Power Swing - "speed of the fault"

[521AB]

Usually power swing protection for distance relays is achieved by measuring the the speed of the impedance trajectory (DELTA Z(t) / DELTA t).
If the speed is faster than a setting value, it is a fault, if it is slower, it is detected as power swing and distance protection is blocked.
Any idea if there is any study or any statistics giving a reasonable order of magnitude for the "fault"? This could make easier the setting calculations for power swing protection!
Thanks!

 

[Alex68]

Usually the power swing protections have a "tomato" shape.
The function counts the time between crossing the outer and the inner curves.
To achieve a correct timing of the function you need several transient stability studies, but you can consider a travelling time of 40 - 50 ms.

 

[521AB]

Alex68.
Thanks for your answer.
I know that there are studies (from EDF? I'm not sure) about the speed of the fault. It would be interesting to know if they are available or if their results are known.
I have also to say that you can set the outer border amd also the inner border, in practice you set the DELTA Z. Just setting the time is not enough.
Additionally, after the first power swing, the second "race" is usually faster than the first one, so there is an additional timer for the "second race". Not easy to set if you do not have any system study.
If we could know that a fault has a speed of -let's say- >=500 ohm/s (based on a wide statistics), than it would be enough to set the power swing to 400 ohm/s amd a lot of problems would be solved.

 

[Alex68]

I'm sorry, I've never heard anything about these studies.
I am very interested in it and I'd like to have a copy.

I have set several out of step protections of synchro generators. I prefer a double blinder characteristic instead of a tomato one. Because you can trip only if the impedance is in the 2nd or 3rd quadrant and you are sure of the pole slip. Using a tomato shape you trip considering only the time to cross the borders, even if it is a stable power swing.
Moreover with a double blinder characteristic you can trip the circuit breaker far from the 180 degrees of voltages.

But I know that a tomato shape is preferred for line protections.

 

[521AB]

Tomato shape?
I have always used a double quadrilateral characteristic (Zin border and Zout border) eventually with load encroachment area to "avoid" the load.
I do not understand what this tomato shape is...

 

[Alex68]

You can find an example in this manual:

http://www.geindustrial.com/multilin/manuals/lpso.htm

 

[521AB]

I see now, it is the MHO characteristic (or called circular).
Actually I do really prefere to use the quadrilateral characteristic, even for Power Swing and not only for distance protection.
Siemens and ABB, for instance, do have such characteristics since lot of years. Anyway it appears to me that GE also measures the time from Zout to Zin; it's just using circular characteristics for the two zones.

You mentioned "trip" from pole slip.
I see some confusion in the terminology here.
My understanding is so far the following:
- Power swing protection: detects a stable oscillation and blocks impedance protection
- Pole slip protection: detects unstable oscillations, can count them and can trip at any position of the impedance trajectory inside the characteristic and after a certain number of oscillations. It can also detect a stable oscillation and block distance protection, but this is not its main purpose. Its main purpose is to open the circuit breaker. It can also be used in some significant lines in the middle of the netwqork, so far away from generators, to "split" the electrical system in dangerous situation.
It is of course used to protect the generators (historically this was its original use).
- Out of step. Sometimes it means pole slip, sometimes it means power swing. Not really clear to me.

My regards

 

[Alex68]

You can add also "loss of step", which is the same as "out of step".
The meaning is simple: when a "pole slip" occurs, the generator experiences a "loss of step" and it is "out of step".
We are discussing about the same phenomena from different points of view. You look from the grid and I look from the generator.

The tomato shape is a combination of internal and external circles. The aim is the same as the quadrilateral shape: detecting a power swing and blocking the minimum impedance function, when the travelling time between the external and internal borders is higher than the setting.

 

[521AB]

Which means "out of step" means "pole slip" --> unstable oscillation --> CB open!
Right?