non load break switch

[protoslash]

We have a simple non load break switch on a 44kV line (circuit switcher, in-line disconnect w/e you call it).

The switch is energized but not carrying any load downstream. My understanding is that the switch can be opened manually as it is technically not breaking load current. However, when we did this, the switch drew arc and faulted.

One of my theory is charging current due to mutual capacitance of 3 phase line that is causing current to flow in the switch. If there is very little current (<1A), what would happen if the switch is opened? Should it still be able to handle it?

 

[davidbeach]

It depends on the line drop rating of the switch. High-speed whips can be very handy in raising the line drop rating.

Circuit switcher is actually a thing, quite different from a disconnect switch.

 

[cuky2000]

Disconnect switch for this voltage level and above are rated for very small current interruption and even capacitive current stored in the line could exceed the switching rating creating a catastrophic failure in the equipment as well a system fault.
There are accessories commercially available that will help to mitigate this event.

 

[protoslash]

thanks for the replies. One more related question:

If two adjacent feeders are framed in parallel on the same pole. One feeder is loaded while the other is not. Would the loaded feeder phase conductor induce current onto the other feeder phase conductor?

My feeling is no. While there is induced voltage, unlike neutral conductors that are multi-grounded, the phase conductor cannot complete a close circuit for the induced voltage to drive a current. Thoughts?

 

[davidbeach]

Charging current could be there. The capacitance of the line has to be considered.

 

[Palletjack]

you say drew arc and faulted. What exactly did it do?
I don’t understand what faulted.
I open 115KV lines with no load often with a non load break disconnect. The line is 6 miles long. Yes it’s loud, but doesn’t open the breaker 80 feet on the line side of it.

We open 25 kVA lines with non load break even more often with no problems.



As a side note, I opened a non load break disconnect (on a motor operator) that was feeding a loaded substation supplying four circuits. Was that loud? Very. Did it break the load? Sort of... Right after the 115KV breaker opened 18 miles away.

 

[cuky2000]

To visualize the issue here is an oversimplified calc. for the total line charging current of a 115 kV line:
a) Line charging current is ~0.25 A/km.
b) Line length =6 mi (9.65 km).
c) Total charging current ~ 2.4A > Switch Rating Capability.

For reference, see the graph and the enclosed Link

 

[Bill West]

Protoslash, I have to raise a big caution here. We were cutting a series capacitor station into the middle of an existing 500kv two circuit stretch, 250 miles long. One line was out, grounded at the adjacent substations and at one point in the capacitor station, the parallel line was under load. The capacitor station isolation and bypass disconnects were all in place and connected but the drop to one of them had to be cut and reconnected to correct its sag. As soon as the poor fellow sitting on top of the grounded disconnect cut through the drop, induction made it live from the distant substation. The jolt caused a fatal fall to the ground.

So while this discussion is fine for 44kv switch capabilities and parallel feeders we shouldn't let it cause us to miss any safety concerns about parallel induction. What was particularly disturbing in this case was that the grounding rules seemed to be being obeyed until the moment of the cut. My broad lesson was that there is no such thing as a dead 500kv line and more broadly the old chestnut that it isn't dead until it's grounded.

Bill

 

[Palletjack]

To visualize the issue here is an oversimplified calc. for the total line charging current of a 115 kV line:
a) Line charging current is ~0.25 A/km.
b) Line length =6 mi (9.65 km).
c) Total charging current ~ 2.4A > Switch Rating Capability.

For reference, see the graph and the enclosed Link

I understand all that very well.
I have ours calculated already with other factors added in such as humidity levels and under built inductance.

My question to the OP was what faulted.
From an engineering perspective a fault is abnormal currents. I do not consider charging currents abnormal currents.
Did something trip or did it just arc? Does he have whips on the switch? The switches are rated to break charging currents. At 44KV if it arced a lot and faulted something was still energized.
If this was a sub feeder maybe the sub XF and regulators were still energized.
Who knows until we get more information.

 

[protoslash]

OP here with more info

the disconnect switch is close to the end of a 5km 44kV feeder. the feeder is loaded but since the switch is close to the end with no load downstream, there shouldn't be current through the switch.
photo of switch: [URL unfurl="true"]https://imgur.com/a/M8Wm0HF[/url]

crew opened one phase of the switch with a stick, arced, burned gear, but not hurt. Feeder breaker tripped on instantaneous overcurrent. The fault was on one phase but ionized air developed into two phase to ground. see relay event here: [URL unfurl="true"]https://imgur.com/a/dAVXeaG[/url]

After feeder closed back, crew opened other two phase no problem.

Questions still remains:
-if it arced, then there must still be current. But towards the end of the line, charging current should be small.
-what about current from parallel feeder induction? how does this even work if both feeders are energized (44kV 3wire solid uni-grounded at station)?
-other two phase opened fine under same condition, most likely defective equipment on the bad phase.

thanks for the replies.

 

[davidbeach]

Certainly not what I’d mentally pictured. It arced, the arc wafted into a phase above and the two phases shorted through the arc.

 

[Mbrooke]

I could be wrong- its early for me- but in the waveform you linked the current on the "red" phase seems to spike for about a cycle and half before evolving into a 2 phase fault (red and yellow).


I can only guess there was a line to ground fault initially which evolved into a multi phase fault.

 

[Bill West]

I'm going to stir up some crazy thoughts here. Suppose:
1. there was preexisting downstream leakage to ground on red phase
2. that lead to the switch arcing on opening
3. the arcing caused voltages surges that made the initial leakage into a fault
4. that gives the jump in red phase current in the recording (MBrooke's observation)
5. and made the arc big enough to rise into the yellow phase
6. then the recording shows that L-L-g fault and trips the line
note that it is not just L-L, red is greater than yellow and blue rises somewhat.
is red still feeding some L-g fault downstream of the switch?

The sequence of the fault currents is clear in the recording. Needing a reason for 2 is what leads me to 1. Then 3 can lead to 4 and the rest of show goes boom.

What is really weak here is what leakage (cracked insulator to wood pole? dry ground?) would be enough to upset the switch and yet not so much power at 44kv that it didn't just torch the leakage path? Nonetheless it's whatever upset the red switch that started the whole show. The tail end of a 5km 44kv line doesn't seem likely to have enough capacitance or parallel induction to do that, not even enough inductive kick from the 5km. Can I suggest to ProtoSlash that a close look downstream along the line might be worthwhile. Explaining this one seems worthwhile because the crew doesn't need this sort of experience again.

Bill

 

[Mbrooke]

This is a stretch but it went through my mind- what if the initial rise in current on the red phase and then the blue phase during the L_L_G fault is from reduced voltage causing specific loads to pull more current?

I mean it would make sense if the disconnect was accidentally opened under load, arced in series traveling upward exploding into a 2 phase fault?

Just a wild guess as I can't picture load causing a current spike that large but anything is possible.

Another wild guess is VT or PTs left down the line triggering ferroresonant over voltage.

 

[Palletjack]

Looking at the syncrowave, I see amps, and quite a lot before the event.
Back your curser up and see what the amps were.
On an unloaded line you sine waves on the currents should be non-existent.

 

[protoslash]

Bill, I like your theory.

Crew did find broken insulator on a primary meter downstream of the switch (sry i don't include this piece in the original post). I am going to ask on Monday which phase switch was opened and which phase the broken insulator was found.

Question on point 3, why would arcing cause voltage surge?

Mbrooke, "red" phase current surge was close to 3000A, must be fault not load. can you elaborate why PT could cause ferroresonant over voltage? As stated, there is a primary meter downstream of the broken switch with PT connected.

Palletjack, The feeder is loaded, the switch is not since it is almost the end of the line.

very good discussion folks

 

[Palletjack]

What was the primary metering feeding?
I guess what I’m getting at is was the line completely unloaded?
Some more information seems to be missing for this to happen.

Ferroresonance? Depends on line length and size. Possible if the line was unloaded and you had matching inductance reactance and capacitive reactance in the primary PTs and the conductor.

Possibility this could have also caused the insulator to fail at that point with the increase in voltage during that event.

 

[Bill West]

Thanks Proto. Just for clarity, I can picture 2 types of primary metering set ups, which fits you?
-a pole transformer like tank with all the CTs & PTs inside, oil filled, porcelain bushings.
-a pole rack with individual CTs & PTs, oil filled cans or cast resin style

The metering PT itself is of course an inductance. One place for MBrooke's ferroresonance here would be the magnetizing current increasing sharply on overvoltage, the current pulls the voltage down, the current falls back, repeat. It's the non linearity of the magnetizing curve that leads to the values oscillating back and forth.

In the absence of resistive loads the inductive currents can be harder on a switch because when the sine voltage zero stops powering the arc the current is at max. This is the same as shunt capacitor switching.

For arc induced voltages I was thinking of the switching surges that show at line ends on high transmission voltages and are a reaction between series inductances because of sharp current interruptions during switching. Of course what you likely have here are low line inductances, only the PT might be a meaningful one. As PalletJack mentions the broken bushing may have been okay before the switching but failed because of a voltage spike during it. That would take less speculation than my leakage current idea. Unfortunately the recording can't show us the voltages at the line end but red does show higher frequency spiking at both changes in current.

Without specific line properties and an elaborate switching surge study we may still be stuck guessing here. Too bad because cautioning against dropping metering kits with a switch could be overkill for what is just guess reasoning on a one off case and yet we wouldn't want this to happen again.

Bill

 

[protoslash]

Both failed switch and broken bushing was on the red phase.

The primary meter has exposed CT PT on the pole. Any idea how much excitation current a primary meter instrument transformers pull on a non loaded line?

I ran my system in a model, and it showed 0.6 A of charging current through the switch in question. I will probably leave the sequence of events like this:

1. combination of poor switch condition, charging current, metering instrument current caused the opening of the switch to arc.
2. the arc caused enough transient voltage spike to upset the poor PT insulator downstream. The failed PT location faults SLG. The switch is far from making a contact to ground so I deduced it must be the PT bushing that is faulting initially.
3. The arcing of the switch ionizes air and develops into line to line fault
4. At this point the SLG fault downstream still exist but the ground fault magnitude is reduced since the l-l fault pulls down the voltage. (added a few lines to the waveform)
5. breaker clears all fault in 6 cycles

It is a lot of guessing going on indeed.

 

[Palletjack]

Going to go out on a limb here...
were the PTs connected in Delta?