Replacement SolarEdge inverter producing 10% less power?

[azdave]

increased ambient temps is definitely one factor which contributes to decreased production.

Here is my production graph for one full year with the daily high and low ambient temps included. We average 300 clear-sky days here in Phoenix so most of what you see are cloudless days. You can see my production in May begin falling off as ambient temps rise and summer approaches. Even as the days get longer it does not help overcome the efficiency loss due to heating. This is a simple string inverter system facing due south.
Usage SRP Power - Google Chrome 7292020 85621 AM.jpg

[J.P.M.]

Here is my production graph for one full year with the daily high and low ambient temps included. We average 300 clear-sky days here in Phoenix so most of what you see are cloudless days. You can see my production in May begin falling off as ambient temps rise and summer approaches. Even as the days get longer it does not help overcome the efficiency loss due to heating. This is a simple string inverter system facing due south.
Usage SRP Power - Google Chrome 7292020 85621 AM.jpg

Dave: I sure don't disagree with the idea that heat reduces array output, but my comment to theoak's post related specifically to his statements about output being related to some vague parameter he calls "distance to the sun". I'm only referring to one variable and a small one at that, not about output as f(other variables that affect array output).

As I alluded to in a prior post to this thread, without instrumentation to enable estimates of the major variables that affect array efficiency, over relatively short calendar periods of, say, a month or two, it's near or actually impossible to separate out the effects of varying solar incidence angles on array output from the other things that effect array output.

And, because of the slower changes in solar geometry, particularly near the solstices (were we are now), such effects (due to the solar geometry changes), which are relatively minor over periods of a month or two relative to, the effects of other variables such as clouds/wind, etc., such effects from changes in the solar geometry will be masked by the effects of the other variables.

Additionally, for arrays that are mostly and/or generally equator facing and tilted at some angle to the horizontal, for times of the year about 6 weeks to maybe 2 months on either side of the summer solstice, the daily clear sky insolation will be less on the solstice than for dates on either side of it. The effects are small near the solstice for such orientations but increases as time to/from the solstice increases.

That means for times like now (late July), and contrary to what theoak claims, rather than output dropping because "we're moving away from the sun..." (which BTW, is also incorrect - solar perigee happens ~ July 4), potential (clear sky) solar insolation on most commonly oriented arrays has been increasing since ~ 06/21 and, depending on particular location and array orientation will probably increase until the 1st or 2d week in Aug., again, depending on orientation.

As an example, for my array only, but I'd suggest it's representative of the general trend for commonly oriented arrays, the ratio of day long clear sky (Plane of Array Insolation)/(Global Horizontal Insolation) for a the following dates may be useful to help understand what's going on:

06/21: 0.942
07/21: 0.966
08/21: 1.033
09/21: 1.134

What that says is that GHI decreases relative to the POA insolation (or that POA increases relative to GHI) on a commonly oriented array the further away from the summer solstice for dates relatively close to the solstice.

To be clear and maybe redundant, that's what the effects of the solar geometry will do to potential array output. The effects will be relatively small and so most of the time masked or buried by other driving inputs such as clouds, wind and ambient air temp., array fouling and other stuff, but the fact of the matter is that solar geometric effects will most likely work in a fashion opposite to what theoak thinks and has written.

That's all I was trying to say.

[azdave]

I was just showing a basic graphic for the simple folk here (like me) that output drops very measurably when the panels are really hot. I wasn't expecting the "Spanish Inquisition" as they say.

All good plus I learned something new today.

[J.P.M.]

I was just showing a basic graphic for the simple folk here (like me) that output drops very measurably when the panels are really hot. I wasn't expecting the "Spanish Inquisition" as they say.

All good plus I learned something new today.

I believe I understand and appreciate what you are writing. Not bustin' on you either, thus my comment about temps. being important.

I'm simply trying to be clear about the influence and limits of solar geometry on P.O.A. irradiance/insolation with respect to other inputs and why theoak's statements were incorrect.

I would have let it go with my original comment but theoak compounded his error with more error and not liking factually inaccurate stuff, and in the usual spirit that anyone can take what they want of my stuff and leave the rest, I attempted to fill in the reasons why I believe theoak's statements relating to the effects of solar geometry on array output are incorrect. I criticize theoak's statements. I don't impugn his character or challenge his honesty. Just the accuracy of some of his statements.

[theoak]

From the June solstice to today I have about a 4% drop in direct radiation, for my latitude and tilt assuming perfect conditions and perfect collection. Now, throw in some points for heat ... that gets ball park at the 7% drops folks are seeing from a given day to the given day minus a few weeks when the inverter failed.

https://www.pveducation.org/pvcdrom/...lar-insolation

Of course, any concerns, I would hope the OP would contact their installer and ask. They probably paid a pretty penny and it is totally okay for the installer to earn a bit more of the money paid.

@J.P.M. I totally respect you. At 13k posts you definitely have a lot of knowledge to share as already shown in just this thread. I look forward to learning more.

[Mike90250]

And a lot can be laid onto measurement tolerances, if the old inverter read a couple % high, and the new one reads couple % low easily 5% error there

[J.P.M.]

From the June solstice to today I have about a 4% drop in direct radiation, for my latitude and tilt assuming perfect conditions and perfect collection. Now, throw in some points for heat ... that gets ball park at the 7% drops folks are seeing from a given day to the given day minus a few weeks when the inverter failed.

https://www.pveducation.org/pvcdrom/...lar-insolation

Of course, any concerns, I would hope the OP would contact their installer and ask. They probably paid a pretty penny and it is totally okay for the installer to earn a bit more of the money paid.

@J.P.M. I totally respect you. At 13k posts you definitely have a lot of knowledge to share as already shown in just this thread. I look forward to learning more.

Understood. Not a problem with me and not personal at all. Just clearing incorrect statements.

I'm now wondering how you figure you're getting - or how you determined you're getting ~ 4% less insolation since the solstice. Do you have access to a reasonably reliable and nearby pyranometer ?

Are you possibly confusing Global Horizontal insolation (GHI) with Plane of Array (POA) insolation ?

Unless the array is shaded so as to drastically affect things, for most common array orientations (that is, mostly equator facing) in the northern hemisphere this time of year, while GHI decreases at a slow rate immediately after the solstice for ~ 6 weeks or so. POA usually increases slightly for about the same length of time +/- a bit, maybe a bit longer for most (but not all) common array orientations. POA insolation is what is used for array calcs. unless the solar array is horizontal. Then POA == GHI.

When all's said/done, unless a site has instrumentation to reliably and consistently measure the inputs that affect array performance the most (insolation/irradiance. ambient temp. at the array, wind vector as a minimum), a way to account for any shading, and also a way to measure/estimate array fouling, estimating array or system efficiency is a crapshoot. I watch my array like a hawk with a lot of instrumentation and measurement and I'm pretty sure the best I can do on an either instantaneous or long term estimates of system efficiency is maybe ~ +/- 0.7- 1.0%. Without the instrumentation or diligence, I doubt I'd be able to get much better than +/- 10%.

[theoak]

@J.P.N ... sure its all a crapshoot. We all are guessing at the end of the day.

@Undebt for sanity checks sometimes I use PVOut and search of folks in my postal code. I try to search for systems that are a similar kWp and azumith. If folks around you are trending the same as you I would not be too concerned. Thus far if I see something that I interpret as odd with my system, I will look at similar systems in my area and I see the same general behavior. Once you upload your data, PVOut has some some comparison graphs and charts that let you compare yourself to others around you. Just an idea ...