220.87 - Determining Existing Loads... but with PV

[MazzEE]

If I'm adding a load to an existing building, I understand the getting the peak demand from 1-year's worth of utility data or 30-day metering requirements to prove additional load(s) are acceptable, however with this new exception added in NEC 2020, what then happens when you add new load(s) to a building with PV (renewable energy system)?
The exception negates this method (which makes sense), but then what do you do? Take the output of the PV system and add it to the Calculated Loads as if designing from scratch, then add your new loads in? (I may be missing something, but any direction would close this loop I'm stuck in)

 

[tortuga]

In that case I use the standard calculation.

 

[ramsy]

2020 NFPA-70 220.87
Exception: if the feeder or service has a renewable energy system (i.e., solar photovoltaic or wind electric) or employs any form of peak load shaving, this calculation method shall not be permitted

That disqualifying exception only applies to 220.87(1) in the 2023 NFPA-70, not to (2) & (3).

If you can't find a local AHJ planner who allows the 2023 language, AHJ's may be petitioned for a letter ruling.

 

[jaggedben]

That exception to the exception is too broadly written; if the demand data bypasses or accounts for the renewable energy system, you should still be able to use the data you have. (Also what if you're not using the exception?) But until they fix it you are dependent on the reasonableness of an AHJ, which is to say you may be SOL and have to use the standard calc like tortuga said.

As more and more facilities start to have their own data streams and not be reliant on the utility as the only source of consumption data, this section could really use an update.

 

[wwhitney]

As more and more facilities start to have their own data streams and not be reliant on the utility as the only source of consumption data, this section could really use an update.

I agree 220.87 needs some work, but nothing in the existing wording privileges utility data over other data sources.

Cheers, Wayne

 

[jaggedben]

I agree 220.87 needs some work, but nothing in the existing wording privileges utility data over other data sources.

Cheers, Wayne

Well, the bit about renewable sources that the OP quoted sort of does exactly that, at least indirectly. It assumes that demand data is from a utility meter and that the renewable generator is on the customer side. Both those assumptions are very unwarranted.

I think what this section needs, honestly, is an informational note. And remove that discriminatory language about renewables.

 

[tortuga]

Yeah renewable should be changed to co-generation, for example two identical log processing mills with the same co-generation first one is gas to steam co-generation the other has a waste wood to steam co-gen plant. The wood is considered 'renewable'.

 

[wwhitney]

Well, the bit about renewable sources that the OP quoted sort of does exactly that, at least indirectly. It assumes that demand data is from a utility meter and that the renewable generator is on the customer side. Both those assumptions are very unwarranted.

Not following you, there is no reference to the recording ammeter location (indeed the discussion of feeders suggests it may be on the customer side of the electric meter) or who owns it.

I do agree that the language prohibiting renewable sources should be replaced with language that basically says "if there are interconnected sources beside the primary source, sufficient data must be recorded to compute the feeder or service demand that would have resulted had all the interconnected sources been off."

Which I guess just means for a feeder with only two ends, you need to record all the power going in at the end towards the primary source (doesn't matter whether it actually comes from the primary source), and to record all the power put out by any interconnected sources attached to the other end.

Add an ESS on the other end, and it gets trickier--if your ESS can charge from say PV or the grid (admittedly not super common), and you do your 30 days during the summer when the PV is used exclusively to charge the ESS, then your data won't show any of the possible charging load on your feeder.

I guess you could add language requiring you to remove all ESS charging load from the observed demand data, and then add back in the maximum possible primary source ESS charging load. But you can see that now the correct wording is pretty complicated.

Cheers, Wayne

 

[Elect117]

There is no feasible way to calculate the existing demand with more than one source.

Metering demand is a game of precision. Cheap equipment, lack of calibration, and reads that are not on the same clock are just going to give you bad information.

The section would have to read,

"When calculating demand with more than one source, add the highest recorded value for each source that is interconnected with one year of data where each source is operating in optimal condition."

That number will never help on the calculation though.

 

[wwhitney]

There is no feasible way to calculate the existing demand with more than one source.

I don't follow your point at all. Obviously in some configurations it is easy. E.g. a "supply-side" PV connection will have all of the POCO and PV current flowing through the service disconnect for the loads. So for this configuration single point metering suffices.

Metering demand is a game of precision. Cheap equipment, lack of calibration, and reads that are not on the same clock are just going to give you bad information.

Agreed those are issues, but using a single recorder with multiple channels, of sufficient precision and accuracy, should allow reconstruction of the single source behavior from more complicated multiple source configurations.

Cheers, Wayne

 

[jaggedben]

Not following you, there is no reference to the recording ammeter location (indeed the discussion of feeders suggests it may be on the customer side of the electric meter) or who owns it.

I do agree that the language prohibiting renewable sources should be replaced with language that basically says "if there are interconnected sources beside the primary source, sufficient data must be recorded to compute the feeder or service demand that would have resulted had all the interconnected sources been off."

Which I guess just means for a feeder with only two ends, you need to record all the power going in at the end towards the primary source (doesn't matter whether it actually comes from the primary source), and to record all the power put out by any interconnected sources attached to the other end.

Add an ESS on the other end, and it gets trickier--if your ESS can charge from say PV or the grid (admittedly not super common), and you do your 30 days during the summer when the PV is used exclusively to charge the ESS, then your data won't show any of the possible charging load on your feeder.

I guess you could add language requiring you to remove all ESS charging load from the observed demand data, and then add back in the maximum possible primary source ESS charging load. But you can see that now the correct wording is pretty complicated.

Cheers, Wayne

Well, if you don't follow, at least this long reply of yours bolsters my point that 220.87 is behind the times.

 

[jaggedben]

There is no feasible way to calculate the existing demand with more than one source.

Metering demand is a game of precision. Cheap equipment, lack of calibration, and reads that are not on the same clock are just going to give you bad information.
...

Increasingly it's true that I can download 15 minute interval data for the residential systems install. I can get consumption, ESS and PV in a few minutes and put them into a spreadsheet. Consumption is actually calculated from all three sources/sinks, so that calc is actually done for me.

I don't see why 220.87 shouldn't be written to allow the use of such data. Especially for a dwelling.

 

[Elect117]

I will say this as an opinion because I am not perfectly familiar with the newer metering systems, but there are a bunch of reasons why I could not use that information when asked to provide stamped drawings or review plans for plan check.

1) How do I know if all the inverters are working?
2) Are the panels being kept clean in the summer?
3) The POC and the point of the meter installation must be properly documented. For example, parallel runs and which CTs are for which phases (not cross connected).
4) Buss connected CTs being properly installed, maintained, and tested.
5) Cycle data is on the same average, 5 min average, 3 min average, 15min average, etc.
6) Demand data is on the same clock

Maybe one day if they can get mains to also provide metering and feed that data through cellular then I would agree, but for now, I am not likely to believe the data being sent without a bunch of verification on the above.

220.87 is not just for the existing service or switchgear, it is used on upgrades as well. So to size your service entrance conductors on that data instead of the max of either is an issue.

 

[jaggedben]

I will say this as an opinion because I am not perfectly familiar with the newer metering systems, but there are a bunch of reasons why I could not use that information when asked to provide stamped drawings or review plans for plan check.

I hope you don't automatically trust utility metering. Also no disrespect, but why in general would anyone need your review?

1) How do I know if all the inverters are working?
2) Are the panels being kept clean in the summer?

None of that matters if the measurements are taken with CTs.

3) The POC and the point of the meter installation must be properly documented. For example, parallel runs and which CTs are for which phases (not cross connected).
4) Buss connected CTs being properly installed, maintained, and tested.
5) Cycle data is on the same average, 5 min average, 3 min average, 15min average, etc.
6) Demand data is on the same clock

Of course I agree that all such due diligence is required. But these aren't reasons you can't use the data.
On the systems I install everything is metered by the same device, same clock, same voltage reference, so your points (5) and (6) wouldn't be much concern. I think you may be misunderstanding something about what device(s) provides the data I was talking about.

Maybe one day if they can get mains to also provide metering and feed that data through cellular then I would agree, but for now, I am not likely to believe the data being sent without a bunch of verification on the above.

I don't follow. If you have data from a device that meets an appropriate standard for metering equipment, and the installation has been verified, why does it matter how that data is transmitted to you.

220.87 is not just for the existing service or switchgear, it is used on upgrades as well.

I think we all understand what's at stake.

So to size your service entrance conductors on that data instead of the max of either is an issue.

Sorry, max of either what or what? Again, I feel you misunderstood something about the data that these systems can make available.