arcing distance for 13.2 kv conductor to ground 4

[electricpete]

Let's say I have the geometry as shown attached. It is a 2AWG 13.8 kv conductor 4" away from outside corner of a grounded metal structure. Assume the air is dry, the conductors uninsulated, no surges on the system. Would you expect that an arc of this distance (4") along the path shown is possible or expected ? What would be the basis for a conclusion (facts, analysis, standards)?




=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[magoo2]

I would look up Peek's equation. It works out to be about 500 kV per meter.

For 13.8 kV, I think it would be more like 1 or 2 inches rather than 4 inches. So at 4 inches, I wouldn't expect it to breakdown or flashover.

 

[racobb]

Are you asking for hypothetical purposes or do you really have a case like this?

I don't know of any code that would allow this, so if it exists, you would need to immediately modify the construction to meet the applicable codes.

Alan

Democracy is two wolves and a sheep deciding what to have for dinner. Liberty is a well armed sheep!
Ben Franklin

 

[alehman]

One standard for spacing in 15kV metal-enclosed switchgear is 5 inches, so I would not expect an arc at 4 unless there is a large transient that starts it.

Alan
----
"It’s always fun to do the impossible." - Walt Disney

 

[electricpete]

OK, Peek's law - let me try to understand that as a starting point.

http://en.wikipedia.org/wiki/Peek's_law

ev = mv * gv * delta * r * ln(S/r)
where
ev = voltage to begin corona discharge
mv is conductor roughness factor 0.8 to 1. Say 1 for simplicity.
gv = 30 kv/inch = breakdown strength for air
delta = 1 for air at STP
S = spacing
r = condutor radius

Letting delta = 1, ev=1, dividing by r we have
ev/r = gv * ln(S/r)

ln(S/r) will be higher than 1. So the threshhold voltage/distance ev/r will be higher than gv = breakdown strength of air. That doesn't make sense... the stress concentration factors associated with non-uniform field should make ev/r < gv, should they not? I guess I must be missing something.


=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[electricpete]

alehman - that is uninsulated? Is that spacing line to ground or line to line?

racobb - the question arises during analysis of an actual failure. More details will surely bring the thread in a different direction than the simple initial post question that I am focusing on... I will save full discussion of the failure for another thread.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[electricpete]

magoo - To follow up my previous post, I apply the equation as:

ev/r = gv * ln(S/r)
Let gv = 30 kv/cm = 3e4V/cm = 3e6 V/m
S = 4". R = 0.15" for 2AWG
ev/r = 3e6 V/m * ln(4"/0.15") = 1E7 V /m = 10000 kv/m (again doesn't make sense since it is higher than gv)

If I had done ev/r = 3e6 V/m / (ln(4"/0.0.15") = 900 kV/m
That is closer to your answer, but I had to divide instead of multiply to get there. How did you come up with 600 kv/m?

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[racobb]

So that answers the quesion of hypothetical or not....sounds like a failure. Please post more specifics with regard to the metal structure. Are we talking about switchgear, rigid indoor or outdoor bus, or overhead conductor?

Can't believe it is overhead conductor.

Alan

Democracy is two wolves and a sheep deciding what to have for dinner. Liberty is a well armed sheep!
Ben Franklin

 

[alehman]

The U.S. NEC stipulates 5" phase-to-ground for uninsulated, indoors, up to 13.8kV.

Alan
----
"It’s always fun to do the impossible." - Walt Disney

 

[waross]

Outdoors you have to consider the possibility of insects and small animals partially bridging the distance and starting an arc. For birds and small animals look for a partially cooked carcass on the ground. For cockroaches, butterflies, spiders, etc, there is seldom any evidence.

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

 

[racobb]

electricpete,

Without more info, you are wasting alot of peoples time here.

Alan

Democracy is two wolves and a sheep deciding what to have for dinner. Liberty is a well armed sheep!
Ben Franklin

 

[electricpete]

ln(S/r) will be higher than 1. So the threshhold voltage/distance ev/r will be higher than gv = breakdown strength of air. That doesn't make sense... the stress concentration factors associated with non-uniform field should make ev/r < gv, should they not?

Sorry, I was confusing ev/r with ev/S. As an excuse, you threw me a little off-track expressing your answer as a volt/meter... still not quite sure what is meant by your answer 500 kv/m or where you came up with it.

I think I did figure out how to apply the wikipedia formula again to find voltage:
ev = r* gv * ln(S/r)
Let gv = 75 kv/inch
S = 4". R = 0.15" for 2AWG
ev = 0.15* 75 kv/inch * ln(4"/0.15") = 37kv.
I see they call this critical value of line to neutral voltage.... are they meaning
So would seem if my geometry were two AWG 2 conductors separated by 4", then there is a margin factor of 37 kv / [13.2 kv/sqrt3] ~ 4.8 between actual voltage and voltage required for corona. Or if I put in the factor mv=0.8 then a margin of 3.8. Assuming dry air, no surges etc.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[electricpete]

I just realized that the geometry I drew previously was incorrect (I was going from memory... but memory was a little off). The metal grounded structure has a lip overhanging to the left, so the closest point to the energized conductor would be an outside corner on the bottom of that lip, vs an outside corner facing up. This is shown in attached. I will work on seeing if I can get an estimate of the height of that lip.

racobb - I promise, I will present the full failure scenario in another thread. As I said, it will surely lead in much different directions, but I just wanted to explore this question first for my own education to try to understand it as best I can.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[Skogsgurra]

Paschen's law is normally used over here (Europe). Does it give similar results?

http://home.earthlink.net/~jimlux/hv/paschen.htm

Perhaps the same thing as Peek's law?

Gunnar Englund

http://www.gke.org

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

 

[electricpete]

It looks to me like Peek is derived from Paschen. Paschen assumes a uniform field or known peak field intensity. Peek accounts for the non-uniform field effects of this particular geometry (two conductors).

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[electricpete]

Also I suspect Peek is also geared towards geometries where P*d is much more than the Paschen minimum.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[desertfox]

Hi electricpete

I looked at this formula for Corona discharge and allowing the air density factor and surface condition factor of the conductor to be unity I obtained an RMS value 25.98kv for the onset of Corona discharge to occur in your set up.

http://www.copperinfo.co.uk/busbars/pub22-copper-for-busbars/sec6.htm#Corona Discharge

It goes on to say how the air density varies with location at sea level etc which may bring this figure down.

Under the conditions you describe I would not expect a flashover to occur unless some small insects or rodents have got in, if they have I don't suppose there is much evidence of that left.

regards

desertfox