Inverter duty motor insulation

[thinker]

The VFD system (USA installation) has PWM Active Front End (AFE) and two motor inverters fed from the common DC bus. Inverter 1 is connected to 575V motor, therefore, the AFE input voltage is 575V (after step-up drive isolation transformer 480V/575V). The DC bus voltage is 900VDC. The Inverter 2 is connected to 460V motor (to be built in EU as US version). The motor 2 supplier states that since the DC bus voltage is 900V, the motor insulation (as per standard) should withstand peak voltage 1800V (2 times DC bus voltage). To achieve that, they plan to enhance the existing insulation, and de-rate the motor (thicker insulation-reduced winding copper cross-section to fit the slot-reduced torque/power). We argue, that Inverter 2 will be set up for 480V, so the motor 2 terminal voltage will be 480V and not 575V. Insulation standards are based on terminal voltage value. In our opinion, motor insulation will see increased DC bus potentials to ground, so only insulation to ground should be enhanced and not phase-to-phase and turn-to-turn insulation. Can anybody comment whether we think correctly?

 

[jraef]

He is right. The "480V" programmed output of a VFD is just an RMS value we assign to what the PWM pattern makes the motor react to correctly. The actual pulse voltage will be at the DC bus voltage level and that is what the insulation must be based on. If you think about it, a 480V RMS system is ALWAYS at the same peak voltage of what a 480V VFD would have on the DC bus, so the same will follow through on a 600V class system. A 600VAC RMS system is actually 846V peak.

"Will work for (the memory of) salami"

 

[Skogsgurra]

You need better insulation. Period.

The fact that motor voltage has been set to 480 V doesn't change the DC link voltage and neither the PWM edges. They will still reach around 2*900 V if your cables are of any length (reflections at the high impedance at the motor end causes this).

Let your supplier have it his way. You will be in a really weak situation if you force him to use standard insulation. Not only will you risk the longevity of the motor - you will also void the guarantee.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[thinker]

Thank you very much, Jraef and Skogsgurra, I will follow your advices.

 

[John2025]

How much, if any, difference would a load reactor make? Are the pulses short enough to be attenuated or do I misunderstand completely (not the first time)? I'll admit I don't really understand the reflections.

 

[mikekilroy]

John, you are correct; load reactor would drop the Vpeaks to or less than the lower voltage equiv.

But, gosh, the reactor would probably cost same as just buying the right motor and KISS.

http://www.KilroyWasHere<dot>com

 

[thinker]

If the cost of the reactor is not the issue, would it be reasonable to use such reactor at the inverter output, and NOT to enhance the motor insulation due to higher DC bus voltage?

 

[mikekilroy]

yes.

Example how much it cuts output spikes (the issue):

reactor.jpg(72.5 KB)

https://www.box.com/shared/05khulpaza4t8ou1y10t

http://www.KilroyWasHere<dot>com