## GE previews new high-efficiency traction motor for hybrid and electric vehicle applications

##### 30 July 2012

by Jack Rosebro

 The new prototype motor. Click to enlarge.

General Electric engineers have successfully tested a prototype hybrid and electric vehicle motor with a peak power level of 55 kW and the ability to operate continuously at up to 105 °C (221 °F), using conventional transmission fluid as the motor’s sole cooling agent.

The motor—which GE says is 3 to 5% more efficient than existing motors—could potentially extend the range of a plug-in vehicle. It was developed as part of a $5.6-million US Department of Energy (DOE) project, and performs well over a range of bus voltages, from 200V to 650V. The permanent-magnet motor uses a concentrated (solenoidal) winding, and is extremely compact, with a stator diameter of approximately 23.4 cm (9.2 inches), and a length, including end turns, of approximately 13 cm (5 inches). Power density is about twice that of today’s motors, according to GE. The motor can use a transaxle’s fluid to cool both the rotor and the stator, producing a continuous 30kW of output over a range of 2,800 to 14,000 rpm at 105 °C. Alternately, the motor can use a hybrid vehicle’s engine coolant to limit operating temperature, rather than a separate, dedicated coolant loop, GE Senior Engineer of Electric Machines Ayman El-Refaie noted that many current production motors are limited to a continuous temperature of 65 °C (149 °F). GE also developed new high-resistivity magnets for the motor, which reduce magnetic losses and reduce or eliminate the need to use more costly segmented magnets. General Electric has discussed licensing the motor technology with electric motor manufacturers as well as automakers, and hopes to see commercial application in the automotive sector by 2015. This technology is scalable and flexible enough that it can be leveraged in a number of capacities,” commented El-Refaie, noting that GE plans to explore additional potential product applications, including higher efficiency industrial motors, high-speed oil and gas compressor motors, and generators for aerospace applications. Next steps include a four-year project, during which time GE engineers will seek to produce a motor with similar performance characteristics, yet with no rare-earth magnets. ### Comments Excellent news for future lighter EVs. This would do for an AWD vehicle and could be fitted in or near wheel to reduce/eliminate some of the mechanical parts. Could also be used as a light weight genset for HEVs/PHEVs. The added 5% in efficiency would bring it close to 99% and the 2X gain in power density would bring it close to 6hp/lb. Rare Earth is not so rare. There are many potential mines waiting to be opened whenever the price is right. This is what we really need: one of the big boys to jump in and start mass producing these things. When the price on a good motor is down around$1,500 then we'll be able to use the extra to buy more batteries.

Throw in something like those $160/kWh Zinc-air batteries...and this gets really interesting :-) Would be a great motor to drop in to drive a rear axle even just a mild hybrid on the front axle would give you ~200hp 4WD vehicle which would make up for the hybrid premium. It would also help that its not that fussy about cooling if its going to be rear mounted and only used when required The high coolant temperature allows a smaller radiator and less airflow. This compounds the advantages in weight, cost and efficiency. If we want rare earths, all we really need to do is start using thorium in nuclear fuel. Thorium is a byproduct of rare-earth refining, and is expensive to separate. There is currently no significant market for thorium, so it is effectively a waste product. Turning it into a revenue stream would immediately make a lot of thorium-rich RE deposits profitable to refine. Statement: "Power density is about twice that of today’s motors, according to GE." is very misleading, tells little. If they were sure of superiority of their design, they'd for sure have said how many kW/kg, as some axial flux motor makers did. BTW Bosch already makes similar motors, a whole range. PSA Peugeot Citroen uses Bosch motor to drive rear wheels in their through-the-road hybrid. MG...average good motors have a power density of about 3 hp/lb and the best are up to 5 hp/lb. If this motor is 50% better than today's good units, its power density may be around 4.5 hp/lb. @ Harvey, There are some e-motors used in recently launched EVs with low power density. For example e-motors in Renault EV Fluence, and Kangoo electric. On their specs page (French) I found they use motor of 130 kg developing about 50 kW. Big difference in power densities between those motors (I think made by Continental, synchronous AC with slip rings), and the one from Tesla roadster. @HarveyD, How much is that in metric? About 7.3 kW/kg. Some twenty months ago, GE declared, to highly positive media coverage, they would purchase 25,000 EV's(~6,000/year: an average 500+/month) by 2015. Have over 10,000 EVs been bought or has this followed their trend in tax payments? Sorry for being lazy Anne. Here are the conversions: Assumimg that 1.0 hp = 0.746 Kw and 1 Kg = 2.2 lbs 1. 3 hp/lb = about 4.9236 Kw/Kg 2. 5 hp/lb = about 8.2062 Kw/Kg 3. 4.5 hp/lb = about 7.38738 Kw/Kg Anne: I fully agree with the Metric system with one exception. Could you convince EU to do like Japan and state fuel consumption in Km/l instead of the odd l/100Km Harvey, in Europe another HP is used for vehicle power: http://www.mazda.de/car-configurator/#/configurator/m3/ ... Mazda3 ... den 2.0 l MZR 110 kW (150 PS) oder den 1.6 l MZR-CD 85 kW (115 PS)! and: http://www.mazda.co.uk/showroom/mazda3/specs/ 1598cc in-line 4-cylinder, 16-valve, DOHC, 105ps (77kW) at 6000rpm and 145Nm at 4000rpm There 1 HP = 0.733 kW (1 kw ~ 1.36 HP or PS) See: http://www.viewamerica.net/art/en-horsepower.html and Wiki - Horsepower Another thing - l/100Km or Km/l (or mpg). Usually all consumer goods prices are expressed as say$/lb, $/kg,$/oz, $/bbl, and not kg/$.
European "l/100Km" mostly follows that pattern, ie how many litres to buy for 100 km.
American mpg reminds of vehicle range, like miles per tank. Probably comes from the time when fuel was ridiculously cheap in US, cheaper than drinking water. It used to be "Fill'er up". When fuel is expensive, you may fill just half, if driving infrequently.

Perhaps now that fuel is not that cheap in US, it makes more sense to use quarts per 100 miles, than mpg. It could make it easier to calculate how much fuel you have to buy.

MG...I fully agree that Kw should be used instead of various values of HP but I still prefer the Japanese metric way to state fuel consumption.

I wonder how bad their previous motors were that they could boost their efficiency by 5%?

Likely, the motors were about as good as could be achieved with the e.g. length of flux paths and winding sizes required by the larger size.

-MG I agree, let's see those numbers. Otherwise I too, am not impressed.

Regarding this motor - its maximum rpm is 14000 rpm. Or perhaps more accurately, still only 14,000 rpm.

Should you want to find an AC permanent magnet motor exceeding 14,000 rpm today, then look no further than the Prius in the back lot of your local Toyota dealer. Under the hood you will find not one but two similar machines, one in particular is rated 60Kw.

So, certainly no earth-shaking mechanical improvement there. Anyway the aircraft industry have been using 12000rpm 400Hz motors for years. They are not foreign to GM either. Their first experimental AC powered car, the Electrovair II used a motor which maxxed out at 13,200 rpm and was also liquid cooled and this was way back in 1966.
Some 40yrs later the Tesla Roadster has to manage with a motor that barely exceeds 13,000rpm also. I would point out that since these motors are coupled to single gear ratio reducers, faster would be better.

So GE's research comes up empty in the one parameter that certainly would be of interest.

Let's discuss efficiency. I have put forward that the copper loss i.e. the electrical loss in the stator wiring is constant per unit torque. This holds whether the motor does 10 rpm or 10,000 rpm. If you accept that premise then it follows that efficiency improves with rpm.

For example a 1500 rpm machine having 92% efficiency.
It must therefore have a loss of 8%, you would agree ?

Now let's increase both input frequency and voltage, and speed that machine to 12000 rpm while still providing the same torque and drawing the same current.

At the finish the motor will be putting out eight times the mechanical power while incurring the exact same copper loss it did at the start. The same current remember, the current itself was not changed and remained constant.

So if we have the original electrical loss accompanied by an increase of output power of eight times, this can only mean that the effective efficiency must have increased to a figure of 99%.

But what about the iron loss in the stator shouldn't that increase big time ? Well, No. Build the stator with thinner laminations - like those found in 400Hz aircraft motors - and problem solved.

Next steps include a four-year project, during which time GE engineers will seek to produce a motor with similar performance characteristics, yet with no rare-earth magnets.

The problem with rare-earth magnets is that they are temperature sensitive. If the motor is allowed to get too warm they begin to lose their magnetic properties. Hence the necessary liquid cooling.

The other disavantage with magnets is the fact that they are always "ON", so when coasting in an electric vehicle there is always going to be some drag from that previously mentioned iron loss despite the use of thinner iron laminations. This braking effect offsets their very slightly higher efficiency obtained when motoring.

Ordinary induction motors can run safely at higher temperatures and may obviate the necessity for liquid cooling in some applications. Perhaps GE is going to investigate them next.

Look - in case any entrepreneur happens to be reading this - what the industry or the aftermarket could probably use is a SRM or induction motor design with a 3 : 1 planetary reducer built into the end bell of the motor which can be coupled into any front wheel drive vehicle out there, after removing the clutch and gearbox. Engaging the final differential gives the overall 10 : 1 ratio that a hi-speed machine needs, since most auto gearboxes will not accept anything like 12000rpm in their 1st and 2nd gears.

Now let's increase both input frequency and voltage, and speed that machine to 12000 rpm while still providing the same torque and drawing the same current.

At the finish the motor will be putting out eight times the mechanical power...

No.  You've only doubled the input power (voltage * current); you cannot get 8x the output when losses were only 8% to begin with.

At the same stator current and slip frequency (difference between drive frequency and rotational speed), the rotor current of an induction motor will be the same.  The back EMF reflected to the stator is proportional to the rotational speed of the rotor field, or the rotor speed minus the slip speed.  If you were operating at 4% slip and double the drive frequency while keeping the current and slip the same, you'll go down to 2% slip and increase the back EMF by factor of about 2.04 (slightly more than doubling).  The mechanical power output will increase by about the same factor.

I can quote chapter and verse from induction motor fundamentals, but it's a lot of work to write out in limited HTML so I'd suggest you just look up some on-line coursework instead.

what about the iron loss in the stator shouldn't that increase big time ? Well, No. Build the stator with thinner laminations - like those found in 400Hz aircraft motors - and problem solved.
You spend more volume on insulation, so the net flux density (and motor power density) falls.  As frequencies rise still further, you've got to deal with skin effect in the stator conductors; there's a tradeoff for everything.

Entrepreneurs in this field worth their seed money either are electrical engineers or employ at least one who's worth his salt (or stock options).  While you might come up with some remarkable new way of making an induction motor which achieves spectacular results by operating in a regime that wasn't practical before, you're not going to change the underlying physics.

-E-P thanks for responding. Perhaps I can convince you.
It's times like this a telephone would come in handy !

I wrote:
"Now let's increase both input frequency and voltage, and speed that machine to 12000 rpm while still providing the same torque and drawing the same current.
At the finish the motor will be putting out eight times the mechanical power..."

You responded:
No. You've only doubled the input power (voltage * current); you cannot get 8x the output when losses were only 8% to begin with.

In my example I was multiplying by eight not two. I should have been more explicit. Normally we rewind for a lower V/Hz for EV work so that voltages stay within reason. For our immediate purposes we could assume this to be a 100Vac motor @ 50Hz going to 800Vac @ 400Hz.

An eight fold increase in rpm and voltage, with current and torque held constant.
But You are correct that if I began with 8% loss, then
I really started with 92% output not 100% !

At 400Hz it will be absorbing eight times the electrical energy and outputting 800%-8% =792%

So the hypothetical power ratio is 792/92 = 8.6 not 8.0.
So far, no one has disputed that 792/800 = 99% overall figure however.
You are the first to point out that initial error in my argument however. Thanks.

In the past people that I've run my argument past say they are just uncomfortable that running a motor faster under certain conditions can improve the efficiency of the machine. It kind of goes against the laws of nature and smells of the fictious "free lunch". But to continue.....

Accepting that copper loss is independent of rpm is one thing, but of course no-one with integrity could neglect the effects of the iron loss which are expected initially to be equal to the copper loss at rated power. They would naturally get worse, hysterisis at frequency and eddy currents at frequency squared. In actuality it appears that you don't get anywhere near all the iron loss that text books would have you believe. Sure car alternators get very lossy if you keep constant field and rev the engine. But look at the lamination's thickness. When the bean counters are watching your back you get the thickest and therefore cheapest lams money can buy that still allow you to make spec. But that's in the auto biz.

Industrial motor design is usually to a higher order due to IEC regs and with the newer high efficiency motors even more so. So eight times the hysterisis iron loss works out to be eight times practically nothing. Eddy currents appear to follow a 0.7 exponential law in real life, with the 4 times frequency testing, done privately, so far.

Assuming constant current and torque as before, then inputting 4 times the frequency and receiving four times the rated power is one thing, in practice only a 4^0.7 = 2.6 continuous rating is useable however. In other words motor case temperature remains the same, admittedly with the fan working harder, if you pull only 2.6 times and not 4.
The transient power needed for vehicle acceleration remains available at a true 4 times. And there is still a 3.5 times current increase over that to pullout, but then you're buying into 10 times the copper loss. For the few seconds needed it may be cost effective, if the battery can bear it that is.

Other losses including friction and windage of the rotor are minimal, although removal of the integral fan and installation of an external blower, temperature controlled, has been suggested. I mainly wish to make the point to others that high rpm motors will always show better efficiency figures by virtue of their rpm and it is not something to get too excited over.

That said this GE motor is mighty small for a 30Kw continuous output, too bad it is made out of unobtainium for the rest of us. This is the size of motor they should option for the Nissan Leaf for use in the inner city with gearing for 100Km/hr rather than the 150Km/hr 80Kw it currently has.

I omitted to discuss slip strategies here - but it would take more pages than the moderators would want. Suffice it to say that you describe slip as a %age but I was seeing it as a fixed value for constant torque, independent of rpm. The back emf would be in proportion to the velocity of the rotating flux. Beyond that I haven't given slip too much thought since the modern idea is to keep the stator further away from voltage saturation by employing inverters with much higher current capability and even lower V/Hz motor windings, which allow base speed to move higher up within the total speed range, than would have been done in the past. Optimising for the motor rather than the inverter, if you will. Concern over slip has taken a back seat but I will consider the ramifications I might be overlooking so thanks for that.

Sorry, I mis-read your comment.  I partially read it, deferred a response until later, and jumped into a reply on another partial reading without ever digesting the whole thing.  Several errors resulted.  Mea culpa.  On the other hand...

At 400Hz it will be absorbing eight times the electrical energy and outputting 800%-8% =792%
Almost.  The 792% output needs to be compared to the original (100%-8%)=92% output.  7.92/0.92=8.61.  From this must be subtracted increased iron losses, greater winding resistance due to skin effect, etc.

I'm waiting for some radical advance in motor design to hit the market.  Carbonyl-process colloidal iron composite stator cores with fiber reinforcement, injection-molded at line production speeds.  Extruded copper tape windings with integral cooling channels.  Something that makes a motor able to operate in the dozens or hundreds of kHz range with losses equal or less than today's 60 Hz units and specific torque able to snap hardened steel driveshafts.  Something that goes inside a wheel and weighs less than the iron brake disc and caliper it replaces, but yields 50 kW peak or more.

When the bean counters are watching your back you get the thickest and therefore cheapest lams money can buy that still allow you to make spec. But that's in the auto biz.
Having been in the auto biz and obsessed over a tenth of a cent per unit (and saved \$2/unit at least once), I know whereof you speak.  I'd like to see radical advances rather than incremental ones.  This will come from minds like Tesla, looking at the equations of Maxwell that everyone had looked at before and seeing possibilities that no one had seen before.

I'm not sure I follow you on the other stuff, but maybe you should elucidate.

I'm not sure I follow you on the other stuff, but maybe you should elucidate.

I would prefer to take technical discussions over to AEVA or DIYelectric car where you can still locate the thread or topic months or years later. Unlike say,on GCC.

For AEVA, "How to convert a hybrid" has some thoughts on base speed placement. On the motor controllers forum, I've just started a "14000rpm motor" topic where I intend to discuss the narrow interest of induction motor rewinds for fractional V/Hz and single stage reducers. Perhaps later, if you have the time that is.

T2 and E-P agree more than they admit.

Future mass produced e-motors, with much higher power density, will be fitted in or near each wheel and will provide all the controlled traction and energy recovery required at lower (total) cost than current heavy e-motors with or without transmission.

One can easily guess where those improved, lighter. lower cost motor wheels will be built.

@T2,
Sorry I didn't see your post. I agree with your analysis of efficiency and advantages of high speed e-motors.
I think that motor MG2 of Toyota Prius (latest PSD - 2010 model release) doesn't exceed 14,000 rpm. I have two numbers for max rpm of MG2: 13,500 and 12,400 rpm. MG1 in previous versions had about the same max rpm as MG2.

Bosch has two (scalable with length) versions of their so called SMG (PM based) motors for hybrids, with 180 and 138 mm diameter. The 180 mm one has max rpm of 14,000 rpm, and the 138 mm one has Max rpm = 18,000 rpm. Possibly the reason for lower max rpm of larger diameter motor is higher moment of inertia (goes with square of radius), which causes more bearing load. Higher rpm also have some rotor components retention issues.

I think that for very high rpm motors (13,000 rpm and up) it's preferable that L/D ratio is >1, preferably up to 2 (? cooling issues in the centre of rotor). Tesla roadster's motor looks a bit elongated. Perhaps in elongated motors, rotor cooling at the centre is more difficult in PM rotors than in cage copper rotors.

Regarding coasting (aka sailing/gliding) with PM motors - recently released hybrids from BMW and Mercedes use it (turn of the engine) when gas pedal is released at speeds of up to 150 kmph (forgot the exact number). They both use ZF pre-transmission PM motor, now a 40 kW one.
If they somehow can disconnect coils, then the drag torque would only be caused by currents induced in laminations - probably much smaller drag than from rolling and air resistance.

What do you think of using TWO PHASE IM (TPIM) motor, four pole, instead of 3 phase one (also four pole)?
Found this for motor design: "Criteria also required a minimum number of poles to minimize stator losses.". Wouldn't it then favor 2 over 3 phases ?
It would require less space at the stator circumference to put all the coils, and smaller radius allows higher rpm.

Switched reluctance motors are cheaper and more efficient than most others and need no rare earths. ..HG..

The comments to this entry are closed.