Commercial PV Design Questions

[mddorogi]

Any comments on this are appreciated. My company is going to be installing a 600 kW DC, 510 kW AC system using SolarEdge 480V inverters. There will be 4-120 kW inverters and one 30 kW inverter. A few points are not clear to me, or let's say I get differing statements about what's acceptable and to code. My Master Electrician and myself seem to view things the same way, but a couple of electrical PE's see it differently. Another electrical PE is in the middle somewhere.

The (simple) one-line is attached. The meter in the middle is the solar generation meter required by our utility DTE. It is CT rated, and they also require a visible knife-blade disconnect on each side of that CT cabinet. The questions are:
- is the combiner panels REQUIRED to have a main breaker (some electrical PE's say yes by NEC 705.30(A))
- is either one or both of the 800A switches REQUIRED to be fused?

I'm pretty convinced that my simplistic way of thinking about it is wrong, and I'd like to know why.
In my opinion, the 800A fuses in the dedicated solar switch in the switchgear provide overcurrent protection from the grid side, which is all you really need.
It seems that adding more fuses or a main breaker in the combiner panel, while not harmful, do not add anything other than expense and complexity.

The inverters have a maximum output fault current to 50% higher than the max continuous rating; for example 120 kW inverter is max continuous 144.3A and max output fault 213.3A. The inverters cannot produce anything like the fault current coming from the grid side. Each inverter has its own appropriately sized breaker, and adding a main doesn't seem to accomplish anything.

I admit that I don't understand fault conditions with inverters and what happens in real life when an inverter either gets fried or there is a short in the AC output circuit somewhere. I've never seen that happen.

 

[ggunn]

I can't tell if your system will be connected line side or load side through a backfed breaker. If it's through a BFB none of the discos need to be fused. If it is a line side interconnection one of the discos (usually the one closest to the point of interconnection) must be fused. The panel that combines the outputs of the inverters can be MLO as long as it conforms to 705.12(B)(3)(3).

 

[mddorogi]

Thanks. The system is back-fed through that 800A fused switch in the switchgear, the one showing the 800A class L fuses.

I believe the combiner panel does meet 705.12(B)(3)(3). The four 120A inverters have a max continuous output current of 144.3A and the 30 kW inverter has 36.25A. The total is then (4*144.3+36.25)*1.25) = 766.8A, which is less than the ampacity of the 800A bus bar.

 

[ggunn]

Thanks. The system is back-fed through that 800A fused switch in the switchgear, the one showing the 800A class L fuses.

I believe the combiner panel does meet 705.12(B)(3)(3). The four 120A inverters have a max continuous output current of 144.3A and the 30 kW inverter has 36.25A. The total is then (4*144.3+36.25)*1.25) = 766.8A, which is less than the ampacity of the 800A bus bar.

I'm assuming that the 800A fused switch in the MDP is on the line side of the main breaker; it is the only OCPD needed on the combined PV inverter output. Caveat: check with your AHJ; a couple of the ones we deal with require a fused disconnect where it isn't needed or required by the NEC.

Where did you find a 30kW SolarEdge 480/277V inverter? The smallest we can find is the SE80K-US which we can get SE to derate to 66.6kW if we have to. We haven't been able to get any that have only a single inverter unit.

 

[mddorogi]

Hmm. The 800A fused switch is on the load side of the main switch. Maybe I should see if the switch can be moved to line side. It is a 3000A service. Are you thinking about the 120% rule? Oddly, no inspectors around here seem to enforce that with switchgear like they do when back-feeding in a breaker panel, which never made sense to me.

SolarEdge makes 30 kW, 40 kW, 66 kW, 80 kW, 100 kW, 120 kW, we have used them all. The 30 kW have been very hard to get this past year, but we now have five or six in our warehouse set aside for projects. We even have a 20 kW 480V SolarEdge inverter, still new, but they don't make those any more. SolarEdge commercial inverter supply has been terrible, but getting better. I've made a lot of noise with them and have their ear.

 

[ggunn]

Hmm. The 800A fused switch is on the load side of the main switch. Maybe I should see if the switch can be moved to line side. It is a 3000A service. Are you thinking about the 120% rule? Oddly, no inspectors around here seem to enforce that with switchgear like they do when back-feeding in a breaker panel, which never made sense to me.

It's my turn to say hmm. I don't try to get around the 120% rule even in switchgear. Are you qualifying the busbars under 705.12(B)(3)(5) with "engineering supervision which includes available fault-current and busbar load calculations"?

 

[jaggedben]

... The meter in the middle is the solar generation meter required by our utility DTE. It is CT rated, and they also require a visible knife-blade disconnect on each side of that CT cabinet.

That's quite stupid they require a disconnect on both sides. Sorry that's happening to you.

Do neither the utility or the PEs understand UL1741?

The questions are:
- is the combiner panels REQUIRED to have a main breaker (some electrical PE's say yes by NEC 705.30(A))
- is either one or both of the 800A switches REQUIRED to be fused?

To agree with what's already been said... If you have customer owned OCPD upstream, then neither. Which is what I think you're saying you have.

If you didn't have upstream OCPD , then one or the other of your items would be needed, but not both. Theoretically either disconnect or a main breaker in the panel could serve as your NEC required service disconnect and overcurrent protection. As ggunn said, usually it would make sense that it be the one closest to the utility; some other rules might require it when other details are known.

In no case would both DTE discos be required by the NEC to be fused.

...

I believe the combiner panel does meet 705.12(B)(3)(3). The four 120A inverters have a max continuous output current of 144.3A and the 30 kW inverter has 36.25A. The total is then (4*144.3+36.25)*1.25) = 766.8A, which is less than the ampacity of the 800A bus bar.

It's actually the overcurrent device ratings that matter to 705.12(B)(3)(3), so rounding up could cause issues. Your 120kW are going to require 200A breakers and your 30kW is going to require a 50A. That adds up to 850 so you actually have a code problem here, unfortunately. Not a real safety problem, just a code problem. Maybe a PE stamp and an understanding AHJ get you around this, but that's up to them.

 

[ggunn]

That's quite stupid they require a disconnect on both sides. Sorry that's happening to you.

Do neither the utility or the PEs understand UL1741?

Several AHJ's I deal with require a disco on both sides of the PV meter for some 3 phase systems.

It's actually the overcurrent device ratings that matter to 705.12(B)(3)(3), so rounding up could cause issues. Your 120kW are going to require 200A breakers and your 30kW is going to require a 50A. That adds up to 850 so you actually have a code problem here, unfortunately. Not a real safety problem, just a code problem. Maybe a PE stamp and an understanding AHJ get you around this, but that's up to them.

Can you cite the reference? I always size my system OCPD and conductors by 125% of the total inverter current when there are multiple inverters. I have for years and it has never been questioned.

 

[jaggedben]

Several AHJ's I deal with require a disco on both sides of the PV meter.

My sympathies.

Can you cite the reference? I always size my system OCPD and conductors by 125% of the total inverter current when there are multiple inverters. I have for years and it has never been questioned.

The total system OCPD isn't what matters here. It's the sum of the individual inverter breakers he wants to put on the busbar. 705.12(B)(3)(3) (you brought it up).

 

[ggunn]

Can you cite the reference? I always size my system OCPD and conductors by 125% of the total inverter current when there are multiple inverters. I have for years and it has never been questioned.

Sorry, I didn't read your post closely enough. Yes, he has a code problem if his inverter OCPDs add up to more than the busbar rating of his combiner panel. Around here an inspector would likely fail it.

 

[mddorogi]

Sorry, I didn't read your post closely enough. Yes, he has a code problem if his inverter OCPDs add up to more than the busbar rating of his combiner panel. Around here an inspector would likely fail it.

Thx. I see that. Not aware of any inspectors around here who would catch or enforce that. "Code problem" is fair it seems since the issue is an artifact of available breakers. I could use 180A breakers instead of 200A breakers for the 120 kW 144.3A inverters, if there were 180A breakers available anyway. Then the code problem would disappear.

 

[mddorogi]

That's quite stupid they require a disconnect on both sides. Sorry that's happening to you.

Do neither the utility or the PEs understand UL1741?

The utility company (Detroit Edison)) certainly does not. I have been to the mat with them on this requirement, through the public service commission, and with multiple in-person meetings with DTE interconnection and service planning folks. They simply want a knife blade disconnect on each side of the CT cabinet so they can lock them out and safely install their CT meter. The fact that the switch in the switchgear can be locked is irrelevant to them, even though it's 10 feet away. The fact that the inverters shut off when there's no grid power means nothing to them; they just say that inverter functionality could fail. They simply want a visible knife blade switch on each side.

It is such a huge problem trying to find all of these switches. But, I've had to be careful as it is impossible to change that bureaucracy and I don't want to make enemies there.

 

[mddorogi]

It's my turn to say hmm. I don't try to get around the 120% rule even in switchgear. Are you qualifying the busbars under 705.12(B)(3)(5) with "engineering supervision which includes available fault-current and busbar load calculations"?

Actually I do have a PE that will go through that and sign off. I'd like to run through those calculations myself sometime. I have not done it. The 120% rule also has the "opposite end" requirement built into it, doesn't it? Do you move switches in switchgear around so that a solar back-feed switch is farthest away from the service switch? It is not that easy to move these things around.

 

[ggunn]

Actually I do have a PE that will go through that and sign off. I'd like to run through those calculations myself sometime. I have not done it. The 120% rule also has the "opposite end" requirement built into it, doesn't it? Do you move switches in switchgear around so that a solar back-feed switch is farthest away from the service switch? It is not that easy to move these things around.

I draw the SLDs with the opposite end note. It's mostly a non issue because most of our systems are line side connected.

 

[mddorogi]

I draw the SLDs with the opposite end note. It's mostly a non issue because most of our systems are line side connected.

How do you do that?. We have to be "behind the meter" per our utility regulations and therefore must connect on the load side of the utility CT cabinet, or somewhere else in the switchgear. There is already a spare 800A switch that we plan to use for solar. The grid power comes into the switchgear from below. The main switch and the spare switch (off) are circled in the image below. The utility transformer is in a pen right outside this switchgear, 1500 kVA 13.2 kV/480Y.

Guess we should look a the CT section of this, which I assume is right above the main switch. The utility does not normally like us connecting there, although we have got approval a couple times. The other problem is then we'd need a fused switch within 10' of that connection point, and there is no wall space for that.

 

[pv_n00b]

The (simple) one-line is attached. The meter in the middle is the solar generation meter required by our utility DTE. It is CT rated, and they also require a visible knife-blade disconnect on each side of that CT cabinet. The questions are:
- is the combiner panels REQUIRED to have a main breaker (some electrical PE's say yes by NEC 705.30(A))
- is either one or both of the 800A switches REQUIRED to be fused?

If you took the meter and meter isolation out just to make the circuit look less complicated and just looked at the 800A MLO panel being fed from an 800A feeder it is clearly code compliant. This is a standard design for commercial systems.
The meter isolation disconnects do not need to be fused, they don't even have to be load break rated. They are not providing any overload protection function in the circuit. They just isolate the meter so the utility tech feels safe.

 

[ggunn]

How do you do that?. We have to be "behind the meter" per our utility regulations and therefore must connect on the load side of the utility CT cabinet, or somewhere else in the switchgear. There is already a spare 800A switch that we plan to use for solar. The grid power comes into the switchgear from below. The main switch and the spare switch (off) are circled in the image below. The utility transformer is in a pen right outside this switchgear, 1500 kVA 13.2 kV/480Y.

Guess we should look a the CT section of this, which I assume is right above the main switch. The utility does not normally like us connecting there, although we have got approval a couple times. The other problem is then we'd need a fused switch within 10' of that connection point, and there is no wall space for that.

View attachment 2565418

Of course, every service has its quirks, and connecting on the line side of the main OCPD and behind the meter can be challenging. When I draw up a design I seldom have all the information I would need for a load analysis and available fault current calculation, so if the connection point is in the switchgear I follow 705.12(B)(3)(2) and add the note for the breaker to be at the other end of the busbar from the feed from the utility.

 

[wwhitney]

If you took the meter and meter isolation out just to make the circuit look less complicated and just looked at the 800A MLO panel being fed from an 800A feeder it is clearly code compliant.

It is clear that the 800A MLO busbar can't be overloaded, but it is not code compliant. It doesn't comply with any of the options in 705.12(B)(3). Which is a failure of imagination on the part of the CMP.

[I submitted a PI last cycle to allow 705.12(B)(3)(3) to ignore one breaker of the smallest size possible, as by a counting argument that can't overload the bus as it can't simultaneously be a load and a supply, but it was rejected as too complicated.]

The OP can work around this in a number of ways: use a larger MLO panel. Use a 800A MB panel and interconnect one of the PV inverters via a feeder tap on the 800A feeder. Use two smaller MB panels. Use 5 fused disconnects or enclosed circuit breakers and 5 feeder taps.

Cheers, Wayne

 

[pvgreeze]

It's actually the overcurrent device ratings that matter to 705.12(B)(3)(3), so rounding up could cause issues. Your 120kW are going to require 200A breakers and your 30kW is going to require a 50A. That adds up to 850 so you actually have a code problem here, unfortunately. Not a real safety problem, just a code problem. Maybe a PE stamp and an understanding AHJ get you around this, but that's up to them.

I'm confused...why is 705.12(B)(3)(3) being applied to the combiner panel sizing? Wouldn't that fall under 690, specifically 690.8(A)(1)(e)?

My understanding is that 705.12 is specifically regarding the point of interconnection, which is the 800A fused switch in this instance. 705.12(B)(3)(3) would be invoked in order to determine if the busbar (or any section of the busbar) could be theoretically overloaded by both the utility source (the 3000A main) and the PV source (the 800A fused DS). I always thought 705.12(B)(3)(3) existed as a way to prove that the busbar (presumably 3000A here) could not be overloaded; for example, if that 3000A switchboard has three 600A fused DSs and a 400A fused DS and you throw an 800A fused DS on there, then you are fine because "the sum of the ampere ratings of all [OCPDs]...both load and supply, excluding the rating of the [OCPD] protecting the busbar, shall not exceed the ampacity of the busbar."

705.12 should not be invoked for the combiner panel, as that specifically deals with the point of interconnection and not the PV OCPD sizing, which falls under 690. Unless I'm misunderstanding something...

 

[wwhitney]

I'm confused...why is 705.12(B)(3)(3) being applied to the combiner panel sizing? Wouldn't that fall under 690, specifically 690.8(A)(1)(e)?

690.8(A)(1)(e) applies to the circuit from the AC-side of a single inverter to the first OCPD. 705.12 applies to everything from that OCPD to the utility. [When the utility side of the OCPD is not a service conductor; if it is, 705.11 applies.] So 705.12 applies to the combiner panel. 705.12(B)(3)(3) was added sometime this century specifically to address combiner panels.

Basically, from the point of view of article 705, each inverter is an electric power production source. The combiner panel itself is not an electric power production source.

Cheers, Wayne