480 volt TVSS failures

[Andy32821]

Hi,
I am at a large convention center that has forty 4000 amp 480/277 main switch boards.

Each main switch board is feed from an independent utility owned 12kv to 480/277 transformer.

The transformer has no lightning protection on the secondary side. The main switch board has a N-G main bonding jumper and a TVSS .

Four times in the last two years on different switch boards, the TVSS units have failed in a catastrophic fashion. A lot smoke, some minor internal switch board damage, a completely destroyed TVSS unit and a main breaker trip. Of course management is not too impressed when guests have to be evactuated , fire department has to respond, and power cannot be restored until the TVSS unit is gutted and the switch board internals inspected for damage.

Here’s the interesting part. It is not the MOV’s failing. In fact in all four cases the MOV’s and the sand surrounding the MOV’s are in good shape with no damage. (But all the fuses did open).

Both the manufacturer and an independent R&D center have confirmed this. Instead the arc blasts seem to becoming from facing energized lug connections and fuses that feed the TVSS.

These are year 2000 model TVSS units that (by design) had no insulating barrier between the input lugs, and the TVSS’s disconnect switch input lugs are connected with unprotected #4 awg copper wire directly to the switch board bus. The entire TVSS unit consisting of a disconnect, fuses and TVSS unit is mounted by the manufacturer internal to the switch board. Newer models of this style TVSS have fiber insulation barriers between all exposed energized metal parts.

The manufacturer is going to offer their ideas sometime in the future and we are going to work with the power utility on power quality testing. Both of these are going to be drawn out events that will take us well into peak show season.

I would like some advice on making some meaningful short-term solutions to reduce the potential for the phase to phase arc blast while the engineers come up with a permanent solution.

My plan is to wrap the lugs and fuses with medium voltage insulation sheeting and ty-rap it in place.

Questions:
Is there any technical down side to this?
Would this prevent the arcing?
What type of insulation would be best, both from ease of installing and from insulation qualities?
A retailer name would be great. All I find are importers selling to OEM's.
Any type of material I should avoid? (rubber?)

Any comments or advice will be much appreciated.
Thanks,
Andy

 

[rbulsara]

Sounds like a poorly desinged TVSS.

Short term solution would be to remove the TVSS, rather than messing around with it and be liable for future mishaps. This could even be the long term solution.

Alternative long term solution is to install better TVSS with good history.

Rafiq Bulsara

http://www.srengineersct.com

 

[rbulsara]

By the way, you may want to investigate what may be causing the phase to phase overvoltages that are sufficient to flash over the poorly designed lugs/fuse terminals but not activate the MOV.

It will also help to know magnitude of the overvoltages if being caused by switching of industrial equipment.

Rafiq Bulsara

http://www.srengineersct.com

 

[rbulsara]

Other thing to look at would be the coordination bewteen the main breaker and TVSS fuses. These are all secondary things to do, after fixing the TVSS.

Rafiq Bulsara

http://www.srengineersct.com

 

[krbernier]

the design of that this is disgusting. if any of the bolts holding the lugs loosens up your looking at a phase to phase short. in my opinion the fuses are really scary..not to mention the conductors on the other side are going to pull those lugs down into each other.

first thing id do is go through them all and make sure everything is torqued correctly. id also try to make some sort of fiberglass shield between the phases. good luck

 

[krbernier]

after looking at it again id get rid of those mechanical lugs on the left and do an insulated crimp connection bolted directly to the tabs and id relocate those fuses to a conventional fuseholder setup and just bring wires into that disconnect like id do to the other side.

im willing to bet your problem is loose connections leading to phase to phase / neutral faults

 

[krbernier]

are those fuses or the mov's?

 

[dpc]

So in the photo, I see the special fuses, the special switch, but I'm not seeing the actual surge suppressor?

It seems like an installation problem in the way the components were installed at the factory.

I agree with Rafiq, in the short term, you could open the switch and just do without the TVSS - they seem to be creating more problems than they are solving.

I'm not clear where the failures are occurring. If the faults are happening before the switch, the leads can be disconnected and taped.

I've always had good luck with GE surge suppressors, but again, I'm not sure the TVSS is actually the problem.

 

[Andy32821]

Hi Guys,

If I turn off the TVSS disconnect I still have the bare input lugs that can arc, but no longer have MOV's to try and clamp the voltage.

This may actually case more of an arc chance on the input side of the disconnect.

Right now I can't do much modifying. I have to wait on GE and the power utility to make thier recommendations. They may very well recommend some of your ideas. Installing new TVSS units to today's code will probably cost $200K, by the time I install breakers and mount new TVSS units.

My short term choices are to completely remove the TVSS and put the convention center's electronics at risk, or try a temporary fix.

I am hoping someone can give me some guidance on the temporary plan of insulating the lugs and fuses.
Thanks,

 

[krbernier]

do you have any pictures of the past failures?

from that new picture i am less concerned with the area between the disconnect and the tvss itself as it looks like there is not a great chance of phase to phase contact.

if you remove the mechanical lugs on the left side and use a crimped terminal on the wire with some heavy shrink tubing over it to insulate all but the portion at the end. then just bolt the connector directly to the disconnect tabs your chances of a phase to phase fault would go away in my opinion.


those mechanical lugs like to loosen up where they bolt to the disconnect. i frequently find these have not been adressed during yearly maintenance. it does no good to torque the connection between wire and lug if the lug itself isnt torqued.

the plant i maintain had no prreventitive maintenance program for its switchgear before i was hired on and since i started going through every mcc yearly failures and faults caused by loose connections are no longer happening.

 

[dpc]

Without more accurate information regarding the location of the previous failures, it is difficult to say. If you don't feel opening the disconnect switch will suffice, then disconnect and remove the cable that connects the switch to the main bus.

The TVSS units are not required and it will be safer to operate without them.

I'm not convinced that the failures are caused by surge voltage. You would likely have other evidence of transient voltage levels high enough to arc over (in air) the distances shown in the photo.

 

[Andy32821]

One of the bad units had a loose lug (the first actually) and we thought we found the problem. However all three others had lugs tight and inplace. In every case the event had enough energy to completely melt off one incoming #4 awg copper lead. Once this unprotected lead was loose there is nothing to clear a fault except the main breaker. So I don't know if the event tripped the main, or if a loose lead triped the main. But there is pleanty of arc damage inside the swichboard with the TVSS disconnect switch being the heart of the blast.

Inspections from two different sorces say that the MOV's are in good condition and working, and they say the lugs and fuses show signs of pp arcing.

Lighting season will be over here (in Florida) soon. Sounds like the best I can do is remove the copper leads that feed the TVSS from the bus until the issuse is decided by others.

All the blasts occured early morning, probably about the time for tap changes, but not a common time for lightning.

I tried to read the VSP100 data sheet to see if two of these fuses could pass enough current to allow a #4 wire to melt without blowing. But the data sheet is voodoo to me.

 

[dpc]

These fuses are supposed to allow enough energy to pass to give the surge suppressor a chance to operate. I don't think they would allow enough energy to melt a #4 wire.

So the faults were downstream of these fuses?

 

[Andy32821]

The general consensus from the inspections is the MOV's did what they should have, however something blew all three fuses. At that time or at later time, a phase to phase arc between the metal lugs at the disconnect produced enough of a blast to blow a lead wire off. What triped the main, the event or the lead wire, is anyone's guess.

I have had a lot of pushback at the idea of an arc at this distance. The only other (non arc)thing I can think is the MOV turned on but the fuse did not open allowing the lead wire to melt. But this is unlikely since the MOV's were tested and turned on at the proper voltage. Normal voltage runs about 500 volts. I don't know the MOV turn on voltage, but probably over 700 volts.

Right now, mostly event seems to be pp arc at disconnect switch.

 

[LionelHutz]

I'd agree that it doesn't make much sense that there would be a flashover at that switch. It seems to have decent phase to phase spacings for 480VAC.

Raycap has a supressor called a Strikesorb that is a very capable supressor. If suitable, you could do away with the disconnect and fuses and direct connect these.

http://www.raycapinc.com/index2.htm