Rooftop Solar/SRECs in Washington DC - Advice?

[Lawpup]

Hi all,

I’m new to this forum (and to solar), and I was hoping I could get your collective feedback on two quotes that I received to install solar panels on my roof, as well as some questions that I have about solar in general. I’ve done a lot of reading about solar and specifically solar in D.C., including the “Solar for Dummies” book that is recommended on this forum. D.C. has one of the best SREC (Solar Renewable Energy Credit) markets in the country, with SRECTrade currently having them at a market price of $435 per SREC (with 1kW = 1 SREC), so I’m eager to get this system up and running. That’s definitely an added incentive to set up solar, as I can generate SREC income separate and apart from the money that I’m going to be saving on electric bills.

First off, after doing some research, I found out that D.C. allows residents who install systems that are less than 10kW in size to use estimated production - not actual production - as a basis for calculating how many SRECs you generate in a year. If your system is larger than 10kW, then you have to get a revenue-grade meter and use actual production as the basis for calculating SRECs generated. My understanding is that estimated production is calculated by GATS using the PV Watts modeling. In plugging in the quotes below for my house, the estimates are all between 13,000 and 14,000kWh/year, which equates to 13-14 SRECs, or about $5,200 per year in additional income (based on current market price, and considering commissions (13 SRECs x $400 per SREC = $5,200)).

When I got my first set of quotes from companies, the largest system that I was quoted for was about 10.6kW. Given that any system size over 10kW would require me to use actual production instead of estimated production, and seeing as 10.6kW isn’t really that much greater than 10kW, I decided to request from the installers only proposals that were as close to 10kW as possible while still being under. My thinking here is that calculating SRECs based on estimated production will lead to a more certain/stable stream of income than SRECs based on actual production, as actual production can be affected by solar panel malfunction, uncharacteristically bad weather, or a number of other factors. Plus, I can always choose to switch to using actual production as the basis for SREC calculations down the line if I ultimately find that my system generates a lot more than what is estimated; but being under 10kW at least gives me the option to use estimated production. Does this seem logical to you? Am I missing anything here? I understand that it probably wouldn’t make sense to get a 9.9kW system if my roof could handle something much larger, such as 12 or 13kW, but I’m not sure if the difference between a 10.6kW system and a 9.9kW system is substantial enough to forego the opportunity to use estimated production for SRECs.

Next, I was hoping to get some input on the below quotes and let me know what you think. On its face, Quote #2 appears to be better due to a lower price per watt, slightly larger system, and longer workmanship warranty, using what appear to be comparable or the same materials. I was wondering if there was something that I’m not factoring in here that I should be, such as (1) whether it is better to have fewer panels on a roof (22 in quote 1 vs. 27 in quote 2), (2) whether there is any substantial difference in the optimizers (P505 in quote #1 and P370 in quote #2), and (3) whether, in practice, a 10-year vs. a 25-year workmanship warranty really matters (i.e., can you even realistically get a solar installer to acknowledge the workmanship warranty after year 10, anyway?).

Thanks in advance for all of your input! I truly appreciate it.

Quote #1
- Price (pre-federal tax credit): $24,000
- Size: 9.79kW
- Price per watt: $2.44/watt
- Panels: 22 x REC Alpha Series 445 watt panels (25-year warranty)
- Inverter: SolarEdge 7600H-US HD Inverter (25-year warranty)
- Optimizers: 22 x P505 power optimizers
- Workmanship Warranty: 10 years

Quote #2
- Price (pre-federal tax credit): $23,200
- Size: 9.99kW
- Price per watt: $2.32/watt
- Panels: 27 x REC Alpha Series 370 watt panels (25-year warranty)
- Inverter: SolarEdge 7600H-US HD Inverter (25-year warranty)
- Optimizers: 27 x P370 power optimizers (25-year warranty)
- Workmanship Warranty: 25 years

[bob-n]

The 370 (60 cell panel) is a physically smaller panel than the 445 (72 cell panel). Because of fewer cells, they also have lower output voltage. Otherwise, they are really almost identical. You can learn the specs on these panels with a simple google search.

Because of the different panel output voltage, you need different optimizers. The P370 and P505 are right for those specific configurations. You can learn the specs on these optimizers with a simple google search.

The two installations will look slightly different due to the different panel sizes. It might be slightly wider or slightly longer or slightly offset with one vs the other. Otherwise, the quotes seem virtually identical.

Much more important is the quality of the work. Secondary, but perhaps also important is the endurance and longevity of the installation company, in case you have problems or questions.

Note that a "workmanship warranty" does not cover replacing panels or other components that fail. It only covers mistakes made during installation, such as incorrect roofing that causes a leak. After a few months, they may argue that anything that shows up wasn't their fault. So the 25 vs 10 year warranty difference may be meaningless.

I can't comment on aiming for <10kW vs >10kW. It could be that the SolarEdge monitoring software would meet the requirements of the >10kW readout, or it could be that you need to add an Inepro or WattNode meter for ~$200 extra.

Most people size the solar array to their needs, plus perhaps 15% to 30% for future. Yes, RECs and NetMetering pays you for the excess power you generate, so more power gives you more back. But you need a long time to recoup the capital invested in the hardware and labor, and going too big pushes that time-to-break-even out even further in time. But that's your call. Buy as much power as you like.

I hope this helps.

[J.P.M.]

Q1: How much electricity do you burn through in a year ?

Q2: Who are the vendors ?

Q3: Purchase, lease or PPA ?

Q4: Have you had your roof inspected/serviced in the area where the array will be ?

[Lawpup]

The 370 (60 cell panel) is a physically smaller panel than the 445 (72 cell panel). Because of fewer cells, they also have lower output voltage. Otherwise, they are really almost identical. You can learn the specs on these panels with a simple google search.

Because of the different panel output voltage, you need different optimizers. The P370 and P505 are right for those specific configurations. You can learn the specs on these optimizers with a simple google search.

The two installations will look slightly different due to the different panel sizes. It might be slightly wider or slightly longer or slightly offset with one vs the other. Otherwise, the quotes seem virtually identical.

Much more important is the quality of the work. Secondary, but perhaps also important is the endurance and longevity of the installation company, in case you have problems or questions.

Note that a "workmanship warranty" does not cover replacing panels or other components that fail. It only covers mistakes made during installation, such as incorrect roofing that causes a leak. After a few months, they may argue that anything that shows up wasn't their fault. So the 25 vs 10 year warranty difference may be meaningless.

I can't comment on aiming for <10kW vs >10kW. It could be that the SolarEdge monitoring software would meet the requirements of the >10kW readout, or it could be that you need to add an Inepro or WattNode meter for ~$200 extra.

Most people size the solar array to their needs, plus perhaps 15% to 30% for future. Yes, RECs and NetMetering pays you for the excess power you generate, so more power gives you more back. But you need a long time to recoup the capital invested in the hardware and labor, and going too big pushes that time-to-break-even out even further in time. But that's your call. Buy as much power as you like.

I hope this helps.

Thanks, Bob. Yes, this is helpful. I read about the specs for the panels, and, as you point out, the 445 panel is a bit longer (about 14 inches longer), and with a slightly lower efficiency (21.0% for the 445 and 21.2% for the 370s), but my understanding is that they are both high quality panels. Regarding the optimizers, that's what I figured (different type based on the different size of the panels).

Your emphasis on the quality of the work and the longevity of the installer definitely pushes me more towards the Quote #1, as that installer pointed out installation nuances early on, has been very responsive to all of my questions, and even scheduled a site visit early on. I also tend to agree with your assessment of the workmanship warranty, and how once you're past 2 years, it probably doesn't really matter.

The <10kW vs. >10kW is really the big question for me now. As I mentioned, in D.C., if you have a system that is under 10kW, SRECs are calculated based on estimated production as calculated by PV Watts. If your system is greater than 10kW, your SRECs are calculated based on actual production as reported by a revenue-grade meter. The two options I have for system size are 9.79kW, where I would have the option of using estimated production for SRECs, and 10.235kW, where I would be required to use actual production for calculating SRECs. Does anyone have a sense of whether, in this situation, using estimated production on a slightly smaller system would be a better choice than being required to use actual production on a slightly larger system? Some factors that may go into this are whether PV watts estimates are typically larger or smaller than actual production (or are roughly accurate). I'm probably over-thinking this, but figured I'd ask for opinions. Note that the SREC calculations have no bearing on my own electricity offsets; they are distinct calculations, and electricity offsets can only be based on how much my system actually generates.

[Lawpup]

Q1: How much electricity do you burn through in a year ?

Q2: Who are the vendors ?

Q3: Purchase, lease or PPA ?

Q4: Have you had your roof inspected/serviced in the area where the array will be ?

A1: Don't know, we just moved in about 5 months ago. From September through December, our kWh/month energy consumption has ranged from 720kWh-ish in October to just over 1,000kWh in December.

A2: Both based in Maryland. Solar Energy Services and Sustainable Energy Systems.

A3: Purchase

A4: Roof is about 5 years old, and had some roofers inspect it and do some preventative maintenance on it back in October. Solar Energy Services also took a look at the studs/beams in the attic and thought it looked fine.

[bob-n]

PVWatts is a model. It is as good as the data. The weather data is extremely good. Then it comes down to your inputs for everything else: orientation, angle, panel, efficiency, shading, etc. For my home, PVWatts seems optimistic because it doesn't include shading. Here's a plot of my actual 2020 production and what PVWatts predicted. So if you're in a similar situation, estimated production would be in your favor.

pvwatts.jpg

[Lawpup]

PVWatts is a model. It is as good as the data. The weather data is extremely good. Then it comes down to your inputs for everything else: orientation, angle, panel, efficiency, shading, etc. For my home, PVWatts seems optimistic because it doesn't include shading. Here's a plot of my actual 2020 production and what PVWatts predicted. So if you're in a similar situation, estimated production would be in your favor.

pvwatts.jpg

Thanks, Bob. Based on the chart, it looks like PV watts estimated you would generate almost 1,000kWh more than you actually did. Does that sound right to you?

[J.P.M.]

Thanks, Bob. Based on the chart, it looks like PV watts estimated you would generate almost 1,000kWh more than you actually did. Does that sound right to you?

PVWatts is a model for preliminary design, not necessarily a predictor of performance.

Reason: Even with inputs that are close to or smack on the actual design, mostly but not entirely because weather is variable, actual annual system output will be different than the PVWatts modeled output over any consecutive 365 day period by up to maybe +/- 10 % with any 30 day PVWatts modeled output varying from actual by maybe up to +/- 30%.

Example: After the first year's operation, 10/13/2013 through yesterday, my system actual running 365 day output:
Average: 9,071 kWh/yr.
Min: 8,716 kWh/yr.
Max: 9,571 kWh/yr.
Std. dev.: 183 kWh/yr.

System size: 5.232 kW.

The distribution is a bit skewed toward the high side, probably due to higher output values achieved during the first year's burn in.

My system probably had a tighter +/- than the PVWatts stated annual +/- 10 % tolerance at least partly because I live in a pretty sunny climate which tends to give the distribution a smaller std. deviation. Hypothetically, my system produced about 84% as much electricity over the prior 7 years as it would have if every day had been cloudless.

PVWatts modeled output is 9,102 kWh/yr. using a shade correction factor I plugged in manually from PVWatts' big brother, another model from NREL called "SAM", which model BTW produced an annual modeled output of 9,062 kWh/yr for year 4 of a 7 year life.

[Lawpup]

Thanks, all. J.P.M.: that's helpful. So it looks like for you, the models were pretty spot-on with your actual production.

I spoke with some more folks at Energy Sage and at some SREC brokers, I'm getting the sense that there isn't really any big advantage to using production estimated by PV Watts as a basis for calculating SRECs, as opposed to just using actual production. As such, there isn't really an incentive for me to stay under 10kW in system size if I can go bigger. I told the solar installers I'm working with to give me a quote for the biggest system that they can put on my roof, which I anticipate will be in the 10.25-10.75kW range. Looking forward to signing a contract and getting the ball rolling on this!

[J.P.M.]

Thanks, all. J.P.M.: that's helpful. So it looks like for you, the models were pretty spot-on with your actual production.

I spoke with some more folks at Energy Sage and at some SREC brokers, I'm getting the sense that there isn't really any big advantage to using production estimated by PV Watts as a basis for calculating SRECs, as opposed to just using actual production. As such, there isn't really an incentive for me to stay under 10kW in system size if I can go bigger. I told the solar installers I'm working with to give me a quote for the biggest system that they can put on my roof, which I anticipate will be in the 10.25-10.75kW range. Looking forward to signing a contract and getting the ball rolling on this!

You're welcome.

Using PVWatts modeled output, or the output from any model for that matter, as a basis for SREC credits makes no sense unless it is used to estimate ongoing payments with a trueup compared to actual system production at some interval such as 1X/yr.

Read the PVWatts help screens for a more detailed explanation of how to use and interpret the model's output. Hell, even the irradiance data including intensity and temporal distribution is better than 90% synthetic.

As for optimum system size (rather than the biggest system that can be put in a location, roof or otherwise), that's usually best done by an economic analysis keeping the overall goals in mind.

SRECs aside for a moment, one goal is often to have the lowest long term cost of meeting an electric bill using a mix of PV and conventional grid power, not necessarily cramming as much PV on a roof as is allowed. SRECs change the calculus some, but the usual biggest and probably sanest goal for most homeowners is the still the lowest long term cost of supplying electricity to a home, not starting a mini utility.

[khanh dam]

my 10.2kw SYSTEM IS currently generating about 10% more than pvwatts estimated.
Ohio used to have screcs 2 years ago, now they don't.
Under Biden Presidency SREC credits might increase? I really do not understand how they work. But I do know they can be here one year and gone the next or the price go down a lot, so dont' count on them for future revenue. You are very lucky to have them IMHO.

I'd go with the cheaper price per watt. $800 is a good savings.
Just make sure they use FLASHING for the solar racking.
some companies have switched to stupid rubber pucks that literally just sit on top of the roof, with freezing and heat expansion. I'm guessing they will start leaking 10 years latter.

Solar Edge is a crap inverter. They fail quite often. If you are on Facebook group you can join solar installers group and read complaints. I am amazed Solar Edge is the #1 inverter installed, you will likely need to have it replaced after 12 years warranty is up. I doubt your installers will be around then so figgure another $3000 latter on to keep it running.

I'd get a ground mount or pergola setup if you have the space, Most roof installations of that size are a faceted mess, unless you have an older house with huge flat areas of roofing.
most modern houses buildt in the last 15 years have horrible roofs.

[khanh dam]

when they install it, get on the roof and watch them install it, in addition to roof flashing mounts make sure they put some NoOx on each cable connection to prevent future corrosion

[J.P.M.]

PVWatts is a model. It is as good as the data. The weather data is extremely good.

Take this FWIW.

Further to Bob-n's comment about the weather data PVWatts uses being extremely good: Maybe a bit of clarification might be in order.

The weather data PVWatts (and a lot of other models) uses can produce reasonably useful results from weather and irradiance models it uses if the rest of the inputs to the PVWatts model reflect reality to a reasonable degree. That PVWatts (and other) model) can do that speaks to the weather data's usefulness. But to call the weather data "good" or "extremely good" may not be the best way to describe it.

The weather data has two origins. One is the National Solar Radiation Data Base (NSRDB) which morphed from and shares a lot of data with the old(er) Typical Meteorological Year (TMY) data, both TMY1 and the newer TMY 2 data sets.

Point is, that while both sources seem to be representative for an area and give good results that reflect reality reasonably well - particularly given the chaotic nature of weather - the dirty little secret is that most all of the data is modeled, that is, it's mostly synthetic - not measured. The current data is all modeled using time averaged measurements from geostationary satellites. The older data versions, TMY 1 and 2, over 1,000 stations, were all based on ground visual cloud cover observations and other weather data gathered from 26 airports around the U.S. The rest of the >1,000 station data is a >90% modeled approximation of those 26 stations' data with a lot of that data having a lot of gaps.

That the data bases seem capable of being as reliable as they have been as well as the quite close agreement between the old and new methods seems to speak to their bottom line usefulness. I might call the databases useful and reasonably reliable because I've been working with them for ~ 40+ yrs., but to call them extremely good may mask the stochastic nature of the databases that makes solar irradiance modeling and so solar device output no more accurate than long range weather forecasting vs. saying it'll probably be hotter in the summer than the winter.

See the TMY manual for particulars.