motor overheating with VFD, and incorrect amp readings

[cokeguy]

We have a 350 HP fan driven by a VFD, the drive (saftronics GP10) has input 3% line reactor and DC choke, no reactor at the output. The motor always runs overheated, even though it has been running at 60 Hz for the past year or so. We want to try placing a reactor at the drive's output rather than at the input, to see if harmonics are causing this, but the electrical schematic states that the reactor should go either before or after the VFD, not in both. Is there any particular reason why you should't use reactors at both sides of a 350 HP drive? We have other similar but smaller drives with reactors at both sides, with no problems.

Something that dazzles us is that the drive's display always indicates a current 80-100 amps lower than the clamp-on true RMS ammeters we use. We have 3 identical 350 HP fan/vfd combinations close to each other, and on the 3 of them we have this difference in readings. We use GP10 drives extensively in other parts of the plant, 200 Hp or smaller, and the drive's display always agrees with the clamp-on ammeter.

By the way, motor to VFD distance is about 140 meters.

Any comments? we just changed a motor last week, and the new one is still overheating.

 

[DickDV]

First, it is a waste of time to measure anything on the motor leads. Even true rms meters will not read accurately. In addition, it is very likely that the voltages far exceed CAT III levels so you and your meter are at risk. Believe the drive display.

Second, while reactors are nice cheap harmonic filters, they also drop voltage. For example, a 5% reactor drops 5% of its input voltage at rated current. Two 5% reactors will drop 10% of your supply voltage at full speed and load so it is likely that your motor would be starved for voltage with one input and one output reactor.

Third, you have rather long leads and should investigate some ringing and reflected wave protection for the motor. Even if the motor is inverter rated with excellent insulation, the insulation is probably in trouble.

Finally, are you sure the motor is actually too hot. Class F insulation is good for 110 degrees C at the insulation which would be about 100 degrees C on the exterior of the motor frame. Are you measuring the voltage with temp crayons or a heat meter?

 

[cokeguy]

Thanks for your comments. Frame temperature is measured either with thermal imaging camera or raytek type infrared thermometers, and temperature normally hovers somewhere around above 95 degC, much hotter than all the rest of the similar sized motors, either with VFD or DOL driven, which rarely run above 60 degC.

Thanks for your comments about the ammeter and reactors, we haven't taken that into account.

I understand that ringing and reflected wave prevention will reduce insulation failures, but I don't think it will drastically reduce temperature levels, right?

 

[ScottyUK]

Hi Dick,

IEC 60085 defines Class F as good for 155°C at the insulation, which might well equate to 130°C at the frame, perhaps even higher if there is good thermal contact.

A decent design shouldn't push the insulation right up to the theoretical maximum - many big machines are built using Class F materials but run at Class B limits (130°C) which gives enough margin to ensure a long service life and allows for the odd hot-spot.

Spot on comments about about exceeding the capability of Cat III protection on the instruments. So many people don't even consider the capability of their test gear.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[edison123]

Since you're running at full speed all the time, you could try running the motor without VFD or at least bypassing the VFD once the motor gets to full speed. This would confirm/eliminate the VFD as the villain.

Muthu

http://www.edison.co.in

 

[DickDV]

The information on insulation class temps that I had came from MG1 Part1 where Class B is 85 degrees C, F is 110 C, and H is 135 C.

Now I'm curious about the differences between IEC 60085 and NEMA.

Can anyone explain the differences? Because, as ScottkyUK points out, that's a significantly large difference.

cokeguy, I don't think that reflected waves and ringing will affect the motor temp much. The temp question is separate.

 

[LionelHutz]

With a DC choke you could probably move the line reactor to the load. However, that may not fix anything.

Double check the motor setup parameters. For example, make sure the motor voltage is set to the motor voltage (460V) and not the line utilization voltage (480V).

A good solution would be to get rid of the VFD. When you only run at 60Hz you are throwing away about 15kWh to 20kWh of electricity for every hour one of those fans is running.

 

[DickDV]

It's my mistake on the insulation class ratings. The numbers I had were rise only. Add 40 degrees C ambient to them to get total temp and they math ScottyUK's.

I would say the motor is not overheating based on the numbers measured at the outside frame surface.

 

[cokeguy]

Thanks for your comment. I really can't get rid of the VFD, because we do need to modulate fan speed in certain situations. We will install an inverter duty motor to diminish the probability of premature motor failure, but I will still try to find out why these guy heats up much more than its neighbors and keep you posted.

 

[ScottyUK]

Perhaps just because it's driving a larger fraction of its rated load? The others which run cool sound like they are over-sized for the steady-state load: maybe sized for peak load or for acceleration capability, or maybe just bigger than neccessary.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[MAGTiger]

The voltage dropped between reactor terminals may well be 5% but the drop from source to drive (ignorring line drop) will be far less. The voltage drop at the deactor terms is mainly reactive and subtracts from source voltage vectorially. It will depend somewhat on power factor but at drive input the PF should be above .9 however on the output of the drive the pf will be whatever the motor wants it to be. dv/dt can cause motor heating but it sounds like your heating problem isn't really a problem based on the ratings and measurements.
Depending on the cable length to the motor is is posssibel for the voltage in the cable conductors to exceed 1500v assuming you rated volts are 480vrms

Neil

 

[jraef]

I think LionelHutz brings up a good point, make sure the motor parameters are properly set in the VFD to match the MOTOR NAMEPLATE, not the line voltage. It's a very common commissioning error, and what happens is that the motor windings end up over excited and consuming more current than they should while producing no more work, essentially lowering the efficiency of the system and burning off the excess energy as heat in the motor.

"If I had eight hours to chop down a tree, I'd spend six sharpening my axe." -- Abraham Lincoln
For the best use of Eng-Tips, please click here -> faq731-376

 

[mash98]

may be you have forgotten to auto tune the inverter which usually cause drawing large currents.

 

[krsh]

Beside checking the VFD parameters also check the mechanics of the fan(bearings, propeller, alignment). An error on this side can also trigger the same symptoms.

 

[mash98]

a very valid point by Krsh.

 

[macmckim]

Another possiblity is the motor - based on the thread info, looks like the motor (s) have been replaced in past few years.

If client purchased a high efficiency or premium efficiency design typically the motor's full load speed increases by 5 to 10 RPM or more depending on how old the original motor were. IE: 1750 RPM (old) and 1770 - 1780 (new)
Application is a fan - so the extra 10 -- 20RPM increases the required horsepower (cube rule for fans?)