Eng-Tips is the largest engineering community on the Internet

Intelligent Work Forums for Engineering Professionals

Ground fault on delta-delta system (ungrounded) 3

Status
Not open for further replies.

budhiman

Electrical
Jan 26, 2016
60
0
0
US
I want to understand how to have ground fault detection on a water treatment plant (system is ungrounded). The system is fed by a delta connected secondary utility transformer. I understand the requirement one way to detect is using 59G (measure the voltage). My confusion is:

1. Can this is 59G be used with wye-wye connected PT's by measure the residual voltage (sum of line to neutral voltages) since the respective ground fault bus will have zero line to neutral voltage ? Since all the literature's say wye-broken delta.
2. One of the similar post of un-grounded fault detection mentions to use 67N to trip and 59G to supervise. Why is the67N required when 59G can also trip since there will be imbalance of voltages if ground fault occurs? Is this only required in transmission to limit the tripping in respective zones? In this case since the system is fed by a delta connected secondary utility transformer there will not be any ground fault on the primary.
3. Has anyone seen/worked on a system similar to this fed from delta connected secondary utility transformer feeding MV motors (delta-delta system) and the system requires ground fault detection? What did you do to address this?
4. Is there good article discussing the ungrounded fault detection system?
 
Replies continue below

Recommended for you

No such thing as an ungrounded system, though many are called that. What you have is a capacitively grounded system; it is phase capacitances to ground that determines where the balance point rests. Can result in nasty response to undesired conditions.

That said; wye-wye connected VTs (PTs) require the use of three voltage elements; in a modern numeric relay that's no particular concern. The wye-broken delta connection allows the use of a single voltage sensing element; definitely a concern back in the bad old days.

59G just tells you that there's a fault somewhere, anywhere. The 67N begins to give you a hope of isolating it within a portion of the system.

Make your life far easier and provide a grounding resistor. The 5-10A of high resistance grounded system allows the same continuity of service that used to be justification for capacitively grounded systems without any of the downsides. Given the equipment available today, the 1950's capacitively grounded system can't be justified; too risky, too uncontrolled. Ground it and limit the ground current to a manageable value. See Blackburn for a good description of what can go wrong in a capacitively grounded system.
 
davidbeach said:
Make your life far easier and provide a grounding resistor. The 5-10A of high resistance grounded system allows the same continuity of service that used to be justification for capacitively grounded systems without any of the downsides. Given the equipment available today, the 1950's capacitively grounded system can't be justified; too risky, too uncontrolled. Ground it and limit the ground current to a manageable value. See Blackburn for a good description of what can go wrong in a capacitively grounded system.

I strongly second this suggestion.

The two brands I've come across are Eaton and Post Glover. Both have good documentation.

 
You could add a separate grounding transformer to provide a HRG reference. You might also be able to resistively ground a corner of the delta. But a wye would be the "best" solution if starting from scratch.
 
You can put three small transformers connected in wye:delta on the delta system. You can leave the delta closed and ground the wye point through a suitable resistor to limit the ground fault current below the current rating of the transformer primaries, or you can ground the wye point and break the delta. Insert a suitable current limiting resistor in the broken delta.

Bill
--------------------
"Why not the best?"
Jimmy Carter
 
Let me add my favorites on the subject.
Donald Beeman, Industrial Power system Handbook, Mc Graw hill Book Company, Pages 971, 1955

J R Dunki-Jacobs, F J Shields, Conrad St.Pierre, Industrial Power system Grounding design Handbook, Pages 568

Dunki seems to be the best. Incidentally he was the second generation disciple (disciple's disciple) of legendary Beeman.
 
Does a water treatment plant have any areas that might be explosive? If so, that is the reason for having a delta system. They also are popular in mines, and hospitals, where explosive gases may be present.

One solution is to use a wye connected system, but use a switch for the transformer neutral, so that normally the system can be left ungrounded. But it can be grounded to find ground faults, during times when the explosion risk is determined to be zero.

Other than these special conditions, I would use a high impedance ground as David talked about above.
 
Even with a delta or ungrounded system, there will be current in case of LG fault. One manufacturer has come with a system claiming zero fault current with LG fault. He injects a current in to neutral opposite to fault current. It is claimed this is being used in Australia in areas prone to forest fire.
 
That is an interesting idea. If you come across any further information please share it with us.
Thanks prc.
Yours
Bill

Bill
--------------------
"Why not the best?"
Jimmy Carter
 
Are you sure that the system is ungrounded? This is the rule that I am subject to. I understand that the NEC is similar.
Ground fault detection was originally provided with ground detector lamps. There were also ground detector voltmeters with or in place of the lamps. Both these systems depended on an operator eventually noticing that one phase was grounded. Relays could be used with the lamps to alarm on a line to ground fault.
The high voltages and high frequencies sometimes associated with an arcing ground fault were very rare. However the possible results of a arcing ground fault were serious enough that ungrounded systems fell out of favour. All motors and transformers on a feeder experiencing an arcing ground fault may be subject to severe over voltages and the burnout of multiple motors was common.
Consider a resistance grounded system to limit ground fault currents to about 5 Amps. Three 2.5 KVA, 480 Volt transformers and a suitable resistor will be adequate for a 5 Amp resistance ground on a 480 Volt circuit.

Bill
--------------------
"Why not the best?"
Jimmy Carter
 
FWIW and to follow-up Cranky108's post for a ground resistor that gets added in for tracing....

I've seen an entire system with low resistance grounded delta-wye transformers where the grounding resistor had another resistor, I think of equal size, in parallel, and with a contactor. At the main MCC rooms, the ground fault detection had a switch that introduced a 1 sec pulse timer relay to drive that contactor. Was very useful as you always had a ground resistor installed but the special mode made the actual ground fault pulse. Helped with the electricians putting amp clamps on the tens or even hundreds of cables to trace as the pulse was quite obvious on that one cable.

But, in this case with delta/delta and the capacitive ground as mentioned, not sure that there would be enough ground fault current flowing to trace that way. Still, perhaps in combination with the delta corner grounded resistor idea, this might enhance such a solution.
 
I see that a lot, HenryOhm. It used to be called a "Thumper" circuit.

Bill
--------------------
"Why not the best?"
Jimmy Carter
 
Status
Not open for further replies.
Back
Top