Require high speed alternator/motor design critique.

[KatyDave]

I am building a turbo prop and I would like to incorporate a permanent magnet machine that could act as a starter and an alternator. I am an electrical engineer, but more on the control side. The best idea I could come up with, remembering that this is a “home built” project, is at this link:

http://www.scltd.ca/SCL/Alternator.pdf

Using magnets diametrically magnetized allows the equivalent of a two pole motor (only three windings), reduces the electrical frequency, and should increase torque. Am I wrong?

At 20K RPM the windings are rather short (<5”) to generate 15V, so there will be many parallel windings in each phase. Maximum operating speed will be less than 90K RPM. I plan on using a microcontroller to control a HexFET bridge, so I can limit current in the starter mode and then optimize and control voltage in the alternator mode.

I would appreciate any feedback.

 

[Skogsgurra]

There are quite a few high-speed designs available. The main issues are usually heat build-up in the rotor, balancing and bearing problems.

Heat in the rotor may come as a surprise since a PM rotor is supposed to run without any slip and hence no losses. But that would only be true if your flux were perfectly sinusoidal (which it can be, but needs a lot of effort) and your driving EMF is perfectly sinusoidal, too. You can have that if you use a PWM inverter and filters or if you use an inverter with pure sine-wave output. There are some available, at a price.

The need to keep heat down is because epoxy doesn't work well above, say, 140 centigrades.

Balancing is obvious.

Bearings. Hybrid bearings are the best choice in high-speed applications. There are several reasons. Bearing currents are zero. Centrifugal forces are minimized, they increase forces between balls and outer race in an all-steel bearing.



Gunnar Englund

http://www.gke.org

 

[KatyDave]

I think my windings may be in the wrong direction, should they parallel the axis of the magnet, or radially as shown in pdf?

 

[Skogsgurra]

It is not easy to see how the windings are orientated.

I think that you need a lot more help than a few comments to your drawing if you need to have that type of questions answered.

Sorry, this is just to make you aware that there are a few fundamental techniques that you really need to master before you start building a high-speed motor. It is not something to be taken lightly.

Gunnar Englund

http://www.gke.org

 

[sreid]

Two books to take a look at.

http://www.amazon.com/exec/obidos/t...gy_img_2/104-2300972-6033505?v=glance&s=books

http://www.amazon.com/exec/obidos/t...f=sr_1_3/104-2300972-6033505?v=glance&s=books

 

[mc5w]

You are not going to save any weight by doing this. A starter motor needs to be a series wound, series wound with permanent magnets ( effectively a compound motor ), or a compound wound motor. A permanent magnet motor of the type used for drives would not be able to take the torque pulsations and the switching transistors would use up too much weight. A permanent magnet motor would need to be oversize to take the torque pulsations.

Also, a rotating field conventional 24 volt nominal alternator will work much better than a permanent magnet alternating current motor trying to regenerate power. You also need to cope with varying speed such as the engine speed higher when taking off than when cruising.

There is a 10,000 watt air cooled generator that used to be on the market that used an old style shunt wound 12 volt read 15 volt generator that was also used as the starter motor using a big manual pushbutton to bypass the cutoff relay that disconnects the generator from the battery under reverse power flow conditions.

The other thing is that if you try to do any redesign or modification of an airplane engine you need to get the engine recertified by Federal Aviation Administration.

 

[UKpete]

dlyon, I agree very much with skogsgurra's advice on the problems with high speed pm machines (as a result of sometimes bitter experience) - the three key problem areas are as he states, I would only add rotor dynamics i.e. the occurrence of vibration at critical speeds. The hybrid bearings (ceramic ball) are the normal choice, e.g. from Koyo, or SNFA:

http://www.snfa.com/index.php?id=61

Regarding your design, you are correct in saying that the 2-pole design has the lowest frequency and therefore there are reduced problems from iron losses and the inverter frequency is reduced. However, the high speed PM machines I have seen have been 4-pole, specifically because the end-wingings on 2-pole machines have been too large, increasing the bearing spacing and therefore reducing the critical speed of the rotor. The rotor design will give you problems - the magnets must be maintained in position very accurately at all speeds and temperatures otherwise the balance is lost, different thermal expansivities and modulii of the materials can become a headache; for the same reason the rotor restraint (and carbon-fibre is a good choice) has to be wound on under high enough pre-tension so that the magnets are held firmly against the steel hub at the maximum speed. Consequently I recommend you use a more conventional design, with a cylindrical rotor. This will mean you will need to get the magnets ground, best done by a magnet supplier because it is messy.
NdFeB magnets are also a good choice, better than SmCo because they are lighter, less brittle and have a lower resistivity (to reduce rotor losses).

The stator can be considered as more conventional, with the following provisos:
1. the lamination material must be very thin (0.1 to 0.2mm) to reduce the iron losses
2. use a large number of slots (say 12 or 18 as you have shown) to reduce the harmonics in the rotating field
3. use a one-piece stamping (your laminations look like they are in the wrong plane - they should be normal to the rotor axis) - although small quantities can be wire-cut or laser-cut, the difficult bit will be obtaining a small quantity of the sheet material
4. get as much copper into the slots as possible and keep the end-windings as short as possible so you can keep the bearings close together.

I also recommend you research the background to brushless dc motor design - the two books recommended by sreid are really good, you may prefer Hendershot & Miller because it has more detailed design information in it.

As an aside, lot of this technology has been at the R&D stage in automotive applications (motored turbo assist for overcoming turbo-lag, fuel cell oil-free air compressors, exhaust energy recovery systems etc.), and micro-turbine generators.
Some examples of high speed machines:

http://www.s2m.fr/chap4/index.html

http://www.calnetix.com/Rotordesign.cfm

http://www.bowmanpower.co.uk/turbo_alternators.html

This is a major task, it won't be easy and it won't be cheap.