SolarCity Isn't Looking So Terrible

[sensij]

Ah, so the rep is indeed BSing. Can't trust any of these installers..everyone has their own little lie.

I told the rep SolarCity most likely will give me the regular string inverters and not solar edge with optimizers due to the fact that I have very little shading..and the rep said "but clouds, we get clouds, that's a shade". I didn't think any optimizers would help with that.

Despite J.P.M.'s objections, I can share data that shows a Solaredge system can put out more instantaneous power in cloudy conditions than a string inverter could. The question is not whether or not it produces more, but whether it produces a meaningful amount more over longer periods of time. The answer to that question is much murkier.

[Singuy]

That wasn't what he said, but he can certainly defend himself lol. Have you searched this forum yet for Solar City?

Yup, in fact the title of the post alludes to the notion that everyone on the forum hates SC but from my experience, they are so far the lesser of all evils.

[J.P.M.]

Despite J.P.M.'s objections, I can share data that shows a Solaredge system can put out more instantaneous power in cloudy conditions than a string inverter could. The question is not whether or not it produces more, but whether it produces a meaningful amount more over longer periods of time. The answer to that question is much murkier.

Do you have any ideas why that data may show that ? Same panels ? Same location/orientation? I'd be willing to agree that a Solaredge inverter/optimizer system may be more efficient than any given string inverter system, but I'd suspect that difference would be the same under most all sky conditions. I'm also not sure how you'd measure the output of individual panels in a string inverter system.

However, beyond (or before?) all that, I find the notion that some portions of an array will have more (or less) instantaneous or long term production simply because of cloudy or partly cloudy sky conditions alone to be a hard sell. That would mean that clouds and the vapor pressure of the moisture that is what clouds are would need to have a greater pressure gradient than is possible from the laws of Thermodynamics. Even if such a fine demarcation line were possible, such a fine edge would be diffused by the atmosphere over the distance between most clouds and the ground, similar to the way sharp building shadows cast a less distinct shadow line farther from the building, identical; to the way solar eclipses cast a full and partial shadow on the earth. Or, for another example, observe how a hill casts a shadow onto an adjacent valley or lower elevation a mile or two away. The sharpness of the edges of shadows decrease in definition the farther the shadow is from the object. That's mostly a consequence of the diffusion effects of the atmosphere. Finally, Even if such array differential instantaneous were possible to be caused by partly cloudy conditions, I'm not sure it would be measurable, simply because the edge of a cloud, by virtue of the cloud's velocity relative to the array would seem to me to mean that such an edge would pass over an array in a matter of seconds, both making the required instantaneous measurement of the irradiance impossible with current instrumentation, and also tending to smooth out or smear the readings of the individual panels.

So, I'd appreciate a sharing of the data. I'd either learn something and stand corrected or be able to comment on what's presented from the standpoint of what I think I might know.

BTW, Sensij, I'm not steamed in the least, I just don't think intermittent clouds or full sky cover alone could explain any performance difference by virtue of the differential advantage that the SsolarEdge or, for that matter micro inverter systems may offer. Steady or slowly moving shadows for sure, although of , IMO, somewhat limited financial advantage depending on particulars and costs. And, as for costs, if shading is bad or great enough, at some point I'd question the wisdom of any PV, as it might be a just a bad application that micros or SolarEdge can't salvage. But that's a different subject.

[ButchDeal]

Solaredge system can produce more in low light situations simply due to the built in buck/boost features of the optimizers. This allows the pv modules to operate in a much wider voltage range and can easily be seen in the often very early start up of solaredge systems.
If it is very meaningful amounts of energy would depend on how many hours the modules are under clouds and the inverter system you are comparing to.

[J.P.M.]

Yup, in fact the title of the post alludes to the notion that everyone on the forum hates SC but from my experience, they are so far the lesser of all evils.

Opinions vary. Everyone has a right to their own. Still, having a right to an opinion does not make it a sensible one.

My wife tells me that all the time, so it must be true.

[J.P.M.]

Solaredge system can produce more in low light situations simply due to the built in buck/boost features of the optimizers. This allows the pv modules to operate in a much wider voltage range and can easily be seen in the often very early start up of solaredge systems.
If it is very meaningful amounts of energy would depend on how many hours the modules are under clouds and the inverter system you are comparing to.

I was speaking to Sensij's comment: "I don't think there is much doubt that on a partly cloudy day, giving each panel its own mppt will result in more raw output.", and others after that that spoke to the situation. That idea, i.e., that cloud cover, or broken clouds can impact adjacent instantaneous panel performance, is what I doubt.
I doubt it because the idea seems to imply that adjacent panels will have more instantaneous output by virtue of having greater irradiance over a very short period of time, say, 1 minute or so, or less, which is probably close to the limits of measurement, simply by virtue of each panel having its own mppt. I understand SolarEdge may have other advantages of which you speak, and those may yield more output. I get that. But Sensij's statement made no mention of such reasons, leaving me to think He was addressing the bogus statement e that nthe impl,, okld la M'. The idea seems contrary to common sense and also runs counter to my experience over many years.

I will not try to side track the subject as some others do if/when things heat up sometimes, but I will note that any performance enhancement from buck/boost features is likely of be relatively small in terms of annual output, maybe a couple/few % ? If so, and also given that most any array's output under cloudy skies is likely to be, by my estimate and experience, and from models something of the order of 10 % of annual output. So, a few % of 10 % is what, 0.003 ? or 30 kWh/yr. on a 10,000 kwh/yr. system output ?

[sensij]

I was speaking to Sensij's comment: "I don't think there is much doubt that on a partly cloudy day, giving each panel its own mppt will result in more raw output.", and others after that that spoke to the situation. That idea, i.e., that cloud cover, or broken clouds can impact adjacent instantaneous panel performance, is what I doubt.
I doubt it because the idea seems to imply that adjacent panels will have more instantaneous output by virtue of having greater irradiance over a very short period of time, say, 1 minute or so, or less, which is probably close to the limits of measurement, simply by virtue of each panel having its own mppt. I understand SolarEdge may have other advantages of which you speak, and those may yield more output. I get that. But Sensij's statement made no mention of such reasons, leaving me to think He was addressing the bogus statement e that nthe impl,, okld la M'. The idea seems contrary to common sense and also runs counter to my experience over many years.

The observations I'm describing were made with the Davis sensor, located adjacent to the array. The output of panels most nearly located to the sensor follow the measured irradiance well (accounting for POA vs GHI). The panels further away in the array do not track as well in cloudy conditions.

We've both seen edge of cloud reflections that can elevate POA irradiance over 1000 W/m2 for 5-15 min time periods. What I've seen suggests that those reflections do not necessarily distribute irradiance evenly over a 40 ft array. I don't see as much of a spread in performance on clear days, so I'm asserting some f(clouds) here. My 40 ft, 12 panel array would be a single string on a conventional string inverter.

In a shade free array I think it is more likely that temperature driven performance differences between individual panels due to mostly airflow induced gradients are a better source for any Solaredge claim of greater production than cloud behavior. I think the "early start up" idea is a stretch.

I do not think any of these differences add up to a meaningful total kWh over the course of the year, and do not think that those performance differences in a shade free array would justify choosing their system over a simpler one without seriously considering the long term risks.

The OP's salesman might have been misrepresenting the difference as a sales tool, but even if it is not financially significant, I think a difference can be proven to exist.

Putting together my data will take some time, but I will get it posted eventually.

What we know for sure in a string system is that all panels in the string will operate at the same current. If I can show that the mppt state for the individual panels have a spread in the current generated, and that spread has some f(clouds) to it, I think that is a path to proof of what I've written. If the string inverter is forcing some panels to operate away from their maximum power points, and the Solaredge system is not, then for those conditions, the Solaredge system will outproduce.

[bcroe]

Optimizers will help get max efficiency if some panels are shaded. They will even help if panels are
really badly matched. But clouds are not shade, except possibly during some negligently short
transition of edges covering part of a string. Optimizers maximize efficiency, but they can't find any
extra energy when everything is equally clouded over. Bruce Roe

[J.P.M.]

The observations I'm describing were made with the Davis sensor, located adjacent to the array. The output of panels most nearly located to the sensor follow the measured irradiance well (accounting for POA vs GHI). The panels further away in the array do not track as well in cloudy conditions.

We've both seen edge of cloud reflections that can elevate POA irradiance over 1000 W/m2 for 5-15 min time periods. What I've seen suggests that those reflections do not necessarily distribute irradiance evenly over a 40 ft array. I don't see as much of a spread in performance on clear days, so I'm asserting some f(clouds) here. My 40 ft, 12 panel array would be a single string on a conventional string inverter.

In a shade free array I think it is more likely that temperature driven performance differences between individual panels due to mostly airflow induced gradients are a better source for any Solaredge claim of greater production than cloud behavior. I think the "early start up" idea is a stretch.

I do not think any of these differences add up to a meaningful total kWh over the course of the year, and do not think that those performance differences in a shade free array would justify choosing their system over a simpler one without seriously considering the long term risks.

The OP's salesman might have been misrepresenting the difference as a sales tool, but even if it is not financially significant, I think a difference can be proven to exist.

Putting together my data will take some time, but I will get it posted eventually.

What we know for sure in a string system is that all panels in the string will operate at the same current. If I can show that the mppt state for the individual panels have a spread in the current generated, and that spread has some f(clouds) to it, I think that is a path to proof of what I've written. If the string inverter is forcing some panels to operate away from their maximum power points, and the Solaredge system is not, then for those conditions, the Solaredge system will outproduce.

Thank you for the reply. I apologize if this is an incomplete reply as I'm out the door soon. I'll pick it up on my return.

What is the relative position of the Davis to the array ? E,W,N,S , ?

As you know, I do not have individual panel monitoring. As you also probably remember, I do have well over 100 events of measurements of individual panel temps. on clear days at the time if minimum incidence, along with a bunch of other data collected at the same time over ~ a 12 minute period. In general, the backside panel temp. distribution from one panel to the next follows a profile of that of the temp. dist. over a heated flat plate exposed to fluid flowing over it and around it to a reasonable degree given wind vector variability. With an average wind velocity of 1.9 m/sec (std dev. ~ 0.9 m/sec), measured along the array N-S centerline ~ 4 ft. north of the array with the anemometer cups ~ 6" above the elevation of the top of the array, the difference in panel temps. varies, array leading to trailing edge from between ~ 1.5 C to ~ 7 C. as f(wind velocity). I haven't checked for any dependence of temp. gradient as f(wind direction). There may be some.

If published data on eff./deg. is to nominally be believed (and, off topic, but FWIW, I think I've measured mine to be pretty close to what Sunpower published in 2013), the delta T's I measured would indicate to me a panel output difference from operating temp. of between ~~ 0.6% to ~~ 2.8% over the array, depending on wind velocity.

So, I think I mostly agree with your opinion that most (measureable) individual panel temp. differences over an array, one panel to the next are wind driven, and without putting too many words in your mouth, panel efficiency, so too are the resulting variable portions of panel efficiency differences over an array, with some more or less fixed with laboratory or controlled environment measureable differences not only possible but likely.
.
Whether or not the SolarEdge claims are to be believed - and I'm not critical of what Butch is talking about at this time, at least partly because I haven't looked at what he references and so I'm ignorant about the particulars - but, and meaning no disrespect, I suspect any advantages he speaks of may be somewhat specious advert. hype, or at least dubious, or more theoretical than tested, and may have something to do with claims of, or actual restoration of, some of the efficiency normally lost by panels at lower irradiance conditions, which losses at least partially occur because of the cell's different response to different incoming wavelengths of light and thus efficiency as f(incident irradiance) which will be less favorable to panel efficiency under clouds or diffuse light regimes because of the different wavelength distribution of that light under cloudy vs. clear sky irradiance. But, That's more complicated than can be discussed here.

On cloud reflections: Because of the diffusing nature of the atmosphere and other phenomena I referred to in prior posts to this thread, I'd have a hard time ascribing cloud reflections to any differential P.O.A. irradiance on a relatively small array. Maybe on the big arrays I'll be looking at next week, but not a roofsize array. The projected area of a cloud reflection simply seems too big relative to a small array, and also probably too fast moving to allow any differential irradiance to react with one panel vs. another panel at the other end/edge of a rooftop array.

I also agree with your opinion that most any of these differences we're talking about amount to much, and I'd add that in most situations, it would take some careful analysis and some knowledge beyond what most users have to justify what may be extra expense to garner what's probably, at least in most viable situations, a meaningful and favorable change in cost effectiveness. If micros or optimizers added no cost, I'd still be weighing the unassignable cost of loss of reliability that the added complexity of micros /optimizers adds over a simple string inverter.

As I wrote, adding ~ 30 kWh/yr. to a 10,000 kWh/yr. array doesn't sound like it would be worth it to me. BTW, I got that 30 kWh/yr. from a SWAG of PVWatts output under cloudy skies for two areas I've lived in, one cloudy (Buffalo with 0.178 of system cloudy sky annual fractional output on an array tilted south at latitude under conditions of P.O.A < 250 W/m^2), and one about as sunny as it gets (Albuquerque with, 0.054 of system cloudy sky annual fractional output tilted south at latitude.), with an average of the two being ~ 0.11 of annual output under "cloudy" skies.

It looks to me like the one or more of the following may have occurred between the OP and that peddler: The peddler is full of what's politely called road apples, the peddler is ignorant of what he's saying, the OP is confused by his solar ignorance and misunderstood what the peddler was saying or, all or some parts of all of those possibilities.

[Singuy]

A little update on my neighbor's panels installed by SC. One was installed on shingles 3 months ago, other one on tile roof 5 months ago. Our city was battered by hurricane Irma with a 120mph gust and 100mph sustained for a good 15hrs. We were 30 miles away from the eye, and got hit with the eye wall pretty hard.

So far from surveying these two neighbors, all panels are intact with no problems and caused no roof damage/leaks.

[J.P.M.]

A little update on my neighbor's panels installed by SC. One was installed on shingles 3 months ago, other one on tile roof 5 months ago. Our city was battered by hurricane Irma with a 120mph gust and 100mph sustained for a good 15hrs. We were 30 miles away from the eye, and got hit with the eye wall pretty hard.

So far from surveying these two neighbors, all panels are intact with no problems and caused no roof damage/leaks.

I'm happy for your neighbors. How about other neighbors ? Unfortunately, some panel arrays perhaps failed ?. Most survived ? Your point ?

[Singuy]

I'm happy for your neighbors. How about other neighbors ? Unfortunately, some panel arrays perhaps failed ?. Most survived ? Your point ?

I'm looking for evident of shoddy installation job that surely was put to the test during a major hurricane. These are the only two who used SC in my neighborhood that I know of.

Also you wanted me to come back for future updates...even if I say that SC was excellent, you can always play it off as "well that's just a sample size of one"....so what's the point of doing updates then?

[J.P.M.]

I'm looking for evident of shoddy installation job that surely was put to the test during a major hurricane. These are the only two who used SC in my neighborhood that I know of.

Also you wanted me to come back for future updates...even if I say that SC was excellent, you can always play it off as "well that's just a sample size of one"....so what's the point of doing updates then?

More information is always welcome. Not all of it is conclusive. None of it hurts, provided it's honest.

[Singuy]

No this is not conclusive..technically nothing is conclusive as more and more data can one day may even disprove gravity. Just thought you guys can use some rare data from actual hurricane survivors...not something you see everyday.

[J.P.M.]

No this is not conclusive..technically nothing is conclusive as more and more data can one day may even disprove gravity. Just thought you guys can use some rare data from actual hurricane survivors...not something you see everyday.

As I wrote, more information is always welcome. As for "us guys", I'm probably the most concerned or vociferous among active posters with respect to SolarCity, although I'm sure others may be curious about first hand accounts related to their shenanigans. I think SunEagle may also have gotten a direct, or close to it from Irma. Glad/Hope you did OK.