Autotransformer and line addition- Drop in short circuit current?

[Mbrooke]

Is it physically possible for a major increase in fault current at one substation to actually cause fault current to drop at another nearby?


Here is the scenario. From a 500/345/115kv substation (I will call this substation "A") emanate several 115kv lines which travel about 12 miles to substation "B" From substation "B" 115kv lines travel about 15 miles to substation "C".


For variety of factors a new 345kv lines is being proposed from substation A to another one some distance away. Substation B will tap this new 345kv line, and via a 600MVA transformer connect to the 115kv bus inside substation B.


When all is said and done, computer simulation shows that fault current actually drops at substation C despite there being a good increase at B and even some at A. All the 115kv lines between A, B and C are identical in conductor size and configuration.


Here are the numbers as they are now...

Substation A:

345kv; 31,560 amps --- X/R of 19.5

115kv; 42,700 amps --- X/R of 26


Substation B:

115kv; 12,360 --- X/R of 8.7

Substation C:

115kv; 7460 --- X/R of X/R 11.5


With the 345kv additions...


Substation A:

345kv; 31,645 --- X/R of 19.3

115kv; 45,440 --- X/R of 24.4

Substation B:

345kv autotrafo breaker; 22,800 X/R of 17.2

115kv; 21,815 X/R of 21.7

Substation C:

115kv; 6150 X/R of 12.5

 

[davidbeach]

Seems weird, but I’ve seen many weird things given various system configurations. A couple of things to check.

Are you using the same model and just turning the new equipment on and off? If not, turn off the new in that model and see if you get the same fault values.

How does the prefault voltage at C compare between cases? That new transformer at B might be pulling down the prefault voltage there. An easy trick to get the prefault voltage is to run a very high impedance fault, say 99999 + j99999 Ohms.

 

[Mbrooke]

Thanks, I will give that a try.


FWIW, in steady state peak load analysis the voltage at C is 0.952 pu and 0.967pu at B, 1.008pu at A. Simulated peak with the 345kv system is 1.013 at C, 1.027pu at B and 1.012pu at A.

 

[marks1080]

Is it possible? Yes.

 

[Mbrooke]

What is the physics behind it? I never knew that was possible until now and I have to admit am fascinated by that fact.

 

[marks1080]

Changes in system impedance. Operators will change system configuration all the time to manipulate available fault currents.

Without seeing a model of the system you're talking about I'm not sure I can get more specific. I've seen it enough on my side to not think of it as strange (anymore).

Based on my experience using modeling software, I also wouldn't discount the possibility that whoever built the model made a mistake. Certain transformer configurations are not easy to properly model.

 

[Mbrooke]

Can you give any examples from your side where fault current went down where conventional wisdom said otherwise?


Modeling could indeed be a factor here- 600MVA auto transformer at play so that might be giving the program heart burn.

 

[davidbeach]

What modeling software are you using? Could you extract that portion with a boundary equivalence and obscure any names that aren't just A, B, C, etc?

 

[Mbrooke]

PTI PSS/E

Hence why I question the results. The software is very accurate with the right data plugged in.

 

[jghrist]

Are there other sources at C? If not, then I don't think the fault current would decrease at C with a change that increased fault current at B. The Thevenin equivalent system impedance can be calculated at B from the fault current. If the fault current increases, the Thevenin equivalent impedance will decrease. This Thevinin equivalent impedance can added to the B-C line impedance be used to calculate fault current at C, which will be higher.

 

[Mbrooke]

Ok- you are on to something. There is cogeneration at the distribution level. That might be it!


However I am still curious, and for the sake of discussion- lets say it is not from that. What could be responsible?

 

[wroggent]

What kind of fault was studied?

 

[Mbrooke]

Worse case fault to determine breaker interrupting duty.

 

[wroggent]

Thought it might be. Did the worst case type change? Maybe it used to be SLG but in your new configuration the worst case is now a 3 phase fault due to the introduction of more zero sequence impedance somewhere and the 3 phase fault current may even be slightly higher in the new configuration versus the old.

 

[Mbrooke]

That could be it as well- more for me to consider.

 

[HamburgerHelper]

Mutual coupling can give you odd things like that.