mechanism to reduce the effect of cogging on startup in low wind conditions

[joestue]

I have previously thought about a centrifugal magnet and spring assisted ratcheting clutch mechanism.. Such that once the rotational inertia overcomes the spring tension on the clutch, a magnet provides attractive force. Since wind is a low rpm high torque.. you serrate the clutch dogs and the drum, once they make contact, its instant torque transfer.

a rubber transmission damper line prevents bad things from happening.

anyhow that may be unreliable. both due to the mechanism sticking, failing to operate, and the immediate torque transmission breaking something.

what about a simple lathe dog style coupling, except that the motor and the blades are free to rotate half a turn before the backlash is taken up.

you then add a spring such that the blades will turn backwards away from the motor in low to no wind conditions. as soon as a gust of wind can get enough torque to overcome the cogging force, the blades will rotate half a turn and the inertia will carry it over the cogging hump.

the mechanism won't wear out near as quickly as a more complicated centrifugal clutch. and potentially never would wear out if two springs are used.

one low tension spring pushes the blades backwards away from the motor, the other is short and fat and absorbs the impact.

[Adriaan Kragten]

I would prefer to use a generator with almost no peak on the cogging torque or no cogging torque at all if there is no iron in the coils. But this requires a special generator design and most commercial available PM-generators have a peak on the cogging torque which can be rather large. A problem is also that the measured peak torque is never mentioned in the specification. So you must measure it yourself if the generator has already been bought. If the peak is too high for the chosen windmill rotor, it is a waste of money. That was my experience with the bicycle hub motor as described in KD 745. A centrifugal clutch to make that the rotor starts unloaded seems to complex to me.

[electrondady1]

what problem is  your clutch mechanism is solving ? cogging?
isn't an air core alternator the true solution?

[joestue]

what problem is  your clutch mechanism is solving ? cogging?
isn't an air core alternator the true solution?

if you like spending a grand on magnets, yes.

[Adriaan Kragten]

what problem is  your clutch mechanism is solving ? cogging?
isn't an air core alternator the true solution?

if you like spending a grand on magnets, yes.

If you compare two generators with the same maximum torque level and one with a laminated iron stator and one with no iron in the stator, you will find that the magnet costs of the one with an iron free stator are about a factor four higher. This is because you need much thicker magnets to get an acceptable high magnetic flux in the coils or you need a much bigger armature volume. As the prices of neodymium magnets have been doubled in recent years, this is a strong stimulation for generators with a laminated iron stator. However, if you chose for a generator design with a high peak on the cogging torque, you need a device to make the rotor start unloaded. But even if you can design a working device, this device goes at a certain price and I am afraid that this price consumes all profit of the lower magnet costs. So it is much better to chose a generator design with an iron stator but with a low peak on the cogging torque. I have described several options to realise this. The most prommising option is described in chapter 7 and 8 of report KD 718 as this option also allows the use of the original 4-pole motor winding.

[MattM]

If you want to generate at lower wind levels the only real answer is a wind turbine for that band of wind speeds. 

High starting torque is a feature.  Saves you wear and tear on bearings.  Seems dumb to spin your rotor just to look pretty and not to be actually producing meaningful output.  If its looking pretty without purpose then its lawn art.

[bigrockcandymountain]

If I can succeed in building a generator out of an old motor that has acceptable levels of cogging, then anyone can do it.  It really isn't the issue it sounds like when you read the internet. 

I would also try only using magnets for the N poles if i were to convert another motor. That would lower the magnet cost by half. 

I just looked and the magnet cost for what i did on my motor conversion would be $350 usd today.  If you only covered half, you would be down to $175. That is for a 2kw continuous generator (very conservative rating).

I agree that an air core axial flux is more efficient, with better startup etc but you won't catch me building one any time soon.

[Adriaan Kragten]

The generator as described in chapter 7 of KD 718 makes use of a 3 kW, 4-pole asynchronous motor frame size 100L. It uses 48 magnets size 40 * 7 * 3 mm which cost only € 0.96 a piece including VAT if 200 magnets are ordered. So the total magnet costs are about € 46. This generator will have a maximum torque level which is high enough for a windmill rotor of about 3 m diameter and a design tip speed ratio of 6.5. If you use € 46 for the magnets of a radial or axial flux generator with no iron in the coils, you will get a generator with a much lower maximum torque level and therefore it can drive only a much smaller windmill rotor.

The magnet costs of the iron free radial flux generator as described in my recent report KD 748 are € 92 and so a factor two higher but this generator can be coupled to a windmill rotor with a diameter of only 1.8 m and a design tip speed ratio of 6.25. If you would design a similar generator for a rotor with a diameter of 3 m and the same design tip speed ratio, the required magnet volume would increase with about a factor (3 / 1.8 )^3 = 4.6. This means that the magnet costs will then be about a factor nine higher for the same rotor diameter. So the factor four which I mentioned in my earlier post is much too optimistic. This clearly demonstrates the big advantage of using a laminated iron stator and thin magnets on the magnet costs.

[Adriaan Kragten]

The cogging or sticking torque of a PM-generator at low rpm has two reasons.

The first reason is the friction of the bearings and the oil seal on the generator shaft. Most generators have rubber sealed bearings to make that the grease stays in the bearing. These seals are not good enough to prevent entrance of water at the shaft side if a long lifetime of the bearings is wanted. An extra oil seal is therefore needed at the shaft side. The friction torque of this oil seal is much higher than the friction torque of the bearings, especially if the oil seal is new. The space in between the oil seal and the front bearing has to be filled with grease to create an extra barrier for the penetration of water.  Very accurate measuremnts have been perfomed to the sticking torque of the original VIRYA-3B3 generator. The unloaded Q-n curve is given in figure 2 of public report KD 78. The sticking torque caused by the bearings and the oil seal is 0.4 Nm for n = 0 rpm.

The second reason is the cogging due to the variation of the magnetic flux in the stator stamping. The original VIRYA-3B3 generator had magnets mounted in grooves which make a certain angle with the armature axis an the cogging torque is therefore not fluctuating. The cogging torque due to iron losses is therefore zero for n = 0 rpm. It increases about linear with the rotational speed if the winding is rectified in star. It increases much stronger if the winding is rectified in delta because higher harmonic currents can circulate in the winding for delta connection. So star connection if preferred if a low starting wind speed is wanted. In figure 2 of KD 78 it can be seen that the total unloaded torque is 0.8 Nm for n = 500 rpm for star rectification. This demonstrates that the friction torque of the bearings and the oil seal has a relatively high influence on the total torque if the generator is designed such that the cogging torque isn't fluctuating.

If the generator isn't designed properly, the peak on the cogging torque can be very large (see measurements performed on the Sparta bicycle hub motor as given in report KD 745). So for such a generator, the peak on the cogging torque is the main reason for a high starting wind speed. The starting wind speed can be calculated with formula 8.6 of my public report KD 35. But you must know the starting torque coefficient of the rotor.

[SparWeb]

Joe, if you want to build a mechanism, I believe the best is to pitch the blades, not clutch the shaft.
You can copy something like a Jacobs (plans and patents available for reference) or one of the Mass+spring-return types.
The great thing about the blade pitch is that you can tune it to match the generator not just at one ideal speed, but a wide range of speeds.  Boosts your energy capture.
It used to be the holy grail of matching blades to generators, before MPPT came along.

[Adriaan Kragten]

Joe, if you want to build a mechanism, I believe the best is to pitch the blades, not clutch the shaft.
You can copy something like a Jacobs (plans and patents available for reference) or one of the Mass+spring-return types.
The great thing about the blade pitch is that you can tune it to match the generator not just at one ideal speed, but a wide range of speeds.  Boosts your energy capture.
It used to be the holy grail of matching blades to generators, before MPPT came along.

Technically it is possible to design a pitch control mechanism which starts with a big blade angle, then goes to to normal blade angle corresponding to the optimum tip speed ratio for normal wind speeds and then goes to a small or negative blade angle for high wind speeds to get active stalling. But such a mechanism is certainly more complex and more expensive than a centrifugal clutch in between the rotor and the generator. So you should simply not use a PM-generator with a high peak on the cogging torque. The reason why people still want to do this might be that this kind of PM-generators are cheaply available on the second hand market like for instance direct drive motors of washing machines or bicycle hub motors. But what you save at the generator costs you have to spend much more somewhere else to solve the starting problem.

[SparWeb]

Hello Adriaan,
I don't disagree with your comparison between blade pitch control and electrical modulation of the generator such as MPPT, but what I meant to compare was the clutch mechanism suggested in the OP with blade pitch control.  Maybe I wasn't very clear about that so here is what I really mean:  In a cost or complexity comparison between the 3 strategies, I would put pitch control in the middle between a clutch (more complex and expensive) and MPPT (cheapest and simplest).  I should also note that reliability of the 3 strategies also follows the same trend.  Expect a clutch to fail rather quickly on a wind turbine.

When I drive to work in the morning, I probably operate the clutch in my car about 100 times.  Same on the way back home.  I've had my car for 10 years, and replaced the clutch 3 times.  I've put 300,000 km on it so let's say a clutch lasts 100,000 km on my car.  Or I could say that a clutch lasts about 3 years.  That's about 180,000 clutch operations before I have to replace it. 

Put a clutch on a wind turbine, where it must close and release any time the wind passes through a given speed, and it will operate anywhere between 10 to 100 times in an hour.  In a year that would be between 100,000 to 1 million operations.  That clutch will probably fail in less than a year depending on the nature of the wind and the wind speed setting chosen for it to operate.

Sorry, Joe. 
I can't get behind a clutch idea unless you have a variety that doesn't need friction.  Actually, your idea is more like a sprag clutch, but I don't think they get more life, either.  Also, a sprag fails abruptly, while a friction clutch fails progressively, getting slippy and noisy, which warns you of trouble before anything disastrous happens.

[Adriaan Kragten]

The clutch as mentioned in the first post is meant to overcome the peak in the cogging torque. This peak can easily be a factor three higher than the average cogging torque if the generator isn't well designed. So once the rotor has started, the fly wheel effect of the rotor makes that only the average torque has to be supplied. This means that once the rotor has started, it will only stop if the peak torque of the rotor is lower than the average unloaded torque of the generator. The peak torque coefficient is about a factor ten higher than the starting torque coefficient for a rotor with a design tip speed ratio of 6 (see figure 3 report KD 710). This means that the rotor will stop only at very low wind speeds. So I don't think that the clutch will be activated many times in an hour.

A disadvantage of the use of a clutch is that now the generator bearings can no longer be used to support the rotor. So the rotor needs its own bearings and a shaft in line with the generator shaft. This shaft needs a support at the head frame. So appart from the clutch, you also need these extra components.

[MattM]

Maybe instead of a clutch you could use more of a centrifugal magnetic transmission driven off a nosecone mounted prop.  The coupling (gap between magnets shortens) is triggered by rotating forces to engage the magnetic gears.  No clutch plate to wear out.  Requires your nose cone to reach useful energy state to trigger the main turbine.

Simpler yet, could you use eddy currents to act as your clutch?  Rotate a few magnets over a copper plate.  Once the drag from eddy currents creates enough drag your larger rotor kicks in.  If anything, at a certain point you would create a bunch of heat so you may need a thermal decoupler.

[MattM]

So after a little cursory search, you don't get much torque out of a disc.  On  the report I found, a 60 slot magnet array evenly distributed across a steel disk saw only 3 Nm at approximately 250 rpm.  So it probably makes more sense, from a material perspective, to create an axial ribbon (in a perfect circular belt) and then to run the magnets axially.  Your final tangential speed plays the biggest factor in it rather than simply rpm.  Still sounds much simpler than a clutch.