VFD Failures 2

[Ceast]

One of my customers has recently suffered a rash of AB 480Vac freq drive failures. I've been collecting PQ data on the MCC Main. No anomalies captured to this point. Some of these drives have been in service for a numbers of years without failure.

The drives have been on the recieving end of flash-over and some have exploded. My first thought was the plant is subjected to some serious transients. They have no reports of problems on the 4160Vac equipment. Nothing. Seems the failures are confined to the 480 bus.

I have yet to investigate the status of the surge arresters on the 15kV system but there does not seem to be any arresters or TVSS on the 480 system. THe system is 12470/480 delta/wye with HRG on the 480Vac.

Anyone have any experince with flash over VFD failures?

 

[Skogsgurra]

Certainly have. But that was on a 690 V floating ground bus. Large and small VFDs, very few linear loads.

Nothing spectacular between phases. But phase-ground did have large and sudden transients that caused flash-over to ground.

There was a PFC capacitor bank that switched very often, totally unnecessary since cos(phi1)already was close to 1. After switching the automatic PFC relay off, we have not had any problems.

Another one. Emergency stop cut out the motor (opened breaker between VFD and motor) when running full speed. This was a 1200 kW VFD and motor. Very expensive practice. Used to blow the inverter every tenth - twentieth time. Changed that. No probs after.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[ScottyUK]

Is your logging instrument fast enough to capture high speed transients? Some destructive phemonena last only a few tens of microseconds but contain large amounts of energy. A lot of PQ instruments simply won't see them.

Is the system grounded properly? Have you taken earth electrode resistance measurements? Is there any data from a few years ago or from original build to compare against?


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

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

 

[Ceast]

I collect PQ data with a Dranetz PP 4300 which will capture transients. Would not micro-seconds would be Flicker? THe 4300 will not capture Flicker.

Transients in micro-seconds will cause splattered metal and carbon deposits? The damage looks like a high current ground fault! Metal deposits and discoloration.

I will conduct and extensive ground study next week. I'm somewhat suspicious of the grounding issue as well. The customer has done quite a bit of construction and has added a lot of new electrical gear. Maybe a ground loop was opened. This place has a huge ground system.

Part of the addtions has been 2 cap banks. The funny thing is none of the new gear has been energized.

Thanks

 

[Skogsgurra]

Flicker has nothing to do with it. That is a rather low and slow voltage change.

The 4300 has a problem with not having an adequate anti-alias filter so that it records even harmonics where there are none. But it surely catches energetic transients.

Submicrosecond transients will usually not be energetic enough to do anything. Not even rise the potential enough to cause a flashover. Capacitance in the system filters them out.

What make are the inverters? Might there be regeneration without proper configuration or hardware? That could also cause high DC link voltage. A lot more than one would think possible.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[jraef]

He said AB Gunnar, but I don't think that's germane, especially if they were working fine for a long time.

I have only experienced VFD flash overs twice in my career. First was a long time ago, it was the result of a poorly designed bulk PFC system. The transients it created took out almost all of the VFDs at the same time. AB and Toshiba, but the situation was so severe that I doubt anyone's VFD could have survived.

The other was with a brand called PDL from New Zealand (now owned by Schneider). They had a component problem on their front-ends where in the case of a resistance grounded system, a charge would build up on the chassis frame and eventually flash over to the bridge. I don't remember all the details because I only learned of it in having to replace all of the grounding boards, but it was fairly spectacular when it happened.

So from my experience, I would look for a grounding problem or a newly installed capacitor system that someone either forgot to mention, is unaware of, or have been testing when nobody was looking (hint hint).

Too many times I have seen people do some premature testing because the factory technician was called out too early, they do it after hours, blow something up, then claim they never energized it.


"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

 

[Skogsgurra]

Didn't see those two letters. Some companies need to change their marketing "Rockwell Automation" stands out a lot better.

The "we done nuttin" situation is interesting. Well worth being aware of. But probably not in this case. Been running for a few years, I gather?



Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[controlsdude]

Allen Bradley did have a problem with drives shorting out out of the box a couple of years ago. But this is not your problem. The AB rep claimed it was caused by all the welding that was going on around the equipment I was working on. Of course I was trying to start all the equipment when these people were welding. I have worked on other drives but most of them just open up on the dc bus, not short out. You got to love stuff that explodes on you.

Lightning strikes can cause drives and other equipement to go bad, not this issue?

Are these drives local to the motors or are they installed in a panel with the motors in the field?

Has anyone looked into power transients on the input to this panel?

As far as estop wiring for drives, if your going to estop and open the output from the drive to the motor, does the enable get pulled to the drive? Gunnar pointed this out.

I have seen where if you pull the power to the input to the drive during an estop or just stopping the system that powering the drive up and down does not help it and causes premature failure over time. Do you have a design that does this?

 

[LionelHutz]

Are you sure it's a flashover as the initial cause of the failures? A flashover will often occur if a component in the front-end of a drive fails badly.

The HRG on the 480V system could be the cause. Most drives really don't like operating on a high resistance or floating system, despite what manufacturers will usually tell you. Check for phase to ground protection circuits in the drives that could fail when the power system floats two phases to 480VAC line to ground.

 

[controlsdude]

This is out of the powerflex 70 manual

Unbalanced or Ungrounded Distribution Systems
If phase to ground voltage will exceed 125% of normal line to line
voltage or the supply system is ungrounded, refer to the PowerFlex
Reference Manual, publication PFLEX-RM001….

ATTENTION: PowerFlex 70 drives contain protective MOVs and
common mode capacitors that are referenced to ground. These devices
should be disconnected if the drive is installed on an ungrounded
distribution system. See page 1-11 for jumper locations.

see attached note from pdf

 

[LionelHutz]

Good info Controlsdude. AB is even saying don't do it in that manual. It could very likely be a MOV blowing up and causing the flash over.

 

[Skogsgurra]

Yes! That is probably what happened.

The question is why it happened to many drives after several years of interference-free operation?

I suspect there is something with the PFC capacitors. We probably had same thing in our case (the varistors had exploded, but we never understood that they probably also started the flash-over).

Good eye-opener Dude! PLS for you.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[LionelHutz]

Well, the obvious answer to the question of why now is that something is just now causing a ground fault on the system and shifting 2 line voltages so L-G = L-L on 2 phases instead of L-G = sqrt(3) x L-L.

I've seen some large voltage and current ringing when a capacitor was switched on and that could cause incremental damage to a MOV until it fails. However, the OP says the capacitors and other new gear is not energized yet.