Confused about utility transformer sizing for PV system

[pendragon]

I run a business from home that uses a fair amount of power for scientific computing servers. We have applied to our utility (Xcel) to install a 26 KW solar PV system. We would be on net metering, and the array is sized to roughly match our current usage with an allowance for PV degradation. The solar array output would be fused at 125A just before a line-side tap at our service entrance.

The house was built in 1970 with 200A single phase service that runs underground (4x0 AL 3 wire) to the utility pole where we are on an ancient 50 KVA transformer. The utility is claiming the maximum sized PV system they will allow us to install with the current infrastructure is 3.7 KW. To get anything larger they want us replace their old transformer with a new 100 KVA and bury a new 350AL line to the pole. Our service would remain at 200A.

Our OCPDs will limit the maximum current we can draw from the utility to 200A, and the maximum current we could feed to it to 125A. I don't understand why they are making this demand, other than perhaps trying to get us to foot the bill for a worn out transformer. I've looked through NEC 2017 and the utility's electrical standards, and searched online for something similar, but I've drawn a blank. Is someone able to suggest a less nefarious explanation?

 

[Dennis Alwon]

You clearly need to get someone who knows about solar involved in your setup. With a 20 amp service you cannot back fed the panel with 125 amps--The max back fed breaker allowed is 40 amps on a 200 amp panel with a 200 amp ocdp .

Unfortunately this is not a diy site and I must close this thread

 

[Dennis Alwon]

I am opening this thread after talking with the poster. He has a solar company working on this.

 

[pendragon]

Dennis kindly suggested I provide a few more details about the proposed system, as they might make the design clearer. The PV system would be based on microinverters, and thus its output is AC. The DC sizing of the array is 26.4 KW. Originally the plan was to back-feed the microinverter outputs through breakers in a new main service panel with a 200A OCPD and a 300A busbar. NEC 2017 calculations and the utility's standards document both would allow this. Instead the solar company chose to install a new main panel with a 200A OCPD and 225A busbar, and perform a line-side tap. The microinverter outputs would then be combined and fused at 125A before joining the interconnect between the utility meter and the main panel. While not universal everywhere, this utility accepts either back-feeds or line-side taps for grid-tied PV systems.

My understanding is neither the utility nor the AHJ has an issue with the design. The only issue is the utility wants to replace their existing 50 KVA transformer with a 100 KVA, and upgrade our service loop. Because of the particulars, this will be a considerable expense, and I cannot understand why the 50 KVA transformer is insufficient for a maximum draw from the utility of 200A, as protected by the main panel OCPD, or a maximum feed to the utility of 125A, as protected by the PV fuses into the line-side tap. I of course plan to discuss this with the utility and solar company, but I was hoping for some insight as to what might be going on here. Thanks!

 

[electrofelon]

Their transformer sizing doesn't make any sense. I wonder if this is one of those "solar unfriendly" POCO's?

Edit: perhaps you could ask them why they want such a large transformer?

 

[jaggedben]

Yeah all those details are pretty much irrelevant...This is not about the NEC. The utility's grid is what's at stake and it is not covered by the NEC.

It comes down to the fact that 26kW is a very large system for a residence, and the utility knows the state regulations better than you and your contractor. In California the utility would be on the hook for the transformer upgrade if the system were less than 30kW and/or the transformer was shared by more than one customer. But these regulations vary from state to state and you will need to find an expert on them, or your contractor will need to become one. My first call would probably be to the state public utilities commission to try to find out what the relevant rules are. My second action would probably be to find if there's any consumer group (ratepayer advocacy) who is willing and able to help.

Find out if your transformer is shared with any other customers because that may be key to who is responsible for what.

Finally, they may have given you some leverage by stating the 3.7kW limit, which is ridiculously low. They are supposed to know that you have a 200A service, and they know you've been using quite a lot more than 3.7kW on it. They are not exactly to give you a transformer rated for 200A, but it looks bad for them if they've only given you one for 20A! I would guess there's one of two things going on: a) they should have already upgraded your transformer for the load, regardless of solar, or b) your service shares the transformer with another customer who also has solar, and the transformer can't handle the backfeed of both systems. Either situation strikes me as an argument in favor of some cost-sharing.

My guess is that to avoid paying the full amount for the transformer upgrade, you'll need to gain the sympathy and attention of someone who can add some clout to your case. Either someone at the public utilities commission, or an advocacy group, or possibly an elected official. Hiring a lawyer may or may not pay for itself.

 

[jaggedben]

Actually I'll amend my last post somewhat to echo electrofelon's. If they told you that they existing transformer is 50kVA, it makes zero sense to tell you that you can only backfeed 3.7kW, unless there is already a large amount of solar connected to that transformer.

 

[zbang]

Confusing use of terms.... (let's see if I have this right)

The PoCo doesn't want a system larger than 3.7kw which implies they don't want a backfeed larger than that. This doesn't seem to account for power produced that doesn't leave the customer's property.

The OP/customer wants a larger solar system because they're buying a lot of power from the PoCo ("the array is sized to roughly match our current usage with an allowance for PV degradation"); they'll still buy that power at night unless batteries are involved.

The backfeed concern would only kick in if the customer's actual use goes down markedly during the day.

Does that cover it?

(Are there ways of configuring a system to limit the backfeed?)


If it were me, I'd consider using part of the solar output to feed the DC side of a UPS, which then feeds the computer. OTOH that's more engineering than most people want to take on.

 

[pendragon]

Thanks for your replies. We're near the end of a service branch and all of the properties here are on the large side. The transformer in question is about 100' away from my lot corner and through trees it's hard to see, but it's reasonably possible I'm sharing it with one other property. So that may be the utility's motivation to stick an upgrade on me.

That doesn't fully explain why they want me to pay for a new 100 KVA transformer when a dedicated 50 KVA would do, and especially bury a whole new service loop when the existing one is sufficient for 200A. No one else in this area has a solar PV installation, so that alone can't be the reason.

 

[pendragon]

zbang - I think you've got it right. This system would be on net metering, so as long as we have a long term surplus feeding the PoCo, we shouldn't be buying power from them. But that means during the summer we need to fill the bank.

I've been collecting weather and solar data for the past 13 years, and with the proposed solar array, the maximum peak output we could produce is only around 20 KW (our roof orientation is decent, but not optimal). Our background usage is about 2 KW day and night, with an annualized average of another 0.5 KW caused by peak demands from appliances and non-24/7 servers. However we also are in the process of installing a fully variable speed geothermal heating and cooling system to replace a gas boiler (no current AC), which will cause a very weather-dependent load. That has been sized into the PV design. Our highest production days would be around 160 KWh/day and the lowest around 20 KWh/day. Our annualized average usage is projected at 88 KWh/day, which would be higher when our PV production would be the lowest.

At peak PV production we would struggle to export more than 18 KW, and with the geothermal system running, it would be even less. The suggestion to limit this might be a strategy to consider, because for the most part it would never happen.

We are looking into expanding our use of batteries/UPS/generator, because living at the end of a lightly populated service branch in a metro area means we are always the last on the PoCo's list for fixing power outages. Just a few days after moving in 24 years ago, we had a 7 day power failure in freezing weather, so dealing with that has always been part of our power solution.

 

[pendragon]

With this discussion I'm beginning to wonder if the PoCo has a drastically undersized transformer and is realizing their ruse is playing its last hand. My solar company had to provide them a study justifying the projected increased usage from our new geothermal system to get our PV sizing up to the 26 KW level. Maybe not a coincidence, and maybe that's not a 50 KVA on the pole.

 

[jaggedben]

What they are saying makes little sense based on the information given. You might seriously ask them if their engineer got the decimal point wrong. 37kW instead of 3.7kW.

If you already have 200A rated conductors coming to the house from the transformer there is absolutely no reason those should have to be upgraded.

The term 'service loop' is imprecise. If overhead it's called a service drop and if underground it's called a service lateral.

 

[jaggedben]

There are potentially voltage drop issues that could go beyond the rating of the transformer, too. But overall I think you are right to be suspicious that they are trying to get you to pay for something that should have already done.

I still think you should try to enlist some outside help.

 

[pendragon]

jaggedben - this made little sense to me, so I figured it would be helpful to get a high-level assessment here. Every number/detail of importance is a direct quote from either the PoCo or my solar company's submitted design package.

My solar company expects more information next week, and I guess we'll go from there. I had hoped for more expertise from them, but maybe it's simply a matter of finding the right person. As has been said, this is a large system for a residence and definitely out of the comfort zone for some of the other companies I had asked to bid. I will start probing the avenues you suggested in your first post. Not that it matters, but "service loop" was the term the PoCo engineer wrote. That may reflect on their experience/knowledge. Thanks for your responses.

 

[ggunn]

You clearly need to get someone who knows about solar involved in your setup. With a 200 amp service you cannot back fed the panel with 125 amps--The max back fed breaker allowed is 40 amps on a 200 amp panel with a 200 amp OCPD

He said it was line side interconnected. In that case, per the NEC he is limited only by "the size of the service", but how that is defined is not exactly clear.

 

[synchro]

The harmonic distortion of solar inverters can cause a very small but apparently not negligible amount of heating in a distribution transformer when the photovoltaic backfeed is an appreciable part of its rating:

https://www.researchgate.net/profil...Solar-Panels-on-Distribution-Transformers.pdf

A small increase in temperature to an already marginal transformer could effect reliability because the expected lifetime has an exponential relationship with temperature.

Whether this effect is really that significant or not, it's possible it could be one of the factors in play from the the POCO's point of view. I thought it might be good to be aware of this issue just in case the POCO brings it up, and so you can have a response ready to give them if they do.

 

[Carultch]

He said it was line side interconnected. In that case, per the NEC he is limited only by "the size of the service", but how that is defined is not exactly clear.

Would it also be limited by the size of the service conductors? I see this as a blindspot in the NEC. Consider a 200A service built with #2/0Cu conductors, and suppose you wanted to fully utilize this with 200A of generation. You'd be doing a supply side tap (in #3/0) that needs more ampacity than the #2/0 service conductors themselves. It logically doesn't make sense to be allowed to do this, because the #2/0 service conductors are a chokepoint that don't have the capacity for the full 200A.

The 83% service conductor ampacity rule existing in the first place is another matter entirely. I get that there is load diversity and you rarely draw the full 200A on a 200A service, but in concept that is already accounted for in the load calculations that size the 200A service in the first pace. 175A service conductor ampacity for a 200A service creates a theoretical point of failure in the event that you ever do draw 176A or more, for a serious amount of time. The conductors exceed their ampacity, and the OCPD will never trip.

 

[ggunn]

Would it also be limited by the size of the service conductors? I see this as a blindspot in the NEC. Consider a 200A service built with #2/0Cu conductors, and suppose you wanted to fully utilize this with 200A of generation. You'd be doing a supply side tap (in #3/0) that needs more ampacity than the #2/0 service conductors themselves. It logically doesn't make sense to be allowed to do this, because the #2/0 service conductors are a chokepoint that don't have the capacity for the full 200A.

The 83% service conductor ampacity rule existing in the first place is another matter entirely. I get that there is load diversity and you rarely draw the full 200A on a 200A service, but in concept that is already accounted for in the load calculations that size the 200A service in the first pace. 175A service conductor ampacity for a 200A service creates a theoretical point of failure in the event that you ever do draw 176A or more, for a serious amount of time. The conductors exceed their ampacity, and the OCPD will never trip.

That's the kind of thing I was alluding to when I said that the "size of the service" isn't clearly defined.

 

[gar]

190721-2237 EDT

pendragon:

You are working with a rule based system which may not relate to logic. You need to know what are the legal rules, probably public service commission defined, and how those rules relate to you.

If there is voltage drop from your supply transformer to your home, then the power company is supplying you power no matter how much power your solar system produces.

If the voltage drop is from your home to the power transformer, then you are supplying the grid with power.

If your solar system can never supply more current to the power transformer than it is rated for, then there should be no problem. However, this statement is not correct unless you know more about how the transformer is really rated. Additionally is the distortion that might exist in the current waveform mentioned by synchro.

If you powered your equipment from a battery bank and fed this battery from both your PV and the grid, then you would feed nothing to the grid. Your PV will only provide substantial power for approximately 6 to 8 hours per day. If on average you produce more solar energy than energy you use, then you get no payback from the power company. They don't really like net metering because they lose money on net metering. But if you have a major power outage, then with such a battery system you may be able to at least partially run your equipment during such outage.

Note that only about 1/3 of your electric bill goes to buying energy. Thus, net metering costs the power company to buy power from you. About 3 times what it costs for energy from their primary source. There are other factors in the cost equation, but this is a simple view.

.

 

[pendragon]

I can handle a rule-based system, particularly when I can see the rules. I read through the PoCo's electrical standards document again, and went through the state's statutes and public utilities commission's regulations over the weekend. There's a lot there, but not much that applies to my situation. I'm assuming the PoCo also has internal policies, engineering documents, and cultures that are not necessarily exposed for public consumption. To understand how these apply I will likely need to engage an outside expert familiar with them.

I can well understand why the PoCo could want to push back here. Not only would this be a large residential installation, but if I elect the full net metering option (banking credits forever), this would be a binding contract that would apply as long as I live here. Furthermore as this would be a > 25 KW PV system, the PoCo would have to pay me for RECs at a considerably higher rate than a system < 25 KW (the rate this year is $0.0375/KWh). Assuming the PV system generates at least as much energy as I use, that will cover many times over what I will have to pay in connection and administrative fees.

So I will be a money bleeding proposition for them from the get go, and now even more so with our new geothermal system's electrical requirements, replacing our boiler's gas usage (PoCo also provides gas).