Running AC Motors Over Frequency with VFDs

A couple of thoughts :

1) You could get your motor rewound to operate at a lower voltage. I know a rewiding specialist who assures me that rewinding for 80-100V operation is possible without losing output power.

2) By embedding thermistors in the windings when you rewind the motor you can close-monitor the winding temperature. This will allow you to gun the motor a lot harder for a brief burst, and the controller reads the thermistors and backs the motor drive off to keep the motor from overheating.

Best of luck with your project.

Thanks Paulusgnome.

I am limited a bit bt the controller and batteries so I don't want to have to provide more than 90 Amps motor current if possible. That gives me best acceleration at around 360/200 volta and switching star/delta at about 40k/h.

Sorry I can't help...but it's a well written Q, and we'd love to hear more about the project as it progresses.
I know SFA about big motors
Good luck
Del

Thanks Del. Gosh I was expecting to be flamed something awful for mistreating a motor. My previous questions regarding Lenze VFDs was solved by getting a bigger one when it came up on eBay.

I'll post my progress.

You have a good understanding of the problem, but the solution may or may not be the one you are looking for.

Forgive me in advance for not wanting to bother with conversion of HP and ft-lbs to kW and Nm, it's too early in the morning for that... so I'm not going to give you specific values of torque.

You want an 11kW motor because of the physical size. But you NEED a specific amount of torque and RPMs. What you are asking is if a VFD can perform some sort of magic to make your 11kW motor deliver more than 11kW!

The answer is actually, yes it can, and MAYBE it is enough! You hit on it with the Lenze thing. That is a well known "trick" to VFD applications; configuring a motor for a lower voltage and running it at a higher one. So if you use a dual voltage motor, you connect the motor for the lower voltage (Delta in your case) and you can run it beyond it's base frequency by connecting the VFD to the higher voltage source so that you don't run out of voltage when you exceed base speed. You need to keep the V/Hz ratio constant to avoid over or under fluxing the motor, but beyond that, everything is possible. So as you noticed, the motor power can increase right up to where you again run out of voltage. So in the case of a 50Hz motor run at a voltage that is 1.732 x the lower connection voltage, that means you can increase the kW output right up to 87Hz. Applying that then to your 11kW motor, it means that at 87Hz, your motor can deliver almost 15kW.

So now what you need to do is work the math backwards. Determine your peak torque requirements at whatever speed is involved, and solve for kW at that point. Then you know if your motor is going to be capable of it. Keep in mind also that when you define that PEAK requirement, you can also look at TIME at that level. A good quality Vector VFD can make the motor deliver 200% of rated torque for maybe 30 or 60 seconds, 150% usually for 3 minutes. So if the peak output you are looking for is only for acceleration, you might be able to do that occasionally. But as was mentioned earlier, it is a GREAT idea to make sure you use embedded thermistors or RTDs in the motor to monitor actual motor temperature, especially if you are going to run it into the margins like that. It's also imperative that you have a good cooling system for those motors as well because if they rely upon integral fans, the fan performance curves are designed around normal operating speed, so they under perform when above or below that narrow band.

By the way, don't you have 400/240V motors available there in Oz? I know they do in NZ. I kmow a good source for motors and VFDs for you in NZ if you are looking, and in fact I think he has a salesman in NSW now as well.

Thanks for your post JRaef. Yes we do have 240/415 motors but they are only available in <= 3kW. Over 3kW motors here are 415/720.

The VFD will not be the limiting factor as it is a 30kW with 150% overload for 60 seconds. I have lots of calculations that give me the power requirements and maintaining vehicle speed will not be an issue. It's gradient climb ability and acceleration that I'm trying to pin down and this centers around what peak torque the motor can actually put out if it is rewound for higher frequency operation.

The caution I have been given is that I will exceeding the frequency of the original motor design and will have less-than-calculated torque above 70Hz. The same person who gave me this advice readily explained that it is largely unexplored territory - rewinding ACIMs for higher frequency operation.

How big a car do you intend to fit this motor into? How much current your VFD can handle?

I found that nema does make good motors but they're far too small for for an applications like this.

(I wanted to use one 4go-kart but, not even 4 that I could find them to be powerful enough with some speed.)

Do you intend to rewind a normal three-phase motor? It's a very bad idea nd you'd better dwell on it until something good comes about that's more suitable.

I hate to disappoint you but unless you can lay your hands on a hybrid car gen-motor you'd better build one from scratch using permanent magnets in the outside rotor.

I do believe high voltage low current is a better choice but depends on the type of battery you'll get and try to get a hi-tec one that has high efficiency also!

There are 230V motors commonly available in Asia e.g. Japan and other markets, should be easily available in Aus I would think.

As mentioned, you can connect for higher voltage / frequency to increase power output, this should be oK at levels already outlined. Bear in mind your installation as heat loss will increase with power, of course, if the motor has shaft driven fan, cooling will increase as will power loss to this increased air flow. This can be avoided by replacing with separate cooling if desirable for audible or other reasons.

The performance above the notional max power will decrease as the inverse square ratio of max pullout torque, so a higher pullout motor will have better performance at higher frequencies. e.g. 230V 50Hz winding at 230V 100Hz operation. 240% pullout. Max output = (1/2)2 x 240% = 60%. Torque calcs at 50 % so inside peak which is OK (just)

I would recommend using a current limiter function for higher speeds to ensure motor does not stall in under-volted area of operation. If you have a sophisticated inverter it will calculate torque accurately in higher speeds, but many are prone to calc errors in this region. If you have good torque calcs, you can set and change torque limits at higher speeds to avoid motor stalling.

Motor stalling must be avoided at all times as you will lose control of vehicle and also the motor current, you must stay inside performance envelope for the VSD // Motor combination.

Sounds interesting, I've done similar on centrfuge type applications, 50Hz windings to 200Hz (if lucky) I find there's not much experience of motor control in this mode of operation, I can say you're better off with quality motor as a minimum, minimum of current limiter on inverter helps avoid stalling.

You can connect DC to most inverters but you don't say what supply you have, this will affect the overall system selection and performance.

More information.

The VFD will be powered by a 600 VDC battery pack. The pack will be broken into 60 V sub-packs when the ignition is off. (Battery charging etc is another topic.) At 600 VDC the VFD will produce 420 V phase to phase.

Since the standard motor in Australia is 415/720 Volts 50Hz, the motor will run out of v/f above 1500 RPM so the plan is to rewind it to change the rated speed at 415 volts to 118 Hz. hence the 300/173 volt rewind.

The car weighs 1050kg but will probably weigh in at 1250kg. The VFD is a 30kW Lenze 9329 with overload capacity to 44kW or 89 Amps. In the models I have set up the VFD is the limiting factor with this proposal.

If the motor maintains torque at 118Hz when rewound to 173 V Delta and driven by 415 Volts then we have success - but will it?

BTW The plan is to Star/Delta switch the motor at lower speeds to keep controller (VFD) current low.

The VFD has plenty of features to safely handle the motor - the safety or longevity of the motor is not the issue in this topic - just the projected power levels.

I changed my voltages since the first post to optimise acceleration to 80k/h.

what sort of armature are you going to have?

from the speed controller view the conventional type is just not gonna work well over a speed range you're looking at.

OK, you've got a good supply and you want high torque at maximum speed. In answer to your original query, in all my experience, peak torque at base Hz is less than at lower Hz - I think you can count on 200% - 250% at best based on good motors and 50Hz data given that this would develop 300%+ on lower frequencies in extremis.. I think its more realistic to design closer to 200% and try and find a 100Nm rated output. If I were to guess I estimate you'd lose 5% on rewinding too!

Why do you need such high torque at max when max is about overcoming frictional and windage losses. Isn't it more important lower down in speed range?

I don't see the need to change to star at lower frequencies, this would only reduce performance for little gain. The VSD should control to no load current at good efficiency anyway so I don't see the point.

Hmm, windage at top speed increases the power requirement with the cube(?) of the speed. (double the speed, double the drag, double the distance).

At low speeds you only have acceleration requirements.

Low speeds also have to accommodate climbing hills. Only a 3% gradient takes the power required at 60k/h from 6kw (flat and level) to 12kw. There will be a heavy requirement on having enough torque at low speed to get up say a 10% gradient.

Just musing:

I always find the "acceleration" vs "climbing" analogous, since horizontal acceleration is defined by F=ma, and in hill climbing "a" is a function of the gravitational constant. The net result is you either add kinetic energy or potential energy to the system. Power required depends on how fast you want to change that energy level. In your system, the speed-torque requirement is challenging, in that it could have a "smiley" face, lots of low end torque required, and at top speeds lots of high end torque. On the flat and level the low end torque is short lived, at the high end it is usually sustained. Now, if you are mountain climbing, the torque and power requirements become sustained. (I love my F350 turbo diesel for hauling loads up mountains!) Unfortunately, no electric motor matches all these characteristics and compromises and gearing etc are required to have an acceptable envelope of operation. (Mountain climbing without adequate power is going to be slow!)

Most designs are optimised for the commonest requirements which could be low and medium to high speeds where acceleration may be needed more often. Top speed does not need 200% torque or else it wouldn't be top speed!! (assuming the motor cannot endure 200% continuously)

A design geared to use an asynchronous motor 'field wekened' area would give same top speed continous torque availability but much higher torque at lower speeds - hence my query re 200% torque at top speed.

Selecting the correct gearing and using constant power region of operation is often practical and very useful in such a project - the idea of increasing torque at higher speeds may be OK to overcome drag but doesn't fit standard motor performance allowable.

I understand that there is a concern that the motor is being "pushed". In an EV environment where battery power is limited, the motor will not be called upon to perform continuously as with an industrial situation.

Using a smaller frame and rewinding it to get higher power should be OK provided adequate coling is provided. This of course will be hard to avoid when running at high speed (90-100k/h) which is the only time the motor will be used at high power for around 5 minutes straight.

All motor drives I have used are capable of producing overload capabilities at "top Speed". The Top speed is usually fixed by supply frequency, or max safe operating speed, or commutation limits, or other mechanical or electrical limitations. I have never encountered top speed as being defined when it can no longer produce adequate torque. This is for both AC and DC drives.

The vehicle operation at low speeds lends itself to a series wound DC machine, however, at top speeds in the constant HP range it runs out of ooomph. But these days no one likes to use DC. The machine is expensive and high maintenance.

The AC drive with careful motor selection can give some outstanding performace. A base speed of x rpm at Y frequency with a VFD is no longer a constraint. The motor can be wound to give proper top speed and torque and the frequency is set to accomadate it. In certain industrial situations we have motors running at 15 Hz and rated volts, or 86 Hz at rated volts, or whatever the application demands. A fixed V/Hz curve is no longer a constraint.

I would say the constraint outlined is anticipated performance and thermal capacity of a 132 frame motor rated at 50Hz either used at 87Hz operation or rewound (to 118Hz). The Top speed torque is being defined as 200% of the rewound rating which I think is pushing the envelope (hopefully, not too much).

Rewinding can itself take away some machine characteristics by a few percent, overload performance at rated speed is not so good as at medium frequencies in overload and the motor collapse can be more dramatic if stalled, but my main concern would be thermal capacity given that the original design is 11kW, connected for 18kW then run at 33kW for minutes at a time in high overload. My impression is only that if one could find a frame higher power in the first instance there would be a better chance of success.

Ongoing studies of efficiency and performance in such situations is thin on the ground, I've inferred some data from extensive testing at rated speed and others from a Japanese manufacturer I have detailed information from plus my field experience. This covers many manufacturers' inverters. Most high overload experience I have is from lifting applications where inverter / motor performance during acceleration is a key selection component. This application sounds more similar to a high inertia application like a fan, but this is not normally overloaded so dramatically so I confess ignorance here!

I have many years of experience of many standard and special design motors, as you say, but nevertheless, one can only estimate what may happen with such a selection as, unless you've done it already, its not so easy to predict success with so many changed parameters.

In brief, this is a motor characteristic issue, not a drive one as you're going to be at full voltage pretty much whatever and the inverter is current oversized.

I don't foresee any other problems from my experience, as this size motor is generally well OK for 2 pole 60Hz mechanical speeds apart from the cooloing fan.

The application cannot outline all parameters and as such one can outline experience and perceived pros and cons and, hopefully, this gives Jonny some other insights into his project having all other details to hand.

Hello,

It is possible to get more power out of the motor like JRaef explained. In fact we service motors and we see a lot of motors +/- IEC size 80 that give an output about 7.5 kW. These motors are running at 400 Hz 3 faze 400Volts.

Important is to rewind the motor with a double layer winding, this gives less higher harmonic currents for the frequency drive.

Keep in mind that running an 11 kW, 4 pole motor at 118 Hz its speeds is about 3540 RPM,

• so the rotor has to be dynamical balanced for this speed
• the motor output power will be +/- 26 kW
• so you need a frequency drive capacity for a 30 kW motor.

In the past, we did rewind electrical motors for a voltage as low as 24/42 V delta/star 50Hz 3 phase. So this is not the problem.

The problems I see are

1) how to increase your battery supply voltage to lets say 200 x sqrt2 = 283 Volt ?

• a lot of batteries in series ? ( +/- 27 pieces of 12 volt )
• a step up converter ? ( you will need a convert for +/- 40 kVA )

2) if you rewind you motor for a much smaller voltage 100/173 Volt delta/star

• the current will go linear up, so your frequency drive must supply this current and will be the expensive part.

Why not using two smaller motors and two frequency drives? One for each wheel?

Control of two independent motors is likely to be a software challenge. Tesla Motors spent a lot of engineering time on such issues.

Suggest a sun/planet speed reduction gear set with 3 motors driving 3 planet gears, all operating at the same commanded speed. Gives you the conventional drive train and with overriding clutches reduncancy for get-home if one or two motors or controllers fail.

Just a thought.

IF, you can maintain the volts per Hz to 118Hz, then the nominal NP torque will be available up to the higher frequency. (Yes, you do get more power from the same frame size.)

However, if you cannot maintain the V/Hz, the max torque (pull out torque) will be limited to approximately 1/(118Hz/rated Hz)^2 of the rated pull out torque.

Loose rules of thumb, but it will be in the ball park.

Thanks GW. Yes I will be able to maintain V/Hz. it sounds like I will get rated torque at the higher speed. Thanks.

The neat thing about your application is standard VFD drives running from a DC bus can automatically re-charge those batteries while braking. (But be sure to have your DC bus bypass the standard rectifier front end.) Just use the mechanical brakes for safety back up and parking.

We used to actually use a very large capacitor bank for absorbing the energy from low inertia loads during stopping and starting (press feeder application, about 100HP) and thus reduce the peak demand from the supply system.

The other interesting thing you will observe is at low speeds and high torque, the demand from your DC bus will be minimal, since it only supplies the real power, and not the reactive power for the magnetization current. (Neglecting I^2 R losses that should be small.) The highest current draw will be at high speed with high torque (ie high power).

Good luck, and keep us informed with the results.

If you give us the type of winding for the motor in 50 Hz,

• Diameter of rotor (+/- 120 mm ?)
• Lenght of core (+/- 100 mm ?)
• number of slots (36 ?)
• one or two layer winding (1 ?)
• coil step (1-8-10-12 ?)
• number of wires per slot (+/- 30)
• diameter of cupper wire (+/- 0.80 mm)
• connection parallel or 2 coils in serie ( 2 coils parallel ?)
• actual voltage 400/690V delta/star 50Hz
• new voltage 200/360V delta/star 118Hz

I can give you an idee about the winding for 118 Hz

That's great Rudy but I don't actually have the motor and the manufacturer doesn't give those details. I am ordering the motor and having it shipped directly to a rewinder. I will be speaking to him about the details you have given me so far. Thanks for the detailed response.

You can send me a mail if you need more information? good luck

I'm working on the same type of problem, but it is a motor an a boat.