Fence Around PV Site

Status
Not open for further replies.
B

bwat

EE
Location
NC
Occupation
EE
For the metal fence around a PV site, what section of NEC covers the size of the of the conductor that should be used to bond the fence to the metal frame of the PV racks?

I understand the difference between GEC and EGC, and I want to say this should follow table 250.122 (EGC), but Part VII of 250 which covers DC doesn't comment much on this and really only mentions GEC in 250.166(B).

I'll take NEC interpretations or what you have seen AHJ's think on this.
 
I don’t have time to give you code book section references but I can tell you that there are some rules toward the end of article 250 in the sections dealing with installations over 1000 votes that pertain to requirements to ground a fence and then there are rules generally requiring metal parts that are likely to become energized to be connected.

However it is not a blanket and general rule that all PV system fences need to be grounded or bonded into the grounded system. It’s not always even safer to do so. A trade off that occurs when you bond metal pieces into the system is that you create a scenario where they will become energized if there is a fault, such as at the utility, that causes the neutral voltage to spike. An electrified fence in this scenario could be more hazardous, especially if you just bond the fence and don’t have an underground ground wire grid like they do at substations.

I’ve seen inspectors that are originally requesting it became convinced not to request it when this has been explained by engineers.

In summary, it’s not a blanket rule that you need to do this in a PV application, though there may be a few cases where it’s needed.
 
tom baker said:
Is the fence likely to become energized?
How big is the site?
Click to expand...

Likely? I would say no, but there are a few places where the fence isn’t very far from the racks. Definitely less than 10’.

Possible? It would be hard, but I’d have to say yes. And so I’d like to bond all the non-current carrying metal parts together even though there’s nothing but a slim chance anything could happen with the fence.

This one in particular is 5MW
 
bwat said:
Likely? I would say no, but there are a few places where the fence isn’t very far from the racks. Definitely less than 10’.

Possible? It would be hard, but I’d have to say yes. And so I’d like to bond all the non-current carrying metal parts together even though there’s nothing but a slim chance anything could happen with the fence.

This one in particular is 5MW
Click to expand...
Seems pointless and a waste to me, to be honest.
 
Be careful when looking at 'substation' fence grounding requirements. Most of the references you find will be for locations with exposed conductors in open air.

There are very few requirements for fencing surrounding metal enclosed equipment.
 
electrofelon said:
Seems pointless and a waste to me, to be honest.
Click to expand...
Fair enough. Probably unnecessary and a little overkill, but I'd rather err on that side if it's uncertain.


Fred B said:
Public or private utility? The utility side has different requirements not in the NEC.
Click to expand...
Private



jim dungar said:
Be careful when looking at 'substation' fence grounding requirements. Most of the references you find will be for locations with exposed conductors in open air.

There are very few requirements for fencing surrounding metal enclosed equipment.
Click to expand...
That's what I was finding. I'm familiar with the sections in 250 Part X for over 1kV, and there are MV parts of this site that would apply to that and where we'd apply 250.194(A), but I didn't have a good answer for the lower voltage areas.
 
I don’t know your application but it’s not necessarily going to be safer to bond the entire fence. This especially if you are just going to bond it but not create a grounding grid in the earth around it,
 
5MW system sounds like commercial generation, or code language (large scale photovoltaic electric power production facility) if so, OSHA may have criteria related to workplace safely that might influence fence placement and bonding. OSHA references approach distances at voltages, and does reference fences in relation to electrical substations.
Also for commercial generation criteria included in (NEC 2017) "691.11 Fence Grounding. Fence grounding requirements and details shall be included in the documentation required in 691.6."
That refers to engineered design. From installation perspective, bottom line is if it's not in the plans, you dont do it.

If you are in the process of planning, I would consider the bonding requirements related to a pool, (I think probably the strictest for safety reasons) can give guidance as to likely to be energized or contact simultaneously with an energized source, but not as an enforcement criteria or to subplant any existing other requirements. If a fence is more than 5ft from a pool's inside edge bonding not necessary.
Bonding the fence unnecessarily can create a safety issue that otherwise wouldn't be present in the event that other components were to become energized as others pointed out. I can see it making entry into the fenced area for maintenance a potential shock hazard if bonded to an item that had become energized.
 
Down here in Fl those req’s are spec’d by the utility. I’ll try to attach an example.
 

Attachments

  • 32B9648E-00DF-4D72-8192-58476AF829C8.jpeg
    32B9648E-00DF-4D72-8192-58476AF829C8.jpeg
    547.7 KB · Views: 45
I often see specs/installation drawings for large natural gas / gas storage facilities, the perimter fence is often bonded to a GES/ring/rods along the entire perimeter - every few posts, and always seems to be a waste of money based on some boilerplate spec that should not apply.
 
nhee2 said:
I often see specs/installation drawings for large natural gas / gas storage facilities, the perimter fence is often bonded to a GES/ring/rods along the entire perimeter - every few posts, and always seems to be a waste of money based on some boilerplate spec that should not apply.
Click to expand...

Especially if they are copied from a successul project and/or someone with extensive experience, some of those boilerplate specs are terrifying to change when they can be perceived as anywhere close to safety related. I'll admit I've been in that situation many times on the design side where I'm 99% sure I don't need to do what we always spec out to do and it appears to be above and beyond code, but because I'm not 100%, and it could potentially be dangerous if I'm wrong, that same piece of the spec finds a way to stay in there..
 
bwat said:
Especially if they are copied from a successul project and/or someone with extensive experience, some of those boilerplate specs are terrifying to change when they can be perceived as anywhere close to safety related. I'll admit I've been in that situation many times on the design side where I'm 99% sure I don't need to do what we always spec out to do and it appears to be above and beyond code, but because I'm not 100%, and it could potentially be dangerous if I'm wrong, that same piece of the spec finds a way to stay in there..
Click to expand...
I'm sorta the opposite. Given the amount of BS there is out there, I am more likely to assume something is hogwash and happy to rid it from the earth as quickly as possible, particularly if it's something involving a green wire or a ground rod.
 
electrofelon said:
But that is the stupidest thing ever. The fence posts are already driven or cemented into the ground, why is a ground rod needed??? This ridiculousness needs to stop.
Click to expand...
I agree but looking at the drawing the only possible reason I can think of is they are trying to mitigate the voltage gradient on the public side of the fence IF a fault were to occur. So I guess from their point o view the ground rod is not really to ground the fence but to extend the gradient zone, hence the min distance spec.
 
5mw systems usually connect to the grid with pad mount xfmr’s and at the utility’s 35kv or less distribution systems. They don’t usually have a large scale substation with higher voltages and open air insulated equipment.

You don’t address voltage on the fence from faults or induction unless you have a ground grid in the vicinity of the fence to make the voltage between the hands and feet locations to be near zero. Connecting the fence to the grounded system without the ground grid could actually increase the chance of a shock event when a fault on a utility system occurs. Remember, in medium voltage multigrounded neutral systems, an open (not just a grounded) phase conductor somewhere else on the utility system can cause a voltage rise on the utility neutral. (Of course an open neutral conductor could do this too). So if the fence has little to no liklihood of becoming energized or inductively charged, bonding it when the standard may not require it might actually increase risk of a shock incident if that occurred. No one can fully assess your application here but it’s not always required and not always even beneficial.
 
Status
Not open for further replies.
Top