high RPM / high frequency 50kW PM generator sizing

[maxmitchell]

Short Version:
I am looking for a set of equations to help size a permanent magnet (PM) generator.
The power goal is 50 kW at a speed of about 95,000 RPM.
I suspect that due to the high RPM, it would be best to reduce the pole count on the rotor as much as possible, so probably a single North, and a single South pole.
The stator should have at least 3 phases, but probably ONLY 3 phases.
The target voltage output at this RPM is anything less than 800 V, but probably as close to that as possible.

I have a few books on electric machine design, but I cannot find any equations that will help me roughly size the length or diameter of the PM rotor. Nor can I find any equations to help me find the number of turns, gage, etc of the stator winding.
Can anyone recommend a source of information to help me with this?

Background (in case anyone is wondering):
We are designing a 50 kW turbo-generator for use in a hybrid car. This will fundamentally be a turbocharger (+ combustor) directly connected to a PM generator rotor. I have built 5 turbo-thrust engines using turbochargers before, and I worked for 5 years at a company that produced “micro” turbo-fan engines (high bypass ratio, geared fan).
I am very knowledgeable on the thermodynamics, fluid mechanics and operational characteristics of small turbine engines, but am very novice when it comes to power electronics.

Right now, I have a BorgWarner S366 turbo running on a test stand producing thrust only. The next stage of the project is to design, fabricate, and incorporate a PM generator directly on the turbo shaft, sticking out of the compressor inlet side of the shaft (not near the hot stage).
I ran through the thermodynamic calculations, and the turbine should be capable of producing an “extra” 50 kW of power without exceeding the material gas temperature limits. I say “extra” because this takes into account the power needed from the turbine to power the compressor at this operating point. A rotor speed of ~95,000 RPM is needed for this operational point, thought the turbocharger is rated to over 115,000 RPM.
The generator should also be able to function as a starter motor (hence, my statement about “at least 3 phases” on the stator, since only 2 phases here would not be able to spin up the rotor from a stand-still. When acting as a starter, only about 2 kW should be required to get the turbo up to “idle” speed of ~40,000 RPM.
Voltage target at rated power is around 800V since this looks like the new standard in high-voltage hybrid vehicle powertrains.
The electrical output of the stator would likely go through some sort of simple but beefy rectifier, then used to charge some small array of chemical batteries or capacitors.

Thanks!
Max Mitchell

 

[itsmoked]

Sounds like scads of good fun!

Any chance of a gear reduction? I ask because a brushless DC motor run as a generator would be a plausible choice. They require electronic commutation and usually Field Oriented Control (FOC). FOC requires reading back the inactive phase during powering of the active two phases. At 95k the speeds required for that electrical measurement and subsequent math can have problems. Not sure it could be made to work at 95kRPM but cutting that speed in half would allow it.

I have to think about this some more.. 95k..

Oh and don't cross post in Eng-tips. It gets VERY confusing quickly when you screw that up.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[maxmitchell]

sorry about the double post. I deleted the other one.

One of the primary design constrains of this project is to avoid gear reduction.

That "micro" turbine project I worked on for 5 years...
It sent about 15 kW through a 2-stage gear reduction system. The turbine input shaft spun around 130,000 RPM, and the fan output shaft spun around 10,000 RPM.

Over half the problems we had with that product were related to the gearbox. We had to keep slapping on expensive solutions to get the thing to finally be reliable.
I became convinced that using gears to reduce the speed of something spinning around 100 kRPM could not be made to work economically with modern technology.

 

[itsmoked]

Thanks for the delete. Excellent.

I figured as much. High speed gear boxes are definitely problematic. I'm still thinking about this, there may be a way.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[waross]

Motors and generators are limited in capacity by the "Volts per Hertz Ratio".
Let's have some fun with numbers.
50 kW at 750 Volts is 38.5 amps.
95000 RPM on a 2 pole machine is 1583 Hz.
750 Volts at 1583 Hertz is 0.47 Volts per Hertz.
For size comparison, you want a generator that will produce 0.47 V/Hz x 60 Hz = 28.4 Volts at 38 Amps.
That equates to about 2.5 HP.
The point of this is that if you can find a BLDC, two pole motor that will withstand 95,000 RPM it will meet your specs.
By the way, two phase motors work fine.
They predated three phase motors in a lot of areas.
Three phase is more economical of supply conductors that two phase and two phase was phased out almost 100 years ago in favour of three phase.
You may be able to cheap out on the drive inverter.
If your inverter has enough capacity to handle the load without the motor falling out of sync, you can just ramp up the inverter output slowly.
Millions of synchronous motors have been running without any feed back or position sensing.
Most of them run their entire lives without pulling out of sync.


Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[maxmitchell]

Thank you guys for the feedback!
let me clarify a few points...

While it would be great to find an off-the-shelf solution for my electrical power goals, I am not counting on this happening.
I have a decent machine shop in my garage (CNC mill, lathe, surface grinder, TIG welder, etc), have access to Hofmann balancing equipment, and can procure individual rare-earth magnets and custom stator laminations, if needed.
Right now I am just trying to design a system that would work; No need to conform to an existing electric machine (though, that would be nice, if it exists).

@waross:
If I could find such a BLDC motor, I could upgrade the bearings and 2-plane balance the rotor on a Hofmann to get it mechanically ready for 95kRPM operation. But, would I also need to re-wind the stator to support the higher frequencies?
If so, how-so?

Regarding use as motor vs generator...
I assume I should optimize this thing to be a 50 kW generator at 95kRPM, but use some sort of relay to put it on a "starting circuit", which would only need to pump out ~2kW from some sort of modest inverter. Once idling (at 40kRPM!), I would use the relay to switch to "generator mode" and instead of pumping electricity back into the inverter, I would send it to my beefy rectifier.
Is that also what you are picturing?

How does a 2-phase stator deal with start-up? How do we know what direction we are spinning in when we first start off?

P.S.
I attached a photo of the thrust-only (for now) turbine engine test bed. Please ignore the mess, I recently had to break it down, and only managed to get it half put back together!

 

[waross]

Three phases are displaced 120 electrical degrees.
Two phases are displaced 90 degrees.
Two phase worked the same as three phase but it was driven out of the marketplace economically.
Many people call center tapped single phase two phase.
The arguments go on interminably.
Try running a true two phase motor on single phase that is called two phase.
Good luck with that.
If you can find a two pole 28 Volt BLDC motor try it as is.
What ever voltage your motor is, if you rewind it with heavier wire and fewer turns it is a good starting point.
If the available motor is 4 pole, your frequency will double.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[davidbeach]

Ultimately, I think that if you can find an electrical machine capable of operation at 95,000 rpm the rest of everything else is trivial.

 

[waross]

I have been wondering about that also, David.
maxmitchell;
A suggestion for you.
See what you can do about spinning two magnets at 95,000 RPM.
If you can accomplish that then as David said, we can help with the trivia.
Starting; Can you use air starting or some other starting method?
One challenge at a time.
Let's get a generator up and running any way we can.
Then we can work on using the generator as a starter.
How are you starting the turbo now?

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[davidbeach]

Micro turbines. They were a thing once upon a time. Probably solved everything you’ve concerned with 15-20 years ago.

 

[waross]

Thanks for that tip, David.

Typical Microturbine Construction

Microturbines are a simple form of gas turbine, usually featuring a radial compressor and turbine rotors and often using just one stage of each. They typically recover exhaust energy to preheat compressed inlet air, thereby increasing electrical efficiency compared with a simple-cycle machine. The air-to-air heat exchanger is termed a “recuperator,” and the entire system is typically called a recuperated cycle.

Figure 2 shows a cutaway view of a Capstone 65-kW microturbine illustrating how these major components are arranged in a commercial product. The assembly is often called a “turbogenerator,” as it includes all the microturbine components plus the generator. The single shaft of turbine, compressor, and generator rotates at high speed—96,000 rpm in the case of the Capstone C65 turbogenerator. Generator output is therefore high-frequency AC, which must be conditioned using power electronics to provide a useable 50 or 60 Hertz electrical output.

Link to Complete Article

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[maxmitchell]

I think that I can make a rotor of PMs spin that fast. I might have to wrap it in Kevlar or carbon fiber to take the centrifugal forces. I know, that will hurt my "air" gap, but this is part of the plan.

I just have no earthly idea how axially long it should be, given a (centrifugally limited) Outer Diameter (OD), based on material mechanical limits.

I don't want it super long, or then I'm going to have "shaft whip" problems.

My (very crude) estimate is 30mm OD x 150mm long.
Does that seem in the ballpark?

...I think I can keep something like that from exploding...

We must maximize OD (without exploding centrifugally @95kRPM) so as to reduce length.
Lower length increases rotor stiffness --> reduces 2nd order (G2) vibration.
This is information from when I worked at Garrett Turbo. We had big G2 balance problems on EcoBoost turbos, at the time.


Also, I'm not planning a recuperator or regenerator...
I hit 22% thermal efficiency on my current simple setup, and that's good enough for now.

Regarding starting:
I am using a 60V electric leaf blower to start now. It works quite well...gets her up to 40kPRM without trouble, once I feed in a little propane. 600 degC EGT during start sequence...no problem.

 

[waross]

Cut and try.
Pick a size that you like and calculate the theoretical output.
Then scale up the length and/or diameter to get the output you need.
One issue that we must address is eddy current loss in the stator laminations.
Any suggestions as to means to reduce eddy currents?

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[waross]

Years ago I reviewed the specs of an aero derivative turbine of about 750 kW.
It was for sale used.
The shaft had been extended to allow more space between the compressor and the turbine.
The air flow ducting had been extended to the side similar to your combustion chamber.
The added length of the combustion path afforded enough transit time to completely combust the fuel; finely divided sawdust.
My boss wouldn't let me buy it.

Bill
--------------------
"Why not the best?"
Jimmy Carter