Earth fault detection on ungrounded system 1

[dfdt]

I would like to hear your comments/views on following
We have single 6.6kV distribution feeder being supplied by ungrounded source via 110 kV(ungrounded star wdg)/6.6kV(delta wdg)transformer. Although technically speaking if we have earth fault on this feeder the earth fault current would be 3 times of steady state phase to ground capacitive current as voltages of healthy phases will rise to √3 times. My question is that "can we use this rise in voltage (from VLL/√3 to VLL) for detecting earth faults on ungrounded system?

 

[davidbeach]

Yes.

 

[dfdt]

thanks David for prompt response
and only issue I see is that it's bit unreliable as it is dependent of feeder capacitance and does transposing or lack of transposing will cause any issue?

 

[davidbeach]

It's actually a widely used technique. Voltage is the only way to detect ground faults on ungrounded (capacitively grounded) systems. But the system capacitance and charging current has nothing to do with the detection. One phase is zero volts to ground, by definition, and so the other phases will be at phase-phase voltage to ground, even with no capacitance and no current flow. After all, you're not measuring current. It can be detected with as few as a single VT, but three VTs connected in broken delta (what the IEC folks insist on calling open delta, but actually quite different from an open delta) is by far the preferred method. Single VT requires a 27/59 relay, the broken delta only requires a single 59 relay, but whether included with the relay or separate it should also include a loading resistor to help avoid ferroresonance.

 

[dfdt]

I agree the right terminology is broken delta as open delta in my opinion describes V connected VTs. Now in this application we will be sensing earth fault simply by monitoring over voltage across star connected VT secondary winding rather than voltage across broken delta as VT only has one secondary winding connected in star formation.