Power factor correction unit totally blown up 2

[BigBill53]

Hi all,
I have a bit of an unusual one here, and I was wondering if anybody has come across something like it (or maybe a PFC expert could point me in the right direction).

I recently took a look at a power factor correction unit that had pretty much totally blown up. All of the capacitors have failed (either burst open or are short circuited). Some of the resistors on the contactors are burnt. In many places there are burn marks on the busbar - very strange for a 415V system where the busbar is spaced at around 40mm apart.

Some other relevant facts:
[ul]
[li]It wasn't a lightning strike - it was during the middle of the day and clear weather[/li]
[li]The harmonics on this site are practically zero - probably the only source of harmonics is the fluoro lights in the offices. Because of the remoteness of the site and the lack of competent electricians, the management stipulated that the entire plant be run on star/delta and autotransformer starters. All they have is electric motors.[/li]
[li]No other equipment was damaged at the time[/li]
[li]The PFC unit was second hand. The last owner used it for around 2 months and then it sat unused for 5 years[/li]
[li]The ambient temperature at the time was not excessive for this area (maybe 30 degrees celsius)[/li]
[/ul]

I have no idea what has happened here. The customer needs this to run their entire plant at the same time (their current transformer is too small) but I don't want to recommend they install another one if this is going to happen again. Any help would be greatly appreciated.

 

[sibeen]

Obviously having a guess here. Some of the capacitors that I've come across that have been used in PFC equipment are rather crap *technical term*.

As some of the capacitors have burst open, I wouldn't be surprised if a plasma was formed by the escaping gasses which caused the burn marks that you saw.

I take it that there was blocking inductors in this unit?

 

[jraef]

•The PFC unit was second hand. The last owner used it for around 2 months and then it sat unused for 5 years

Without a small mopunt of leakage current to maintain it, capacitors over time lose their film of oxide that covers the surface of the positive plates and acts as an insulator. When it's gone and you energize it, you get a dead short across the plate layers. After sitting unpowered for a year or more, they must be put though a procedure called "reforming" before being fully enegized. If you did not know to do that, then that is the likely cause of your failure. I have seen them literally explode on VFDs that people have bought on eBay for cheap, not knowing that they had been in storage for years.

"Dear future generations: Please accept our apologies. We were rolling drunk on petroleum."
— Kilgore Trout (via Kurt Vonnegut)
For the best use of Eng-Tips, please click here -> faq731-376

 

[Skogsgurra]

DC link, yes. But PFC? No.

Gunnar Englund

http://www.gke.org

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

 

[dpc]

If you have a lot of harmonic current, this can cause overloads in the capacitors. Nearby VFDs or other non-linear loads can be an issue for caps. Especially if they are not fused properly.

 

[jraef]

DC link, yes. But PFC? No.

Huh. I could have sworn I have seen it in the instruction manuals of PFC controllers. It's been a long time though. I don't like PFC controllers, I prefer at-load correction. Kind of makes sense though now that you mention it, there is no + or - side of a PFC cap...

"Dear future generations: Please accept our apologies. We were rolling drunk on petroleum."
— Kilgore Trout (via Kurt Vonnegut)
For the best use of Eng-Tips, please click here -> faq731-376

 

[jraef]

Well, I looked up the old project and mfr of the APFC system used on these induction furnaces I worked on, then dug up their manual on-line. I was right, it is mentioned, but they were referring to the SWITCHED MODE POWER SUPPLY that provided the control power! Funny how I had that morsel of misinformation stuck in my mind somewhere, crossed up with other morsels about VFD caps...

Unfortunately, I think I wrote the PM manual for that project and told them about capacitor reforming if they even have the system off line and in storage. Oh well, it wont likely hurt anything.

"Dear future generations: Please accept our apologies. We were rolling drunk on petroleum."
— Kilgore Trout (via Kurt Vonnegut)
For the best use of Eng-Tips, please click here -> faq731-376

 

[ausphil]

Large rodent remnants?

Ausphil

 

[LionelHutz]

Your description sounds like an overload occurred. Did someone measure the harmonics, specificially on the power going to the capacitor bank, when the bank was installed to ensure the capacitors were not running overloaded? Believing you have little for harmonics in the plant doesn't means anything if you hit close enough to a resonant frequency where some level of harmonics exist. Just because you don't have harmonic producing equipment doesn't mean that utility neighbours do not.

jraef - PFCC are film type and not electrolytic.

 

[Overvoltage]

That was my first thought, too, Lionel. In fact I've seen something similar in practice where a customer's PFC with harmonic filtering was tuned right into the resonant frequency of their feeder. Wasn't pretty. However, the OP stated this:

•The harmonics on this site are practically zero - probably the only source of harmonics is the fluoro lights in the offices. Because of the remoteness of the site and the lack of competent electricians, the management stipulated that the entire plant be run on star/delta and autotransformer starters. All they have is electric motors.

So, that tends to push me in the other direction if they're at a remote site (so likely few other harmonic generators nearby or even on the same circuit), and they don't have many harmonics of their own.

Interesting thread...and I'm at a loss.

 

[rcw retired EE]

OP says plant is all star-delta. Was the PFC equipment connected on the star or delta side?- It is probably connected on the LV side.

If the system is delta is it running ungrounded? Is the PFC system expecting to see a grounded wye system but was exposed to full phase-phase voltage across wye connected capcitors during a ground fault?

If the system is ungrounded, could ferroresonance come into play?

 

[BigBill53]

Hey,
Thanks for the comments.

I will check out the harmonics on the plant, but I would bet my first born child that it is almost zero. The entire plant only consists of electric motors, and they only have star/delta and autotransformer starters. The plant was fully running at the time the board tripped and power went out, so I'm fairly sure we can eliminate the starters. As for harmonics from neighbours, well, there aren't any neighbours. They are the only customer on that feeder (the other customer went bankrupt about 6 months ago).

The PFC unit is connected on the busbars of the incoming supply to the plant, doing correction for the whole installation. Pretty standard sort of stuff.

 

[LionelHutz]

The capacitor didn't just fail for no reason. You either got a bad capacitor or it was overloaded. Simply fixing or replacing the bank with another the same runs a high risk of a repeat failure.

Have you calculated the expected resonant frequency of the bank for each different switched combination of capacitance?

You can measure for harmonics on the system which might give some indications of what could be the problem but it won't show you what the capacitors were actually subjected to. It doesn't take much to overload the capacitors if the cap bank and power system were tuned to the right resonant frequency.

You may want to look at a de-tuned replacement bank if you can't figure out a likely cause of the failure. At the very least, look at higher voltage rated capacitors de-rated for the system voltage.

 

[waross]

Did this happen on the first energization or after some time of service?
For all the caps to burst, They must have all been online. Is this a reasonable assumption?
I suspect an over voltage on the feeder. A switching surge or other event that sent a high voltage transient down the feeder.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[ScottyUK]

Adding a capacitance to a long feeder can cause signficant voltage rise at the load end under the right (wrong?) circumstances. I think the overvoltage theory may have merit: as voltage rises the (lagging) magnetising current increases causing the PFC controller to add more capacitance to compensate, resulting in a voltage rise... add in a poorly configured PFC controller perhaps?

 

[waross]

Good point Scotty.
A further explanation:
On a long HV feeder, the voltage drop may be predominantly reactive. The source voltage may be high to compensate for the line voltage drop.
As the capacitors reduce the reactive component of the current the reactive component of the voltage drop decreases.
If the capacitor bank is poorly configured, the bank may supply VARs to the source. With the reactive current going towards the source the reactive component of the voltage drop becomes a voltage rise. In extreme cases the voltage at the load may exceed the supply voltage.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[dpc]

If the caps are on the primary, the harmonics could be coming from somewhere else. I would advise on checking the harmonics as close to the caps as practical. It would be good to rule this issue out.

 

[PHovnanian]

What does the utility feed look like? Is there any possibility that they opened one phase (a fuse on a fault or single phase switching)?

 

[ghostbuster7]

Many times we have found the contactors "machine gunning" to be the cause of the cap bank failures.

 

[BigBill53]

Hey all,
Thanks for the many comments. I think we have identified the source of the problem. Well, maybe the source of the problem, maybe a symptom.

When we took a close look at the busbars we could see that there was tracking near the standoffs. The tracking was unusual though - the burn marks only appeared to be on the busbar, not down the standoffs and across the back of the board to the other side. In other words, it was as if there were sparks occurring between the busbars.

Now, we aren't sure why this was happening but it definitely has been happening. Whole chunks of copper are missing out of these busbars and it looks like it has been happening for some time.