Ungrounded systems

[Catman800]

We have a large campus power system that consists of 3 substation transformers that are considered the service point of the system. Primary line voltage is 34.5kv and is transformed down to 4160v and distributed to many 4160/480 volt outdoor transformers with adjacent load centers. These transformers range from 500 to 1500 KVA wye/ wye connected with HO and XO ungrounded (floating). From these load centers, feeders are brought into various buildings such as office, warehouses and vehicle maintenance buildings. The load centers have ground fault light indication.
I understand that this installation does not meet the requirements of 250.20, 250.21… We are trying to remedy the issue and I am looking for help. The wiring method from the load center is RGC with three phase conductors and no neutral and no EGC, We derive a neutral from a transformer inside the building to pick up 277 volt loads where needed. Article 250.97 is followed, for the most part, for the grounding and bonding part of things. Panels inside the buildings are 30 plus years old and need of some upgrading, which leads me to a question on 250.24c. My understanding is that a neutral conductor has to be run with the phase conductors and terminated with the service equipment with the main bonding jumper even if the neutral conductor is not used. This is for services, does this apply to feeders. Next question is how unsafe ungrounded systems really are? Is there a shock hazard when equipment becomes unintentionally grounded? I have read and understand what can happen when one of the other phases starts to fault are or does fault. Any help from the experts would be greatly appreciated on this fixing this installation.

 

[augie47]

I may be wrong, but I don't see your system as being in violation so long as XO is not connected to a grounding electrode at the transformer.
If, when grounded, the voltage would not exceed 150 volts you would violate 250.20(B)(1) but thats not your case, nor is (B)(2)or (B)(3).

 

[Dennis Alwon]

A neutral does not need to be run with feeders if the panel does not require a neutral. As you stated it is for services not feeders. However if you decide to run a neutral for future uses then it must be size at least as large as the EGC that would be required.

 

[augie47]

Catman,
Back to my post for a moment.
If I'm reading 250.21(B)(1) correctly it says a system must be grounded where the system can be grounded so the max voltage does not exceed 150 volts to ground.
As you state, yours would be 277 so the "must be" grounded does not apply.
IMHO

 

[erickench]

In an ungrounded system the OCPD would trip if you have a double line-to-ground fault otherwise the ground fault indication light would alert you if there is a single line-to-ground fault. Is there someone on duty observing the ground fault indicator light? This is the only way that you would know if there is a problem and it has to be corrected.

 

[Catman800]

In an ungrounded system the OCPD would trip if you have a double line-to-ground fault otherwise the ground fault indication light would alert you if there is a single line-to-ground fault. Is there someone on duty observing the ground fault indicator light? This is the only way that you would know if there is a problem and it has to be corrected.

No to the first question. We have recently established rounds once a day to check these lights. One of the load centers has no indication at all.
I'm trying to understand 250.20 - 250.21 and augie47 brings up a good point with
250.20 (B) (1). I still feel this installation is an illegal one, however, more importantly, I feel this is an unsafe and hazardous one. Does anyone have any first hand experience with these systems, and any examples and what can happen within these systems?

 

[augie47]

In my experience, the most common "danger" comes from inadequate grounding/bonding and/or failure to monitor the system.
If one phase goes to ground, you obviously have full 480v potential to ground.
You have ground indication and you are monitoring the system to see such faults.
The system has the advantage of allowing you to trouble-shoot such problems without loss of power.

 

[Catman800]

[The system has the advantage of allowing you to trouble-shoot such problems without loss of power.[/QUOTE]

Is this necessary in an office buildings, vehicle maintenance garage, and warehouses?

 

[iwire]

The system has the advantage of allowing you to trouble-shoot such problems without loss of power.

Is this necessary in an office buildings, vehicle maintenance garage, and warehouses?

Not in my opinion and hot work is becoming less of an option anyway. It is allowed for troubleshooting but you would still have wear the appropriate type of PPE for the circuit you would be working on.

 

[erickench]

NEC 250.21(A) gives a list of systems that are permitted to be grounded but are not required to be grounded. If your system is on this list then you could ground the system save your worries.

 

[Catman800]

NEC 250.21(A) gives a list of systems that are permitted to be grounded but are not required to be grounded. If your system is on this list then you could ground the system save your worries.

I do not believe our systems fall under any of those conditions. XO on the secondary side of the transformer is open. It?s there, but nothing tied to it. I would love to ground them but I believe there are more problems tied to just grounding them and where do you stop. I feel that you have to start at the source, (i.e.) the transformer and work from there.
issues
1: We have about 20 of these systems
2: Once you leave the 42 circuit outdoor load center, multiple feeds enter one building. 225.33
3: Two different load centers feed the same building with multiple feeds into the building. 225.33 225.34
5: Load centers are only 3 wire load centers
6: RGC pops into the bottom of the one or two of the load centers, out of the dirt and is not bonded to the load center
7: Load centers are 30 plus years old and are rusting

All these issues may be nothing at all to worry about, and I just may be trying to make things better than they really have to be. I'll try to get some pictures together to show for a better understanding.

 

[kwired]

I do not believe our systems fall under any of those conditions. XO on the secondary side of the transformer is open. It?s there, but nothing tied to it. I would love to ground them but I believe there are more problems tied to just grounding them and where do you stop. I feel that you have to start at the source, (i.e.) the transformer and work from there.
issues
1: We have about 20 of these systems
2: Once you leave the 42 circuit outdoor load center, multiple feeds enter one building. 225.33
3: Two different load centers feed the same building with multiple feeds into the building. 225.33 225.34
5: Load centers are only 3 wire load centers
6: RGC pops into the bottom of the one or two of the load centers, out of the dirt and is not bonded to the load center
7: Load centers are 30 plus years old and are rusting

All these issues may be nothing at all to worry about, and I just may be trying to make things better than they really have to be. I'll try to get some pictures together to show for a better understanding.

You either have a grounded system with an additional grounding conductor run everywhere or an ungrounded system with your current ground fault monitoring. You can't have both or a combination of both. If you ground the XO or even a phase at the transformer you have a ground reference and your entire system will have this reference - it will also mess up the ground fault detection system - you will need grounding conductors in place of the ground fault dectection system.

If you separately derive a system at a structure like 120/208 then you have a new system and certainly can ground that system, but the supply is still ungrounded.

 

[SG-1]

The worse case is a re-striking ground fault.

A High Resistance Ground or Low Resistance Grounding scheme may be an answer to your concerns. These can be purchased as seperate units and come in low & medium voltage models.