AC de-combiner and code issues

[BandGap1.1eV]

I've got an interesting scenario that I've yet to encounter before. A large commercial facility was made "solar-ready" by locating sub panels throughout the facility with enough theoretical space to land the output of an inverter in each one. The potential rooftop array can support say, 4 inverters worth of PV modules.

For utility and incentive program purposes, I need to combine the output of the inverters so I can meter it and provide an open blade disconnect at ground level. The client then wants to de-combine the outputs with a dedicated PV distribution panel that then ties individual output breakers into the individual sub-panels located throughout the facility. The sub-panels feed other loads, and all of them are eventually fed by a 3000A switchgear.

Aside from the backfed breakers not being at the end of the bus, this feels like a violation of the NEC, but I can't quite determine where to start. Since there is no feasible way to throttle the output of the inverters to match the loads in each sub-panel, and there's no way to direct electrons once they get into the AC PV distribution panel, I can envision a scenario where breakers start tripping in a cascading manner because the inverters start ganging up on loads, tripping that particular breaker in the PV distribution panel, then going to the next, then the next, then the next, etc.

Then if the voltage drop values of the feeders between the sub-panels and the main gear are not dead nuts across the board, could that cause issues too with one inverter sub-panel feeding another?

I've designed plenty of systems that interconnect to individual sub panels across a complex, but never had to combine them first, then try to uncombine them.

Again, it feels wrong. Just not sure where to start. And for what it's worth, the inverter sizes are not uniform.

Snippet of part of the drawing attached.

 

[jaggedben]

Let's start with my offering condolences that you apparently have a client that supremely doesn't get it and you can't reason with them or walk away from it. I mean, you can start with it being stupid, stupid, stupid, I don't think I have to really convince you of that. So what else are you looking for? A code reference? I would say it violates the rules for parallel conductors. Because you are paralleling the inverters output to the service, that's what you'd be doing. Sorry I'm blanking on the code reference right now, but it's in chapter 3 somewhere.

 

[ggunn]

You cannot unsalt the soup. Maybe, and I do mean maybe, you could use a six pole switch to interrupt all three circuits with one handle throw but keep them separate? Eaton makes them.

 

[BandGap1.1eV]

Let's start with my offering condolences that you apparently have a client that supremely doesn't get it and you can't reason with them or walk away from it. I mean, you can start with it being stupid, stupid, stupid, I don't think I have to really convince you of that. So what else are you looking for? A code reference? I would say it violates the rules for parallel conductors. Because you are paralleling the inverters output to the service, that's what you'd be doing. Sorry I'm blanking on the code reference right now, but it's in chapter 3 somewhere.

I've explained to them how foolish it is and showed them plenty of analogous systems with the "right" way to accomplish this. They keep coming back to "but this is solar ready". The GC is insisting that they can't get the inverter conductors back to the main gear because of the convoluted routing required. I pointed directly to the sub-panels and said "you already have the routing figured out with the feeders for the sub-panel, just follow the same path" to which they came back with "but this is already solar ready".

The facepalms have given me a mild concussion.

I just need to find a code reference that indicates this is a violation. As stubborn as they are, I think they will understand "violation".

 

[wwhitney]

How about 240.8, fuses or circuit breakers in parallel?

What do they say to "the "solar readiness" was designed wrong"?

Cheers, Wayne

 

[BandGap1.1eV]

How about 240.8, fuses or circuit breakers in parallel?

What do they say to "the "solar readiness" was designed wrong"?

Cheers, Wayne

That could work too. Parallel feeders (310.10) and parallel breakers. They are violating both. I've already told them about 310.10. If they come back with "but this is solar ready" I may have to walk. Not worth dealing with this much ignorance.

 

[winnie]

Your customer has a system, and they probably spend a bunch of money some years ago to future proof. Now they are being told that this system has no value. Of course they are going to push back and try to protect the value of their sunk costs. This is a perfect 'Dale Carnegie' situation. You have to lead them to making a good choice rather than triggering their defensiveness by telling them that they are making a bad choice, and in doing so you need to listen to them and perhaps find a useful middle way.

"Clearly, this 'solar ready' system was ahead of its time. It was designed so that inverters at separate parts of the facility could be conveniently connected to subpanels distributed throughout the facility, making for a more efficient system. Unfortunately this system is not compatible with the metering requirements for the various incentives that make a solar installation most attractive. Additionally the utility requires a PV disconnect which this cleanly distributed system makes difficult."

Ok, what happens if you install a PV system per the original intent? Well you don't have a single metering point for incentives, and you also have to somehow meet the utility disconnect requirements.

Does the utility have a way to have distributed metering? What does the ROI do _without_ this metering and the incentives?

Are there alternatives to an 'open blade disconnect' which will satisfy the utility? What about panel level rapid shutdown, which is arguably safer than having an XXX volt DC disconnect?

If the only approved approach is to bring all of the PV to a single location for metering and disconnection, then will only add cost, inefficiency and likely code violations to separate things out again.

-Jon

 

[gadfly56]

It sounds like they want to power certain of their loads directly with the PV panels. We know from discussions here that this is possible, but I don't see how they can have a metered connection, and do this as well, at this scale. Do they not understand that to get credit for their PV, they need to be grid-tied? And at that point, electrons are completely fungible; no such thing as "our" electrons and "their" electrons.

 

[BandGap1.1eV]

"Clearly, this 'solar ready' system was ahead of its time. It was designed so that inverters at separate parts of the facility could be conveniently connected to subpanels distributed throughout the facility, making for a more efficient system. Unfortunately this system is not compatible with the metering requirements for the various incentives that make a solar installation most attractive. Additionally the utility requires a PV disconnect which this cleanly distributed system makes difficult."

This configuration is 100% feasible if we were under the previous incentive program that allowed for customer owned metering. However we're ~3 years past that. Current program only allows a single generating unit to be installed on one building, otherwise you could game the system and take advantage of higher incentive rates for smaller systems and just have 2, 5, 10, however many rooftop generating units on a single building.

"But it's solar ready"

 

[jaggedben]

They're SOL. And they want to kill the messenger. I feel for you.

 

[winnie]

This configuration is 100% feasible if we were under the previous incentive program that allowed for customer owned metering. However we're ~3 years past that. Current program only allows a single generating unit to be installed on one building, otherwise you could game the system and take advantage of higher incentive rates for smaller systems and just have 2, 5, 10, however many rooftop generating units on a single building.

"But it's solar ready"

That seems to me to be an accounting issue. If 4 separate PV systems with 4 separate meters, all at a single building were treated as a single system with the total of the 4 meters treated as a single output, then there would be no gaming of the system.

Don't get me wrong, in your original post you were asking about the customer's desire to somehow graft a single point meter and disconnect into an existing distributed system. I think you are rightly horrified by that approach; even if you could somehow work around the code issues such a system would be more expensive and less efficient than simply doing the single point install.

My questions are:
Is there any way that you could have the utility do the multiple meters while counting as a single meter?

Could CTs be used so that multiple separate PV systems 'report ' to a single meter?

If you do the installation as originally planned and forego the metered incentives how does that change ROI?

Jon

 

[ggunn]

That could work too. Parallel feeders (310.10) and parallel breakers. They are violating both. I've already told them about 310.10. If they come back with "but this is solar ready" I may have to walk. Not worth dealing with this much ignorance.

You might try "just because something was code compliant when it was installed years ago doesn't necessarily mean it is compliant today. If I install something today it must be compliant to the code as it is today, and there's no way for me to do what you want with what is there."

Not that I think what they are asking you to do would ever have been compliant, but still...

 

[winnie]

You might try "just because something was code compliant when it was installed years ago doesn't necessarily mean it is compliant today. If I install something today it must be compliant to the code as it is today, and there's no way for me to do what you want with what is there."

Not that I think what they are asking you to do would ever have been compliant, but still...

As described the original system was code compliant and solar could still be installed using the original system in an NEC compliant fashion.

The kicker is the metering requirements for the incentive program. The crazy nonsense installation the customer is asking for is to somehow meet the incentive requirements and use the old system.

Jon

 

[BandGap1.1eV]

That seems to me to be an accounting issue. If 4 separate PV systems with 4 separate meters, all at a single building were treated as a single system with the total of the 4 meters treated as a single output, then there would be no gaming of the system.

Don't get me wrong, in your original post you were asking about the customer's desire to somehow graft a single point meter and disconnect into an existing distributed system. I think you are rightly horrified by that approach; even if you could somehow work around the code issues such a system would be more expensive and less efficient than simply doing the single point install.

My questions are:
Is there any way that you could have the utility do the multiple meters while counting as a single meter?

Could CTs be used so that multiple separate PV systems 'report ' to a single meter?

If you do the installation as originally planned and forego the metered incentives how does that change ROI?

Jon

ROI in this case depends on third party ownership since this is a public facility.

The utility in this case is in the stone ages. They are unconsciously regressive with their Distributed Generation policy. If Lenny from of Mice And Men was a businessman, he would be the CEO of this utility. On a different project the utility tried to charge the end client twice for the interconnection equipment (transformer, load break, metering, poles, etc), once from the DG group and once from the New Service Work Request Dept. We called them on it and the best they came back with was "Whoops".

Asking them for an exception to their metering policy has a less chance of success than squeezing milk from a rock.

 

[pv_n00b]

This is such an excellent case that proves that:

• You can give the customer exactly what they ask for but not what they need and blame them for not taking your advice when it fails.
• You can make an NEC-compliant system that fails to operate because the NEC ensures an electrical system is safe but does nothing to make sure it is operational.