50N vs 50G Currents 1

[JMatt]

In most SEL relays, there are two different ways to measure ground fault current. One way (SEL relay word 50N) uses the current as measured by a core-balance CT physically wired to the In terminals. The other (50G) uses the mathematical summation of the three phase CT's.

My question is this:
Assuming the core-balance CTs were connected at the residual point of a wye-secondary transformer and the phase CT's are connected downstream inside the main compartment of a switchgear, is it possible to get a 50N pickup without a 50G pickup?

To put my question another way...can I expect the mathematical summation of phase currents measured downstream of a single-line-to-ground fault to be non-zero?

 

[antigfk]

JMatt,

If you take a look at IEEE 242 "buff" book, table 4-2 you will see the following:

50G is "ground fault protection (relay current transformer in a system neutral circuit)" - such as a CT in series between a transformer neutral point and ground

50GS is "ground fault protection (relay CT is toroidal or ground sensor" - this is the core-balance CT that you are talking about. The core-balance CT wraps all three phases (and neutral if applicable).

50N is "ground fault protection (relay coil connected in residual CT circuit" - ie. the neutral circuit of three individual phase CTs wired in a WYE.

These are industry standard naming conventions that some manufacturers do not always comply with.

To answer your question... it depends on the situation. Say you have a transformer with 480V WYE with X0 bushing CT connected to a 50/51G device. Assume cables connect the 480V bushings to an MCC main breaker, and there are no motors- only heaters on the MCC, but each heater branch has 50/51GS. If a SLG fault occurs in the MCC, only the 50/51G would "see" a fault. Say you add a starter bucket and feed a motor with the MCC. The motor is a pontential source of ground fault. If there is a ground fault at the MCC, the utility and and motor (motor just for a few cycles) will feed the ground fault, and the 50/51GS at the motor starter and the 50/51G at the transformer would "see" ground fault current.

 

[dpc]

If both are measured at the same location, the values should be equal, in theory.

But the residual CT value is based on phasor summation of digitized values of raw CT input data that has probably already been filtered for dc offset and harmonics. The flux summation CT may respond differently to various levels of dc offset and harmonic currents. If the fault current is high enough to cause any of the CTs to saturate, then anything is possible.

Short answer - it's *possible* for one to operate but the not the other, even with identical settings. This would not be normal however.

"The more the universe seems comprehensible, the more it also seems pointless." -- Steven Weinberg

 

[jghrist]

Assuming the core-balance CTs were connected at the residual point of a wye-secondary transformer and the phase CT's are connected downstream inside the main compartment of a switchgear, is it possible to get a 50N pickup without a 50G pickup?

I'm not sure what you mean by "residual point of a wye-connected secondary". Core-balance CTs would be around all conductors of a circuit for sensitive ground fault protection. The relay would not respond to unbalanced load (as long as the neutral, if used, was also included within the CT and was not grounded after passing through the CT). In this case, you could get a 50N (using the core-balance CT) pickup without a 50G (sum of phases) pickup, but not the other way around.

If the core-balance CT was in the transformer neutral or around the phase conductors only, the relay would respond to unbalanced load and would see the same current as the sum of the phase currents. 50G would see the same current as 50N.

Another option for using the 50N would be to run the CT neutral through the In element. 50G and 50N would be the same in this case, but you would have some redundancy at no cost with the extra element. This redundancy is of little consequence IMHO, because if the relay summation doesn't work, there is a major problem with the relay anyway and the separate element probably won't work either.

 

[JMatt]

antigfk:

True SEL does exactly the opposite of IEEE 242. A 50N element is asserted sensing current from the "In" input. 50G is a mathematical summation.

all:
My hope is that both the 50N element AND the 50G will pickup for faults on the SWGR bus, but only the 50N element will pickup for faults within the transformer windings. In this way, I can decide to lockout the tie breaker for bus faults or ONLY lockout the main breaker for transormer faults.
This plan is foiled if load current on the unfaulted phases will cause 50G to pickup. antigfk makes a good point about motor contribution to the fault, but the time delay for tripping on a SLG on this HRG system will be long enough so that motor contribution won't be a problem.

 

[cranky108]

If I understand the question, it is possible to have zero sequence current on the X0 bushing, but not the sum of the three X1, X2, X3 bushings. It's simple, it's called a transformer fault.

If you also read the instruction manual, the In input can also be used for directional. In which case all your feeder relays could be connected to the transformer X0 bushing CT.

Or like in the 351R, the In is used as a more sensitive ground detection method.

 

[dpc]

If this is truly a high-resistance grounded system, it is not typical to use overcurrent devices for ground fault protection anyway, since primary current will generally be only one or two amps.

I'm still not clear on what your actual question is. Maybe a one-line sketch might help us to at least be answering the same question.

"The more the universe seems comprehensible, the more it also seems pointless." -- Steven Weinberg

 

[JMatt]

Parameter clarification: I meant to say that this is a low resistance grounded system.

cranky108 says:
"If I understand the question, it is possible to have zero sequence current on the X0 bushing, but not the sum of the three X1, X2, X3 bushings. It's simple, it's called a transformer fault."

This is certainly true on an unloaded system where your unfaulted phases are open circuits. My question is: what does the summation of the X1, X2 and X3 bushings look like if a SLG fault occurs inside the transformer while it is carrying load? The two unfaulted phases continue to carry current to the load, but are those two phase currents 180 deg. out-of-phase? If so, the 50G will not pickup, and my scheme will work. If those current vectors do not sum to zero, the 50G may pickup and erroneously indicate that the SLG fault detected is downstream of the phase CT's.

The load response is the key to whether I'll get a 50G pickup...and I haven't figured out how to calculate it.

 

[rbulsara]

Jmatt:
I presume your system load circuits have no neutrals (that is N is just grounded at the source transfromer) and there is no generator source on the load side.

(if the loads have netural, neutral CTs needs to be included along with phase CTs while summing to detect ground faults).


Now coming to your last question, as long as there is no phase grounded downstream of the CTs, all phase currents through those CTs shall sum to zero, as there is no path for any current to leave the circuit, even if there is a fault upstrem.

Ohterwise for any ground fault, all of downstream breakers with GF protection would open, which obviously is not the case.

Things get complicated with mutiple sources, but that is not your question.

 

[JMatt]

That is the kind of insightful affirmation that I was looking for! That makes perfect sense. Cheers, rbulsara.

 

[SJBatTCE]

I actually had this problem in a motor protection application. The 50G would pick up during starting, but there would be no 50N (core balance CT) current. The DC component of the phases would not balance in the calculation, but no return current was present, I assume because the generator absorbed it for starting. If you have a core balance CT for this type of application, don't use the 50G element as it is less reliable and requires a much higher setting to be selective. If you don't have a core balance CT you're stuck with using the 50G, but I guess it's better than nothing.