micro-inverters or optimizers when shading is major issue ?

[jflorey2]

1. Will micro-inverters or optimizers give me the best efficiency (least performance reduction with partial shading), or are they pretty much equivalent ?
2. Is SolarEdge's system (optimizers and inverters) superior to other optimizers, as I've been led to believe ?

They are pretty much equivalent from an efficiency perspective. I would strongly suggest getting half-cell panels if shading will be a big issue. SolarEdge is about equivalent to Enphase in terms of efficiency. If you want to go with Enphase you might want to wait for the IQ8 inverters, since they add intentional islanding.

[RShackleford]

Can you post a photo of what kind of shading issue you have?

What's a good way to try to photo this complex 3D scene ? Maybe just stand at where I'm thinking of doing the installation (probably a carport, likely just a shed and not actually a parking spot) and point the camera due south, something like that ?

There's a lot of misinformation that shading one section of a string effects the entire string. This is 100% NOT TRUE. Bypass diodes effectively 'delete' the shaded portion and the unshaded panels continue producing unaffected. Each string does need to be on an independent MPPT but most inverters now have 2-4 MPP channels.

I guess that'd be a reason to not go with micro-inverters. Wire up one or more optimizer-less panels to a non-SE string inverter, insert Tigo optimizers if and when needed. I imagine that inserting an optimizer into the string would be simple (especially for my shed/carport with no real "roof" underneath) and Tigo claims you only need to do it for the "problem" panels.

[nwdiver]

I guess that'd be a reason to not go with micro-inverters. Wire up one or more optimizer-less panels to a non-SE string inverter, insert Tigo optimizers if and when needed. I imagine that inserting an optimizer into the string would be simple (especially for my shed/carport with no real "roof" underneath) and Tigo claims you only need to do it for the "problem" panels.

Not even sure that would be worth the investment. If this is on a shed or carport you may not need rapid shutdown. Tigo can be retrofitted. If I was in your shoes I would just throw up a SMA string inverter. If you're disappointed with the loss of production you can always add optimizers later.

[Ampster]

What's a good way to try to photo this complex 3D scene ? Maybe just stand at where I'm thinking of doing the installation (probably a carport, likely just a shed and not actually a parking spot) and point the camera due south, something like that ?

That is a start. There is very complicated digital camera that can plot the sun at all times of the year and give you a fairly good analysis. I saw a solar installer use it in a forested location in Mendocino. He did not share the details.

Google Solar roof is free and can give a gross analysis. I used it as an initial screening process when I was considering a home purchase two years ago.

Also there is a device called a solar pathfinder that costs about $300 if you don't want to wait all year for actual results or want more specifics than the Google project or if your site has not been included in the Google project.

[J.P.M.]

What's a good way to try to photo this complex 3D scene ? Maybe just stand at where I'm thinking of doing the installation (probably a carport, likely just a shed and not actually a parking spot) and point the camera due south, something like that ?

I've used the solar pathfinder and it seems to give decent results. Snoop around on the net and you can find places that'll rent you one.

I've also done something similar to that described in chap. 3 of the tutorial at :www.thesolarplanner.com/array_placement3.html. Truth to tell, that site uses some stuff I worked on/with some folks at ISES (International Solar Energy Society), and mods. to the sunpath chart, back in the day when I was with their resource assessment div.

Not a plug for the solarplanner site, but that site has a lot of other decent info as well, although I don't agree with all of it, and it seems a bit less altruistic that they'd like folks to think, but you can learn a lot from it - just avoid the hype and don't take all of it for gospel. The shade analysis method described there will get you what's similar to the solar pathfinder but will take some time.

Also, but not for the feint of heart, SAM, the NREL model has a decent shade analysis module, but that'll probably take a bit of time learning to use the whole model which is like PVWatts on steroids. Unless you know what you're doing before you walk in, you can shoot yourself in the foot.

Keep in mind that any method to estimate shading is just that - an estimate - and in most applications that's good enough.

[RShackleford]

I've used the solar pathfinder and it seems to give decent results. Snoop around on the net and you can find places that'll rent you one.

"Google solar roof" not available in my location.

I should've mentioned, I did get a bid from an installer of turnkey systems: $16K (before tax credits) for 4.34kw, 14 Silfab Mono 310 panels, SolarEdge P320 optimizers, inverter, monitor.

They did use a fancy electronic widget to measure sun availability (probably the device you'all mention); the bid doesn't include any percentage availability, but does state that the system should generate 4452 kwh in the first year. I wanna say the percentage was somewhere in the 70s, but I guess there's a way to compute it from the kw versus kwh figures. I dunno how biased this might be: would they estimate high to make me think my payback will be shorter, but at the risk of my complaining of underperformance, or vice versa ?

I should also mention, they did this on the sunny portion of my roof, before I decided to do a free-standing thing like a carport, and that would be in a somewhat less-shaded location. FWIW, here is a photo from the new (freestanding) location. (This is October 14, latitude is 36 north). I will definitely take down the largest tree you see to the left and to the right, and could take down more if I choose (all my property, and I can drop all but the big left-side one on my own).

Actually, when I try to upload the image (an 83KB JPG) and click "Post Reply", I get: There has been a database error, and the current page cannot be displayed. Site staff have been notified.

[J.P.M.]

"Google solar roof" not available in my location.

I should've mentioned, I did get a bid from an installer of turnkey systems: $16K (before tax credits) for 4.34kw, 14 Silfab Mono 310 panels, SolarEdge P320 optimizers, inverter, monitor.

They did use a fancy electronic widget to measure sun availability (probably the device you'all mention); the bid doesn't include any percentage availability, but does state that the system should generate 4452 kwh in the first year. I wanna say the percentage was somewhere in the 70s, but I guess there's a way to compute it from the kw versus kwh figures. I dunno how biased this might be: would they estimate high to make me think my payback will be shorter, but at the risk of my complaining of underperformance, or vice versa ?

I should also mention, they did this on the sunny portion of my roof, before I decided to do a free-standing thing like a carport, and that would be in a somewhat less-shaded location. FWIW, here is a photo from the new (freestanding) location. (This is October 14, latitude is 36 north). I will definitely take down the largest tree you see to the left and to the right, and could take down more if I choose (all my property, and I can drop all but the big left-side one on my own).

Actually, when I try to upload the image (an 83KB JPG) and click "Post Reply", I get: There has been a database error, and the current page cannot be displayed. Site staff have been notified.

Sounds like you haven't done any PVWatts runs for your location yet. You'll need a decent est. of long term annual system output if/before a shade analysis is meaningful. PVWatts is as good a way as any to get one, at least for preliminary design.

[RShackleford]

Sounds like you haven't done any PVWatts runs for your location yet. You'll need a decent est. of long term annual system output if/before a shade analysis is meaningful. PVWatts is as good a way as any to get one, at least for preliminary design.

It just wants address for getting latitude, and weather, right, doesn't do site analysis beyond that ?

PVWatts says my 4.34kw system would get 7344 kwh/year; assumes 36 degree tilt (my lattitude), "standard" module type and "open rack", and 0% system losses. So it sounds like that installer's quote of 4452 kwh/year is assuming only about 61% overall system efficiency, so probably not p*ssing in my ear.

[foo1bar]

PVWatts says my 4.34kw system would get 7344 kwh/year; assumes 36 degree tilt (my lattitude), "standard" module type and "open rack", and 0% system losses. So it sounds like that installer's quote of 4452 kwh/year is assuming only about 61% overall system efficiency, so probably not p*ssing in my ear.

I thought you said their estimate of 4452kwh was from your roof?
i would see what it says when putting in what they likely used for your roof.
(probably not 36 degree tilt, nor perfectly due south. Probably 10% system losses or maybe the default of 14%. "Fixed (Roof Mount)", not open rack.)

FWIW, if they're trying to sell the system based on "this is how much power you'll generate", that estimate may be high to help show their system to be so much better than the POCO.
OTOH, if they're selling it with a "we guarantee you'll generate at least 4452kwh, and if it doesn't we'll pay you $0.20/kwh", then their guaranteed production is going to be low.
(at least that's my somewhat cynical view/experience)

FWIW, I used an app on my phone. I think it was called "Solar Shading" to get solar shade estimate. It only sort-of-worked.... I think a lot of the blame rested with the phone I had for how difficult it was to get an accurate or even reasonable treeline out of it.

[J.P.M.]

It just wants address for getting latitude, and weather, right, doesn't do site analysis beyond that ?

PVWatts says my 4.34kw system would get 7344 kwh/year; assumes 36 degree tilt (my lattitude), "standard" module type and "open rack", and 0% system losses. So it sounds like that installer's quote of 4452 kwh/year is assuming only about 61% overall system efficiency, so probably not p*ssing in my ear.

Short answer to your question is no.

PVWatts needs more than that. Read the help screens for particulars.

You are confused about its requirements and output.

Read all the help screens for particulars.

The model's input asks for address (or zip if in the U.S.) to get latitude, longitude and elevation above M.S.L for nearby locations that have TMY or other solar and weather database locations to get irradiance and weather data for the model. You may get several nearby locations to choose from, depending on what you may know about locations (Example, my closest location is 15 or so miles from me and near the ocean with weather much different than mine, but it's what the model will default to. I instead use Miramar MCAS with weather closer what my site sees.
You will need to input the array's orientation.

The model's output is probably +/- 10% of what an actual and correctly modeled system's annual (or 365 running day) output will be. For 1 month or any 30 day periods that's probably something like +/- 30 % or maybe a bit better. Most of the variation is due to weather not being the same from one period to the next.

The system loss parameter default is 14 %. Many users find that a 10 % system loss factor gets closer to actual system output. Putting in a "0.0" or "0" will cause model output to increase and be off.

I'd run (or rerun if you've already run the model) after a review/read of how the input is done. Then compare that to what the vendor quoted after you find out what was done. Get your own model output using good input and you'll be on firmer ground.

Sounds to me like the installer's 4452 kWh/yr. might be based on shading. If so, and if that's a reasonable approx. of what's likely to happen, IMO, that's a very poor location for a PV array. If it was me, and it sure ain't, I'd either find a better location, or chop down a lot of trees or scrap the idea.

The installer or vendor may not be pissing in your ear but IMO, he's not doing you any favors by not asking or maybe screaming at you "What were you thinking" after seeing that much shading. Reason: A 40 % or so shading loss is usually enough make the project uneconomical by most criteria.

But, not my money/application/life.

BTW, I haven't seen any of your goals for all this PV. So, if I could ask, what do you expect to get from all of this ? Apologies if I missed that.

[RShackleford]

I thought you said their estimate of 4452kwh was from your roof?
i would see what it says when putting in what they likely used for your roof.
(probably not 36 degree tilt, nor perfectly due south. Probably 10% system losses or maybe the default of 14%. "Fixed (Roof Mount)", not open rack.)

The 4452kwh was on the quote I got, and that was on the roof. Quote doesn't specify panel tilt, but the guy stated they usually angle it the same as the number of degrees of latitude (which makes sense to me). Doesn't specify azimuth either but I assume they'd make it due south (the roof points only 15 degrees east of due south, so shouldn't be hard to do). Roof-mount was only 1.5% lower than open-rack, PVWatts says because the panels don't tend to stay as cool. I used 0% losses just as a baseline, so I could see what kind of overall losses they must have been assuming (61% efficiency, it appears, or 39% loss); so that 39% loss must assume their estimate of shading losses.

Anyhow, if I redo PVWatts with roof-mount, 165-degrees azimuth, and 14% losses, I get 6211 kwh/year. So clearly their 4452 kwh/year estimate includes significant shading losses.

Oh, and if I make the azimuth 180 degrees, it only goes up to 6249 (that's amazingly little difference IMHO). Back to 165 azimuth, if I change the tilt by +/- 5 degrees (31 or 41) I get 6208 and 6179 respectively. I amazed the sensitivity to those angles is so small; so really I should let my trees be the driver for my azimuth, and probably for my elevation too.

Anyhow, bottom line, I feel I can assume 5000 kwh/year, since their 4452 is more or less guaranteed, and my shed/carport location will be a bit sunnier. Complicated to figure my annual dollar savings, because I'm on a 3-tier "EV" plan, and the tiers affect not just what I don't use from the POCO, but also what I sell back to them. Per kwh rate (in cents) is about 32 on summer afternoons and winter mornings, but only about 6 the rest of the time (and only about 3 at night, irrelevant of course).

FWIW, if they're trying to sell the system based on "this is how much power you'll generate", that estimate may be high to help show their system to be so much better than the POCO.
OTOH, if they're selling it with a "we guarantee you'll generate at least 4452kwh, and if it doesn't we'll pay you $0.20/kwh", then their guaranteed production is going to be low...

The latter, it appears. The fine print says "guarantees annual solar production to be within 10% of proposed production as calculated by PV Watts for a period of five years or the company will reimburse the retail value of the difference in kWh." But again, they don't say what parameters they used in PVWatts.

FWIW, I used an app on my phone. I think it was called "Solar Shading" to get solar shade estimate. It only sort-of-worked.... I think a lot of the blame rested with the phone I had for how difficult it was to get an accurate or even reasonable treeline out of it.

I'll check that out, thanks.

[RShackleford]

I think addressed most of this in my response to @foo1bar.

The model's input asks for address (or zip if in the U.S.) to get latitude, longitude and elevation above M.S.L for nearby locations that have TMY or other solar and weather database locations to get irradiance and weather data for the model. You may get several nearby locations to choose from ...

It ony gives me 3 choices (I can't tell what they are), and they all come out the same. I'm in a pretty populated area of NC, so I imagine it's pretty accurate.

Putting in a "0.0" or "0" will cause model output to increase and be off.

Yeah, I did that just to see what overall losses the quote was assuming.

I'd run (or rerun if you've already run the model) after a review/read of how the input is done.

See response to foo1bar. I think I've done it properly now.

Sounds to me like the installer's 4452 kWh/yr. might be based on shading.

No doubt.

If so, and if that's a reasonable approx. of what's likely to happen, IMO, that's a very poor location for a PV array. If it was me, and it sure ain't, I'd either find a better location, or chop down a lot of trees or scrap the idea.

The installer or vendor may not be pissing in your ear but IMO, he's not doing you any favors by not asking or maybe screaming at you "What were you thinking" after seeing that much shading. Reason: A 40 % or so shading loss is usually enough make the project uneconomical by most criteria.

To be fair to installer, his initial take was "this isn't a good site"; I asked him to do a quote anyhow. Bear in mind this quote is before the 30% federal tax credit. And that I'll be saving/selling a lot of electricity at a $0.30+/kwh rate (net metering on a tiered plan).

I will DIY, so I'll save a lot - I'd be curious of estimates as to how much I'd save with sweat equity.

Also, I can definitely cut down more trees.

BTW, I haven't seen any of your goals for all this PV. So, if I could ask, what do you expect to get from all of this ? Apologies if I missed that.

It's definitely not all about money. Partly to put my money where my mouth is on energy/environmental stuff. Partly for some self-sufficiency. And yes, I realize I won't get electricity when the grid is down, and that it'll cost a bit more to get that. My plan is to start out with this grid-tied system, and add multi-mode later when hopefully there are better/cheaper options for that.

[foo1bar]

Anyhow, if I redo PVWatts with roof-mount, 165-degrees azimuth, and 14% losses, I get 6211 kwh/year. So clearly their 4452 kwh/year estimate includes significant shading losses.

Oh, and if I make the azimuth 180 degrees, it only goes up to 6249 (that's amazingly little difference IMHO). Back to 165 azimuth, if I change the tilt by +/- 5 degrees (31 or 41) I get 6208 and 6179 respectively. I amazed the sensitivity to those angles is so small; so really I should let my trees be the driver for my azimuth, and probably for my elevation too.

I would expect they used roughly your roof angle. Typical suburban type house is more likely 18.4 degree or 22.6 degrees.

So I would start with the results from that to back-calculate what they used for losses due to shading.
Possibly they even have the information in the paperwork they gave you on what they used for azimuth and tilt.

Anyhow, bottom line, I feel I can assume 5000 kwh/year, since their 4452 is more or less guaranteed, and my shed/carport location will be a bit sunnier.

You can probably assume even more than that. Your shed/carport/solar-structure is (probably) going to be constructed to point due south and probably going to be constructed with a 36 degree tilt, so it's likely you have >15% difference just from that. (6179kwh vs 7344kwh implies >15% difference)

Complicated to figure my annual dollar savings, because I'm on a 3-tier "EV" plan, and the tiers affect not just what I don't use from the POCO, but also what I sell back to them.

Do you get to sell back to the POCO at their price?
(Some locations you do, some you don't - depends on POCO and the laws for where you are)

If you do, I think you can get an hour-by-hour guesstimate from PVWatts, as a spreadsheet, then apply the rates to it to see what it gets you.

$.32/kwh is a fairly high rate - so probably solar is economically viable even with some morning or evening shading.

[RShackleford]

Ok, I set losses to 10%. Roof-mounting. Set tilt to my roof's slope (5/12 or 22.6 degrees) and azimuth to the direction the roof faces (15 degrees east of due south, or 165) - assuming they ain't doing any fancy mounting. I come up with 6408, so they're assuming about 30.5% shading loss. Oh, I assumed "standard" panel type; would that be the SilFab Mono 310 they spec'd ? So then I set the azimuth to 180, the tilt to 36, and specify open-rack mounting, and I get 6656. Put in that same loss factor for shading and I get 4624 kwh/year. But since open-rack location will be sunnier, maybe 5000.

Yeah, the sell-back occurs at the retail price, and with TOD applied, so I'll be selling back at $0.32/kwh on summer afternoons (to the extent it's not running my AC, except I can do what I do now, keep the house chilly overnight, good for sleeping, and turn AC off at beginning of peak (1pm), and house stays nice till end of peak (6pm). Yeah, PVWatts will give hourly data, but I'm gonna have to add estimates of my hourly consumption (tricky), and do some Excel kungfu (I'm good at), to get meaningful dollars from that.

[Ampster]

.......................
It's definitely not all about money. Partly to put my money where my mouth is on energy/environmental stuff. Partly for some self-sufficiency. And yes, I realize I won't get electricity when the grid is down, and that it'll cost a bit more to get that. My plan is to start out with this grid-tied system, and add multi-mode later when hopefully there are better/cheaper options for that.

Those are reasonable goals even though some may disagree. You didn't mention hedging your energy costs in the future and at least in California almost everyone agrees that the Power Companies are going to be raising rates in the future as they have in the past.
If what you mean by multi-mode is to be able to add batteries to leverage your solar when the grid is down then you have options with both Solar Edge and Enphase. If you use SolarEdge I would suggest you buy a StoreEdge inverter which would give you the capability of adding an LG Chem battery or the SolarEdge battery that is supposedly coming out in 2020. Enphase has announced their IQ-8 system which also includes batteries and offers some backward compatibility with IQ-7 micro inverters. You are not limited to those options because there are still bimodal/hybrid inverter systems that can be AC coupled to any system but that option may be more expensive than the prospective ones that SolarEdge and Enphase may come out with in 2020.