Time-graded Overcurrent Protection

[Cerkit]

Hi,

Can someone please explain why it is not possible to use time-graded over current protection on ring main system which has multiple sources?

Is it because the it is difficult to distinguish between the fault currents in different sections of the ring?

Thanks

 

[veritas]

With multiple sources it can be very tricky. I would need to see the actual network to give more specific comment. In general if fault current can go either way at a relay location then directional protection is usually used. You would also need to consider the change in fault levels and relay sensitivity if some of the sources are out of service.

 

[Cerkit]

I don't have a specific network. Juat trying to understand why it couldn't be applied. Even with directional overcurrent relays everywhere I have read says that it is not possible to obtain adequate grading. But I'm not entirely sure why.

Thanks

 

[Cerkit]

In the attached for example, If I added another infeed.
Is the issue because of the change in fault current if one source was out of service or more that the fault current varies significantly when looking clockwise along the ring?

 

[veritas]

The setup in your diagram looks fine for the given source. However, if you introduced as source at say the 6 o'clock position then dramas! Say you had a fault between 1 and 6' with the two sources in service. the breaker at 1 would trip as it is set to 0.1. Ideally the breaker at 6' should trip as well. But the one at 1' will trip first as it is set to 0.1. So selectivity is lost.

I would say the issue is more than change in fault current distribution. in this instance. However, I would certainly also pay attention to the relative strengths of the sources as they would influence the sensitivity (pickup) settings of the relays.

I would most likely go for unit protection in the form of cable diff as my main protection and then a time/current graded backup scheme which can be customised by using different settings groups depending on the source configuration.

 

[Cerkit]

Hi. Yeah sorry I should have clarified that I wouldn't maintain the same times as in the single source example, but assuming same topology and introduce a second second. From what you are saying the fault level distribution will be the issue. So the 5oclock circuit breaker would have issues grading with the 6oclock breaker (looking clockwise) as the 6 oclock breaker will see a significantly higher fault current from the contribution of the second infeed.
Does that sound right?

 

[HamburgerHelper]

Infeed, changing current direction, .

One to keep in mind on a ring bus is that you don't know how the current is going to go around the ring since there is very little impedance. I have seen weird splits on the phases. I have seen phase to phase faults produce splits that look like zero sequence currents on some CTs due to how the fault current pathed around the ring bus. When you have two paths on a bus to a fault, you will not know how the current will splits If you have one path to the fault or a appreciable amount of impedance on each path so that slight variations due to connection impedance, spacing, or equipment don't significantly change the flows, then it will be more predictable. So, I don't even think it is really doable with a single source on a ring bus.

At the transmission level,which is not at all single sourced, there are a lot of compromises made to coordinate ground time overcurrent at the transmission level. Enough so that I often questioned the point of even coordinating time overcurrent curves if it is going to take 1-2 seconds (an eternity at the transmission level) for them to operate and that will only happen if the communication assisted scheme fails. Some utilities use ground distance as the primary ground protection and ground overcurrent just as a really slow backup that doesn't necessarily need to coordinate.

------------------------------------------------------------------------------------------
If you can't explain it to a six year old, you don't understand it yourself.

 

[LFAVianna]

The maily reason to dont use an overcurrent time-graded without any directional control (ANSI 50/51 functions) is the dificulty to obtain selectivity and coordination since, as above mentioned, you will have different currents flowing by the ring with a large range of current values. Overcurrent elements with directional control are used to turn a ring system into a radially system, because only the relays with the same direction will have need to be coordinated.
There are another functions to garantee these type of protection but it will depend of the size of the system, like short-circuit levels and impedance data.

 

[marks1080]

I would still expect to see some 50/51/67 functionality to beef up what is probably a distance or differential protection. For example, what would you use for a stub-bus or line test protection?

 

[jghrist]

Consider two o/c devices A & B in series with sources on both sides. If a fault occurred at F1, you would want A to trip before B so that the section between A & B would stay in service. If a fault occurred at F2, you would want B to trip before A. It's difficult to achieve both without directional relaying.

S1---F1----A------------B-----F2-----S2