How to hunt for a grounded wye-delta transformer?

[bacon4life]

We recently had a 7000 amp line ground fault one feeder that also caused an adjacent feeder to trip on about 1500 amps of 3I0 current (500 amps on each phase). One possible cause would be grounded wye-delta connected transformer bank somewhere on the adjacent feeder. Other than inspecting every transformer on the feeder, are there any other ways to identify which portion of the feeder might have the incorrectly connected transformer bank? Or are there other ways for ground fault current to show up on the phase conductors of an different circuit. This is a 12 kV grounded wye distribution system.

Feeder 4 fault currents:
A-7000 <0 degrees
B-0
C-0

Feeder 1 fault currents:
A-500 <180 degrees
B-500 <180 degrees
C-500 <180 degrees

 

[HamburgerHelper]

Mutual inductance?

 

[waross]

Hello bacon4life;
This is one of my pet subjects.
If the fault caused the source voltage to drop on one or more phases, or caused a phase angle displacement on one or more phases any wye/delta bank fed from that source will have issues.
The wye delta bank has the ability to transfer power from the healthy phases to try to "heal" the problem phase. The current will be limited by 3 times the impedance rating of the transformer bank.
If you have any voltage regulators on the circuit you may try dropping the voltage of one phase by a few percent and checking for disproportionate neutral current.
Look for a bank with one fused cutout dropped out but no complaints from the customer. When one phase is lost, the wye:delta bank acts as an open delta connection to feed the load while the open delta back feeds the de-energized phase. Once a fuse clears a phase, the back feed stops and the transformer continues to serve its load as an open delta bank.
You may get help from your billing department. They may be able to identify customers with three phase services at 240 Volts.


Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[magoo2]

If you had a grounded-wye delta bank, then I'd expect to blow a tap fuse before tripping a circuit breaker unless you don't fuse your taps. If you don't fuse your taps, then you probably will have to if this occurs again.

Another possibility is if you have a sufficient amount of air conditioning load and the ground fault on the other feeder is fairly close to the substation. Close-in faults produce high voltage dips. With a high voltage dip at the substation bus, you'll also dip that same phase on the adjacent feeder. Air conditioning compressors will stall under low voltage and if you have enough units connected to that phase, you'll think you have a ground fault condition. I normally estimate the locked rotor current for a/c units to be between 30 and 40 amps per ton. It would only take 50 units at 30 A/ton to produce your 1500 A of 3I0 current. If you don't have much a/c load, then you can quickly dismiss this idea.

 

[davidbeach]

Pick a section of the circuit, open one phase, ground that phase on the load side of the open, let sit 20 minutes, remove ground, close switch, see if you get a call about a burned out customer transformer. If no burnt out transformer, repeat elsewhere on the feeder. ;-) (actually, don't really do that)

Probably a 208Y/120V customer that bought a piece of 480V equipment, and got a transformer that has those two voltages on it.

 

[waross]

Following David's suggestion, open one phase and check for feedback voltage on that phase.
Also, if your field crews have a high voltage clamp meter, when one phase is opened, the current will increase on the other two phases as they pick up the back-fed load.
Thanks for the mental "nudge" David.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[davidbeach]

My somewhat tongue-in-cheek answer above is based on an incident we had a few years back. There was a ground fault on an underground tap, which blew a fuse. Customer beyond the fuse was served at 208Y/120V through a grounded wye/grounded wye service transformer. That should have be it; they should have just called up and complained about part of their service being out. But instead, they called to make a claim regarding their transformer we "burned out". Turns out they had bought some piece of 480V equipment and hired Low Ball Electrical to hook it up; contractor found a 480 delta/208 wye transformer; hooked it up, including grounding the wye point of the 208V winding. I figured their #6 neutral/ground bond had several hundred amps on it from the time our fuse blew until that bond functioned as a fuse. That several hundred amps were also in the transformer windings but not enough to trip the breaker in time to protect the bond.

 

[bacon4life]

Magoo- we have a fuse saving damaging scheme with an instantaneous element set for 3I0 = 1400A. In all the event reports of LG faults at this substation, a tap fuse has cleared the fault while the 5 cycle distribution breaker was still in the process of opening.
I assume you mean single phase air conditioning. This is a commercial/light industrial area with mostly 3 phase load. There are some refrigerated warehouses, but here in western Washington there is very little air conditioning.

Waross-Do I need find and use the zero sequence impedance of the bank as well? Or is the zero and positive sequence impedance usually about the same? When I tried to draw a the sequence network, it seemed like the current was limited by 3*Z0, but simulating the fault in Aspen Oneliner seems to give me a number three times as high. Which value is correct for the example of a 300 kVA bank made up of single phase 100 kVA 2% impedance transformers?
A) 3*I0 = 3*300 kVA/sqrt(3)/12.47 kV/(3*2%) = 693 Amps
B) 3*I0 = 3*300 kVA/sqrt(3)/12.47 kV/2% = 2083 Amps

Typically if a fuse blows, the first thing we do is open all three phases if there are any three phase loads downstream to prevent damage from single phasing. Would opening it just long enough to take a voltage reading (2 minutes) short enough to ensure not damage occurs? I don't quite understand what the ammeter would do. If we opened one phase, 3 phase motor without electronic phase loss protection would draw more current on the remaining phases, whether or not there is a back feeding transformer.

Unfortunately we only have 3 phase LTC regulation. Perhaps we could put up single phase cap banks to raise the voltage on just one phase? Would motor and VFDs also shift load to the higher voltage phase?

 

[waross]

Hi. Yes, motors will also feed the missing phase. The wye:delta banks that used to challenge me were mostly residential, single phase loads. We had two industrial plants with wye:delta services. Our tiny system was protected by fuses, we didn't have the luxury of relay protection. After an outage, the generators would not take the block loading of the entire load. The operators would go out to the transformer bank and pull the feeder fuses. When they had the generators up and sync'd they would energize the feeders, one phase at a time. Now these were pole top fused cutouts. With one operator holding the flashlight, the other operator would use the "Hot Stick" to close the first phase. One third of the lights would come on, and the single energized secondary of the wye:delta bank would back feed both of the other phases in series. The back fed voltages would be approximately 50% of normal voltage. Then the operator would get his Hot Stick disengaged from the first cutout and inserted in the second fuse holder. With two phases energized, the wye:delta banks would back feed nominally full voltage, but in fact, due to line impedance and transformer regulation, the voltage would be a few percent down.
You can imagine what that would do to all the refrigerators and freezers trying to start on the second and third phase. Complaints about burned out refrigerators were common.
A number of times I was in my room with a single incandescent lamp during an outage. The lamp would come on with about half voltage. Then it would jump up to almost full brilliance. Finally the third phase would come on and the lamp would get a little brighter.
Impedance: The books say that the zero sequence impedance will be three times the positive sequence impedance.
However, we basically have an open delta feeding a single winding. Given the phase relationships of the two transformer currents in the open delta, the virtual transformer has the same voltage drop characteristics as either one of the windings on its own. In this case where two transformers act as an open delta to feed a third transformer, I suspect that the back feed impedance may be twice the positive sequence impedance rather than three times.
I may be wrong.

Bill
--------------------
"Why not the best?"
Jimmy Carter