewave, unity power factor load on the generator.

So what i'm saying is generators designed for sinewave outputs can have lower iron and copper losses than generators designed for a trapezoidal waveform optimized for a 6 diode bridge rectifier.. but i digress... that's a last 1% issue.

Short answer is the generator can deliver dc at almost the same efficiency as it can ac.
A further advantage of the ac-dc-ac system is the generator can run at a much higher frequency than 60 hz.... neglecting iron loss, a motor's power output at constant efficiency follows rpm. double the rpm, or double the poles and the frequency, you double the shaft hp. there is a limit of course, I said "neglecting iron loss" for a reason.

If you have a lossless CVT transmission.. which for imaginary purposes only, you can approach 98% efficiency with an 18 speed bicycle, and the 10% step in gear ratios is close enough to infinite for instructional purposes, well now the generator has to run at constant speed and constant volts per hz for back feeding the grid..continuously.

But in low wind conditions, the prop isn't generating enough power to spin the generator due to its constant losses.
But it can still generate enough power to feed a dc to ac inverter.

Alternatively you can wind a synchronous or induction motor for multiple poles and run the generator at one half the nominal rpm..
However at one half nominal rpm, you have 1/8th the shaft hp available from the turbine due to the cubic power available from the air.
By cutting the generator speed in half, you've cut bearing friction in half, you've cut windage losses to a tenth.. but you haven't really cut the electrical losses in half, and furthermore the multipole winding takes up very valuable space in the generator costing you full speed efficiency.

Alternatively ways around this include:
Ac to ac inverters. they are called matrix converters, and theoretically they offer about half the electrical losses of a dedicated ac to dc to ac converter of traditional methods.


alternatively you can have a wind turbine with multiple generators and each generator has a clutch.

imagine a flywheel from a car.. and its starter.. usually you have a 20:1 gear ratio or something like that..
but now imagine you have a flywheel with 20 starters. 10 on each side. each one has a clutch.
there have been some attempts at building wind turbines using this method btw.
so now you don't have to have the frictional losses of all of them running at all times. instead, you have clutch wear.

[teelo888]

I don't know a lot about electricity and generators, not nearly as much as you do, but what I'm getting out of this is that a single discrete gear ratio is better?

It really seems to me like it would be ideal to be able to actively alter a gear ratio, not have to rectify and then invert, and keep a generator in sync and directly tied to the grid; as long as one is able to take advantage of extra torque and higher wind speed. The clutch idea with an additional generator like you mentioned is actually something I had in mind that seems like a good idea.

we often have to pay more attention to harmonic current (drawn by the rectifier) inducing losses in the generator

I'm a bit confused by this. Why would the harmonics induce losses in the output of a PMG? Or do you just mean inducing losses after the rectifier? The way it's worded it makes it seem like it actually alters the power the generator produces.

Anyways, I actually have no idea whether a single phase or three phase setup (you can tell how little I know about this field) would be ideal for this, but for something industrial scale, I was thinking of something like a 32 pole single phase whereby the controller kept the generator running at a constant 225rpm with a 5-45rpm rotor speed operational range. I know even less about average residential wind setups, but I do know they turn much faster (on up to around 400rpm or so?).

Perhaps the drawback of being unable to turn the generator at a higher frequency (because you are keeping it bound to grid frequency) is really the nail in the coffin of the CVT concept, even if it does prevent you from having to convert the power a couple of times.

From the way you describe everything, there is practically no loss when converting from AC-DC-AC; at least far less than I was led to believe from what I had read most elsewhere.

[SparWeb]

The solution that's better than either fixed gear ratio or variable blade-generator ratios is the variable pitch prop.
In that scenario, the aerodynamics of the blade are optimised, and the pitch-control mechanism keeps the blades close to the optimal point.
The optimal point is determined, partly, by the most efficient speed of the electric generator. 
In larger wind turbines, the blades must turn slower but generators want to turn at the same speed.  This system still benefits from a gearbox because it can be lighter than using low-speed generator or blades that must operate faster (hence stronger, hence heavier).  General Electric, Enercon, and Vestas have made big steps in creating direct drive WT's to eliminate the gearboxes altogether.

CVT's are practical in systems where gear ratio must be changed frequently, which is why they are growing in popularity in vehicles.  The last time I talked to an engineer who had some experience with them, he pointed out that efficiency only goes up with cost, and the cost curve is steeper.  As automakers are pushed toward higher efficiency, and some turn to CVT's to solve this problem, there may be innovations that turn the cost equation around.  But currently
While the case could be made that a wind turbine's gear ratio could be changed regularly to suit changing wind speeds, I think the weight of an economical CVT gearbox would be prohibitive. 
Correct me if you know of some lightweight examples!


All mechanical power conversions, just like electrical AC-DC conversions, do have losses (Law of thermodynamics).

[joestue]

Why would the harmonics induce losses in the output of a PMG? Or do you just mean inducing losses after the rectifier? The way it's worded it makes it seem like it actually alters the power the generator produces.

The harmonic content drawn by the rectifier causes extra copper losses in the generator. For a three phase rectifier with a resistive load, the extra losses are on the order of 5% more than they should be. so when you factor in 2 volts of diode losses... at 480 vac you have about a 99.5% efficient system.. and if the generator was 98% efficient back feeding the grid, now its 97.9 percent efficient feeding the three phase rectifier.
this adds up to an efficiency of 97.4%

Single phase motors have a power density of about two thirds what a three phase motor has, so there's no point in building them.

generators have constant losses just to keep them spinning, this loss ideally would be about half what the full load losses are.
for a very large generator they are close to zero in terms of reaching 98% efficiency. but they are not close to zero when it comes to keeping them cool.

The direct drive generators being produced in the megawatt and larger sized wind turbines require a lot of prasodymium  because the magnets will reach 80C and higher temperatures.

and the power density of those direct drive generators is horrible. the 7 megawatt direct drive neodymium generator by GE, if i'm recalling the same turbine.. is larger in diameter and thicker than the 120 megawatt generators in the grand coulee damn. (granted the hydro turbines are spinning at about 60 times the rpm though).

From the way you describe everything, there is practically no loss when converting from AC-DC-AC; at least far less than I was led to believe from what I had read most elsewhere.

losses are relative,
given a large enough system 95% efficiency is not good enough, but it is easily attainable.

most off the shelf VFD's are about 95% efficient.
most grid tie inverters sold to residences are advertising 98% efficient.. they also cost a whole lot more than a 95% efficient VFD. but they have to do another power conversion step between the low voltage panels and the grid.

[Ungrounded Lightning Rod]

It took me a little while to "get" this, too.  But after a bit it sunk in.  So now I'll rant in your direction and see if I can save you some time.

In a renewable energy situation, such as a wind generator, thinking of "efficiency" in terms of percentage of power from the prop making it to the outlet is thinking the wrong thing.

OK, say you lose 21% there.  Make your blades 10% longer and you've got as much output power as if the slightly smaller prop-to-outlet system was running a perfect 100%.  There's lots of wind, sun, and moving-water energy out there.  You just need a system harvesting enough that the amount it delivers after its internal losses is enough for your needs.

The "efficiency" you're REALLY looking for is power delivered divided by cost of system.  That cost includes not just the up-front costs, but also the costs of maintenance, downtime, and increased risk that you never get it done.  But let's just stick to up-front, out-of-pocket buy, build, and set up costs, in money and labor.

Suppose your transmission is perfect:  No losses, no downtime, no extra maintenance or special lubricants, etc.  Suppose, on the average, it gets you an additional 21% of power by reducing losses and/or tuning the mill better to the minute-by-minute wind conditions.  But also suppose it costs you an extra 50%.

Now suppose, instead, you get your extra 21% of power by making your blades 10% longer and adjusting the other parameters accordingly.  Slightly bigger/more magnets.  A little more wire.  It's probably less than a 10% incremental cost, due to a number of fixed costs dominating and producing economy of scale.  But some of the incrementals are a tad higher than linear because the longer blades turn a bit more slowly for a given TSR, and some are proportional to power rather than size.  So let's be unrealistically pesimistic and call it a 12% price boost.

Deciding between 12% vs. 50% extra cost to get 21% extra power should be a no-brainer.  Then look at the rest of the incidentals:  The transmission WILL take more maintenance, WILL produce more downtime (if only while you do the extra maintenance), and will give the mill more ways to fail and more ways to be out-of-tune.  Oops!

Thus you see us gravitating to dog-simple designs, to keep the complexity and investment down, the reliability and watthours-per-dollar up.

[DamonHD]

Nicely put, ULR!

I'd say there are circumstances where actual efficiency may matter, eg solar PV on a small property (ie constrained by roof space) to cover usage, but even then you can usually get better improvements elsewhere, eg through conservation efforts, than sweating a few percent on the nominal conversion efficiency.

(Says the man with Sanyo HIT PV, and aerogel in the walls!)

Rgds

Damon

[Ungrounded Lightning Rod]

Nicely put, ULR!

Thanks.

I'd say there are circumstances where actual efficiency may matter, eg solar PV on a small property (ie constrained by roof space) to cover usage, but even then you can usually get better improvements elsewhere, eg through conservation efforts, than sweating a few percent on the nominal conversion efficiency.

Sure.  Or limited water flow that can be diverted through a hydro system, or limited area for a wind turbine, and so on.

There's usually no point in throwing away power unnecessarily (though some of our designs deliberately discard power in order to achieve some other, more valuable, objective, such as lowering the wind speed of cutin without sacrificing performance in high winds, or simultaneously avoiding runaway mills and overcharged batteries.) There are a number of subtle things that can be done, at little or no cost, that may reduce losses or increase collection, WHEN IT'S USEFUL, and these may give you substantial return on investment.

It's just that complexifying a design, boosting its cost, to chase a few percent of losses, is usually nowhere near as productive as keeping it simple and making it bigger.

The thing about renewable energy is that your fuel costs are zero, leaving you with just the cost of the collection system.  That is a much different economic regime than, say, a fossil fuel plant, where every bit of fuel costs, and the scale is so large that it's worth chasing almost every fractional percent of loss and turning it into revenue-generating power on the wires.  Our situation produces a need for a different way to think about efficiency and tradeoffs.

For us it's raining soup.  We just need to buy, or build, a big enough bucket.

[teelo888]

First off, thank you to everyone so much for all of this great information, and thank you for not hand-waving away a newbie like me that has asked a [probably stupid] question.

Sparweb, I totally agree that variable pitch props are definitely the answer on the wind speed high-end. From what little I know about them, they seem like an elegantly simple way of altering turbine speed.

I think the weight of an economical CVT gearbox would be prohibitive. 

I honestly don't really know of any industrial scale CVTs, so it would be hard to judge. My CVT design is probably comparable to the weight of a traditional gearbox, so it probably wouldn't save any weight for sure.

ULR, thanks for framing this issue in that perspective. I'll have to say that, yeah, you're probably right. I suppose those of us with the psyche that is always trying to find ways to improve the efficiency of a design tend to get caught up in that and lose sight of the financial or business implications of those design changes.

[teelo888]

Wanted to segregate off this last thought into its own reply and get a bit of insight into any advantages the following would have.

Assume that the idea of bounding the generator to grid frequency is no longer something that I'm advocating, you all have talked me out of it. So for our purposes here, assume this is a normal setup (apart from the gearbox) that is allowed to output power at whatever frequency.

(With the respect to only industrial sized applications) What if one were to use a CVT for the primary purpose of boosting output at wind speeds at and just above rotor cut in speed? Would this not be advantageous assuming that the CVT design had no significant inherent frictional or differential slippage losses? Or, does a typical generator provide enough resistance that it wouldn't be physically possible to "gear up" the generator speed in low wind to artificially produce power that would be achieved higher wind speeds?

For example, in my CVT design, it is possible to go from a 1:3 ratio to a 3:1 ratio (with everything in between, of course). Thus, theoretically, the CVT could truncate to a position that would allow the least amount of resistance for the rotor to cut in, and as soon as sufficient momentum is achieved the CVT could begin its transition to a higher gear as long as the electronics determine that there is enough torque available to do so. This seems like it could greatly improve the left side of a turbine power curve, where output power is inhibited by lower wind speed even though the turbine is turning, and from what I have seen it looks like most turbines don't reach maximum power output until ~13m/s wind.

[metalmangler]

Hi Folks
The reason why the large wind turbins are going direct drive after so many gear box failures is something that is well
understood in another field(aviation), and it is this, that as the blades turn, they must pass the tower, and when
they do the blade that is passing the tower decelerates(very slightly) and then acelerates again after it has passed, this introduces a extreamly short lived torque spike, that aproches infinity.......over and over,day after day,year after...crunch, very basic metal fatigue situation.
For a long time this was a problem that was just delt with useing an evolutionary process of keep useing whatever worked, it has only
been fairly recently that equipment capable of measuring these very short duration spikes in the torque have existed..
Its where the indivual gear teath come together that the failures occur. not that there is much evidence left when this happens.
In aviation the torque spikes are caused by the individual combustion events, wooden props were the only thing that would withstand
the torque spikes for many years, and still today there are many aircraft that have very specific restrictions at operating at certainRPM,s.
In come the variable pitch props, now you can tune the engine for smooth running at one rpm, and let the prop take care of power inputs.Jet turbines of course dont suffer from this and have very reliable gear boxes.
Another factor against transmissions is that wind turbines are 100 percent duty cycle machines.Higher duty cycle means fewer,heavyer
parts.
The reasons that cvt,s and all other transmissions work in cars is that there is always some kind of clutch taking out some of the
torque spikes,which for a given power happen in quicker,smaller events and that cars are very low duty cycle machines...
As thie phenominon of these toque spikes was historacly only of interest to piston engine aircraft its not that suprising that the
enginears buildin the gear boxes for big wind didnt know about it, or if they did they were trained to ignore "transient spikes",
almost anything else useing a gear box this big would be electricaly driven and therefore would only experience a transient durring
infrequent starts and stops, not once every second amd a half....
Metalmangler

[Mary B]

Keep in mind there is very little energy available at lower wind speeds and the complexity needed to try and capture it is not cost effective...

[SparWeb]

This is link is a well-written primer on wind power.  Readable and comprehensive.

http://drømstørre.dk/wp-content/wind/miller/windpower%20web/en/tour/wres/index.htm

After reading that, you will appreciate the limitations of low-wind-speed power, and the futility of chasing it.

Commercial wind turbines (successful ones) already can claim to capture a significant fraction of the available wind energy through most of their power curve.  Any argument about improving the technology of wind turbines will revolve around cost, or simplicity (same thing, really), or durability, not efficiency.

[teelo888]

Keep in mind there is very little energy available at lower wind speeds and the complexity needed to try and capture it is not cost effective...

So after I read your comment, I immediately typed out this reply:

Ok, I'm mindful of the fact that there is less potential energy in a slower wind. Let's also assume for the sake of this discussion that costs and complexity are not relevant to our interests for now. I'm wondering if an industrial scale generator provides enough rotational resistance to make it impossible to dynamically gear up in lower wind speeds? It seems that, for example, if you had a constant 8m/s wind, the power output capabilities at that time would not be limited by torque, as I would think that the rotor would have plenty of torque available to turn the generator at a much higher speed if the gearing was different (i.e. that the generator didn't inherently provide so much resistance as to slow or stop the rotor).

However, before I submitted it I had this lingering feeling that what I'm suggesting is trying to break the law of conservation of energy. I went back to the Betz Limit Wikipedia article, and somehow from there ended up here: https://www.windynation.com/jzv/inf/how-much-power-will-wind-turbine-produce which outlines the basic "power in wind" formula which ultimately led me to https://en.wikipedia.org/wiki/Wind_turbine_aerodynamics. I found some power output numbers here: http://www.wind-power-program.com/large_turbines.htm, specifically this image , and applied this wind power formula to these numbers after looking up the specs on the Vestas V90 turbine (these graph numbers are supposedly from the V90).

What I learned is that this turbine leaves very little on the table, even in low wind. If someone has some more reliable data for an industrial scale turbine, that would be awesome, but for the sake of hopefully providing some insight to other newbies like myself, lets take a look at the numbers we are given.

Wind power = 0.5 * air density * swept area * wind velocity³
Because the Vestas V90 in question was offshore, we will use standard air density at sea level and at 15C which is 1.225kg/m³
The V90 has a swept area of 6362m²
The graph looks like the V90 was logged at producing about 900kW at 8m/s wind speed, so lets figure the theoretical amount of power the wind contains.

P = 0.5 * 1.225 * 6362 * 8³
P = 1,995,123W or P = 1,995kW of potential wind power in 6362m² of 8m/s wind.

Looks like the turbine is missing an awful lot, right? Well, this is neglecting the Betz Limit, which we must take into account. According to Betz, of this 1,995kW, we are only able to physically capture 1,177kW. Once again, the Vestas V90 produced ~900kW at this wind speed, and when we divide total potential by output (1995/900), we get 45.11%. According to that webpage, this fellow says "The peak efficiency is close to 45% and the design achieves this efficiency in the range from 8 to 10 metres per second," thus the estimate of 900kW must have been close.

What I learned was that modern turbines already do one hell of a job capturing energy in low speed wind. I don't know if other turbines have similar outputs as the V90 in the 8-10m/s range, but if they do, there is very little energy that isn't being captured.

To go back to my idea of altering the gear ratio at low wind speed, I now feel that this would have an insignificant effect on total turbine output. Furthermore, I now believe that altering the gear ratio in low wind and expecting the rotor to not encounter an additional amount of resistance (of a sum that didn't allow power output to increase) is a bit crazy. The only thing I was right about was that what I was suggesting was trying to neglect the law of conservation of energy. The generator must provide some amount of resistance that prevents the rotor from turning faster, and trying to maintain the same rotor speed and alter gearing such that you ended up turning the generator faster would seemingly be making an attempt to create more power from nothing.

If anyone that is more familiar with this topic than I am could provide some more insight, I would greatly appreciate it. You fellows have been tremendously helpful thus far, and once again, thank you so much.

[SparWeb]

Taking a Vestas as your example, you have selected a machine that has been engineered to the last millimeter (hence a good example to show that there is very very little energy "left on the table", so to speak).  The market for the Vestas mega-turbines accepts complex controls and elaborate maintenance because there is a gain in performance (at both ends of the wind speed spectrum, in fact).  This allows the machine to be optimized across a broad band of wind speeds.

I've geeked-out on my turbines in the past, with instruments and data logs, performance analysis, and bench tests to corroborate certain analytical points.  It was all fun, but all it proved is that small wind turbines with fixed-pitch props are less efficient than the biggies with variable pitch, yaw drives, and microprocessor controls.  Something I already knew.  Of use to me was that the prop size was matched well enough to the generator that in medium wind speeds, the generator was turning at the right speed corresponding to the available wind power.  So I have mine optimized for a small wind speed band, only.  I was also able to see that my machine was likely to never run away in a strong wind, so I left it at that.  The current machine has been up and running trouble-free (some maintenance) for many years.  Which is all I really need!

[joestue]

given the wind power available follows the third power of the windspeed.

windspeed= rpm, torque rises to the square of the rpm..

given that electrical machines' energy handling abilities are linear with rpm..

thus what you really want is a gear box that increases the gear ratio with the cube of the rpm.

at 1 mph, generator, turbine turn at 1 rpm. you get 1 volt out at 1 amp.
at 2 mph, generator turns at 8 rpm, turbine spins at 2 rpm, you get 8 volts out at 1 amp
at 3 mph, generator turns at 27 rpm, turbine spins at 3 rpm, you get 27 volts out at 1 amp.

a good compromise would be a gear box that increases the ratio to the square of the rpm instead.

[teelo888]

given the wind power available follows the third power of the windspeed.

windspeed= rpm, torque rises to the square of the rpm..

given that electrical machines' energy handling abilities are linear with rpm..

thus what you really want is a gear box that increases the gear ratio with the cube of the rpm.

at 1 mph, generator, turbine turn at 1 rpm. you get 1 volt out at 1 amp.
at 2 mph, generator turns at 8 rpm, turbine spins at 2 rpm, you get 8 volts out at 1 amp
at 3 mph, generator turns at 27 rpm, turbine spins at 3 rpm, you get 27 volts out at 1 amp.

a good compromise would be a gear box that increases the ratio to the square of the rpm instead.

Do you have a link to a source that describes this in more detail?