3 phase faults

[Bestboy]

My question is very basic.
We had a big discussion last night with my work mates about 3 phase faults. One engineer said that ALL 3 PHASE FAULTS ARE SYMETRICAL. He even explained that a fault in 3 phases would cause all currents to increase symmetrically and the reverse for voltages. I feel saying all 3phase faults are symmetrical is allitle bit radical. My reasoning is that if the fault occurs in all the phases, at the same point, and at the same instant then the fault will be symmetrical….may I call that ‘a perfect fault’? An example to this would be dropping a metal piece at a perfect position on busbars or cutting all the phases at the same point. So I want to believe that it is possible to have 3 phase faults that are not symmetrical because a perfect fault is almost an impossible happening. I would think that the origin of a 3-phase fault is actually between 2 phases. Then the 3rd phase get involved (to make it a 3 phase fault) due to the aftermath of the initial fault i.e. ionisation of air, flash, insulation breakdown, heat, rupture etc. Any input or correction about this?

 

[waross]

You may describe a three phase fault as symetrical and be correct semantically, but not electrically.
An asymetric fault refers to the fault current having a DC component and not being a symetrical sine wave. The term "asymetrical fault current" is not related to the equality or inequality of the currents in the three phases. It refers to the fact that most fault currents have a large DC component and the resulting fault current wave-form is not a symetrical sine wave.
Given that the magnitude of the DC component, or offset, in a fault current, depends on the point in the sine wave that the fault occurs, it is doubtful if a bolted three phase fault will result in equal asymetrical currents, even if the phase to phase faults occur simultaneously, so refering to a bolted three phase fault as symetrical is probably not symantically correct either.
The magnitude of the asymetric current is often greater than the symetric current, and should not be ignored.
yours

 

[jghrist]

If you have access to relay data from actual faults, take a look at them. I've never seen any where all three phases were exactly the same. A symmetrical fault would have zero neutral current - there's always some. If all three phases are much higher than normal, I'd call it a three phase fault. I'd say this is proof that not all three phase faults are symmetrical.

 

[Rocketir]

You can think of three phase faults as an extreme case of high loading. Most of the time you don't have a perfectly balanced system and there will be neutral current. So if you have it during normal conditions why not during faults. The only difference is the impedances.

 

[stevenal]

Symmetry can mean two different things when describing faults. In school we learn first about the symmetrical three phase faults, then about the unsymmetrical faults that can be described by symmetrical components. Somewhere along the the way we also learn about decaying exponential component that can occur in all cases. Also known as "DC offset" or the "asymmetry factor."

Ungrounded and balanced three phase faults will not have any zero or negative sequence current, so they are symmetrical by one definition. Two or more phases will have a decaying exponential term, so the unfiltered current will have an asymmetry factor. Confusing enough?

 

[jghrist]

waross,

I think the term "symmetrical" in this context means all three phases balanced, as in symmetrical components.

 

[Bestboy]

Thanks gents for the quick reply. I am seeing the light now. However I have some questions: -
1stThe impedance we are talking about is it line impedance or load. Because I hear that in normal fault analyses you ignore the load impedance? The faults I am referring to are ones that may happen up-stream....supply side. Therefore making the load impedance immaterial.
2nd question is: - If you will always have neutral currents, then it looks like there are no symmetrical faults in real world. What are symmetrical faults then?
3rd And by the way you are talking about neutral currents. What if its delta connected system? I assume the other side of supply transformer (wherever it is) will be star connected hence neutral and neutral currents. Are we together or em I missing something? I hope you are seeing it my way.

 

[Bestboy]

sorry I missed some replies before sending my 2nd question. Forgive me

 

[jghrist]

There are no truly symmetrical faults in the real world for the reasons mentioned in your original post. We do not live in a perfectly symmetrical world. Metal pieces do not fall on all three phases at the same time. If you looked at the fault records where a high speed grounding switch was closed, you might get a pretty balance fault, at least if you look at filtered output without the dc component. Phases are not perfectly transposed, however, so even in this case, you won't get perfect balance.

If the fault does not involve ground or neutral, you will not have a neutral current. I just haven't seen records of such a fault.

 

[rbulsara]

I would say what waross said in his first post is correct.

Symmetrical 3 phase fault currents are referring to the symmetry of the AC sine wave and not the balance between three phases. 3 phase faults are assumed to be 'balanced' for the purpose of anyalysis and specifying short circuit rating of equipment.

 

[Bestboy]

I need to go sleep and try to consolidate all the obove fair answers tomorrow morning.... I am on +2GMT. Once again thankyou everyone. And for those who still want to contribute you are welcome. Bestboy

 

[stevenal]

"I just haven't seen records of such a fault"

I'm looking at record of such a fault right now, no discernable rise in residual current. Although I'm not yet sure of the cause in this case, these faults are common enough. A branch breaks free in the wind and lands across three phases. Of course 2 phases go high before the third joins them, but the fault eventually becomes a symmetrical three phase one.

If you want a fault that evolves less, try closing into a line with the personal safety grounds left on.

 

[DanDel]

Bestboy, you seem to be confusing a theoretical 3-phase bolted fault for analysis purposes and an actual fault. Your questions:


"1stThe impedance we are talking about is it line impedance or load. Because I hear that in normal fault analyses you ignore the load impedance? The faults I am referring to are ones that may happen up-stream....supply side. Therefore making the load impedance immaterial."

In the theoretical 3-phase bolted fault, there is zero impedance at the fault connection, and the upstream sequence impedance is symmetrical.
In an actual fault, the fault current depends on the actual impedance at the fault, i.e., the impedance of the 'metal piece' you dropped and it's connections to the buses, along with the true upstream system impedance, which is never truly symmetrical.


"2nd question is: - If you will always have neutral currents, then it looks like there are no symmetrical faults in real world. What are symmetrical faults then?"

3-phase symmetrical faults are used to model the 'worst-case' fault in a short circuit study to determine the adequacy of the connected equipment.


"3rd And by the way you are talking about neutral currents. What if its delta connected system? I assume the other side of supply transformer (wherever it is) will be star connected hence neutral and neutral currents. Are we together or em I missing something? I hope you are seeing it my way."

Neutral currents are not modeled in a 3-phase symmetrical fault, whether the system is 3-wire or 4-wire.

 

[stevenal]

To back up my first post, I'll reference Stevenson who introduces the decaying exponential term in a chapter called Symmetrical Three-Phase Faults. To clarify this bit of confusion, symmetrical and unsymmetrical faults refer to balanced three phase faults versus the other faults. Symmetrical and asymmetrical current refer to current without and with a decaying exponential term.

To the original question: I do not believe the three phase fault that evolved from a line to line fault is somehow less symmetrical than one that involved all three phases initially. But the concept of the balanced three phase fault does rely on some engineering approximations such as balanced in magnitude and phase source voltage, balanced source impedance (make sure the lines are transposed and that the fault occurs only at the end of a complete transposition) and balanced fault impedance. None of these things are achievable to perfection in the real world making the perfect symmetrical fault also unachievable. We can get close enough for all practical purposes though.

 

[jghrist]

From the OP:

One engineer said that ALL 3 PHASE FAULTS ARE SYMETRICAL.

This engineer is wrong (not just in spelling;-)) no matter how you define symmetrical, if you define a 3 phase fault as a fault involving 3 phases.

 

[rbulsara]

I suspect the OP misconsturded what the other engineer might have said.

The OP is obvioulsy confused in the meaning of Symmetrical and 'balanced' in this case.

 

[Bestboy]

Quote from Rocketir(3 Aug 06 11:04):-
'..... The only difference is the impedances'

My question about the impedance was triggered by Rocketir’s post otherwise I was not confused. We are all on the same page now.
Thanks to you all for the valid posts and contributions. I am glad it’s all ending with some humour from 'jghrist'

 

[whycliffrussell]

Common fault literature says (forget the semantics…) that there are two general fault categories:

1. Symmetrical Faults:

This simply means 3 phase ‘bolted’ faults…hence in the analysis of such a fault you don’t need to worry about breaking the system down into symmetrical components b/c the system is symmetrical (theoretically at least). We know what in the real world this is never really true…

2. Asymmetrical Faults:

All other faults! L-L, L-G, L-L-G…To perform the analysis on this type of network you must break the system down into symmetrical components (because the faulted system is no inherently symmetrical) to perform analysis.

BOTH 1 and 2 above have:

a. Asymmetrical region:

This happens @ the beginning of the fault. This region generally consists of a sinusoid of sorts riding an exponentially decaying dc offset.

b. Symmetrical region

This region follows the decaying dc offset. Once the dc offset decays to zero and the faulted current follows a sinusoidal pattern which is (more or less) symmetric) about the x-axis.

Anything else is just semantics…It's more important to understand what's 'going-on'.

 

[jghrist]

I would modify what whycliffrussell (why not cliff russell?) says:

I would refer to item 2

2. Asymmetrical Faults:

All other faults! L-L, L-G, L-L-G…To perform the analysis on this type of network you must break the system down into symmetrical components (because the faulted system is no inherently symmetrical) to perform analysis.

as unsymmetrical faults. This is how Stevenson refers to them in "Elements of Power System Analysis". I'd leave the asymmetrical nomenclature to refer to the dc offset.

 

[whycliffrussell]

i agree w/ the unsymmetrical and asymmetrical nomenclature...actually that's what I ment to use when i wrote the post...Thanks.