Best power measurement method for max AC power - solaredge

[Snipely]

Hello all,
In the process of evaluating my system for clipping, I'm trying find the best way to measure true AC output. In the solaredge monitoring site, there are two ways to view - Dashboard daily power plot, and daily playback under the Layout tab. They are somewhat different. In addition, my reported info at pvoutput.org gives a third result. Is any of these more accurate, or does each have a "close enough" accuracy with different methodologies for calculating? I'm attempting to see how close to 5400W I'm getting, which is the published AC maximum for my SE5000A-US inverter.
I've read elsewhere that >nameplate is not even possible despite the number published on the inverter spec sheet (which is higher than the SE5000-H).

https://pvoutput.org/intraday.jsp?id=56013&sid=50713

Thanks

[Ampster]

I can see my clipping as a flat top to what would be a normal sine wave curve. Your DC to AC ratio is now 1.4 to 1 and mine is 1.5 to 1. My panels are all south facing. With your multiple orientations I would not expect your curve to be a symmetrical sine wave curve. Read Bruces posts where he has plotted his generation over the day. He is not concerned about clipping because he gets power at the maximum of his inverters for longer period of the day.

[Snipely]

Yep, I can see my clipping at 5.2 to 5.34 kW, depending on where I look. pvoutput.org, solaredge dashboard, and solaredge playback view all give different answers.
Trying to figure out how close I'm getting to the limit ot 5.4 kW in reality, or maybe just splitting hairs.

https://pvoutput.org/intraday.jsp?id=56013&sid=50713&dt=20200505&gs=0&m=1

[Ampster]

If I paid more attention to Calculus I would be able to calculate the area between the flat top and the presumed curve but my guess is it is insignificant. I actually asked my installer to calculate the annual production based on smaller DC to AC ratio with a bigger inverter and it wasn't worth the extra money. In the long term with cell degradation, it will go away anyhow.
The differences you see between the databases could be different sampling intervals. In the long term weather will be your biggest variable from year to year.

[J.P.M.]

Snipely:

It's a bastardization of the model, but one way to get a bit more info on whether or not your system is clipping:

1.) Run PVWatts for each of your arrays. Use 10 % system losses and the same orientation(s) as your array has. Get the hourly outout.
2.) Combine all 3 array's outputs on to one spreadhheet.
3.) Total all 3 hourly outputs into 1 system output.
4.) Find one or more dates where your weather was clear and PVWatts weather was also clear. Clear days are easiest to match up.Try to use dates where the ambient temps and wind were not too different from one another.
5.) Compare the modeled hourly output to your actual hourly output for those dates. If PVWatts numbers are a lot higher than your array is producing, your system may be clipping.

That method is vey approximate but mat yield some useful insights.

Because your system has 3 orientations - one of which is responsible for close to ~ 60 % of your clear day output - and they are all different sizes, a graph of your output on a clear day will look "flatter" and be skewed so that the point were the output is ~ 50 % of a clear day's output will be centered away from solar noon by an hour or so.

[J.P.M.]

I can see my clipping as a flat top to what would be a normal sine wave curve. Your DC to AC ratio is now 1.4 to 1 and mine is 1.5 to 1. My panels are all south facing. With your multiple orientations I would not expect your curve to be a symmetrical sine wave curve. Read Bruces posts where he has plotted his generation over the day. He is not concerned about clipping because he gets power at the maximum of his inverters for longer period of the day.

So others reading your post don't run off with incorrect information, clear day output from an equator facing PV array does not follow a sinusoidal output. To imply so is to imply incorrect information.

[Ampster]

So others reading your post don't run off with incorrect information, clear day output from an equator facing PV array does not follow a sinusoidal output. To imply so is to imply incorrect information.

@bcroe has probably played with more sine waves than you or I. Maybe he can elighten us about the risk if others running off with that information. Perhaps I should have said, that it looked similar to a parabola but not the same mathematically.

[J.P.M.]

@bcroe has probably played with more sine waves than you or I. Maybe he can elighten us about the risk if others running off with that information. Perhaps I should have said, that it looked similar to a parabola but not the same mathematically.

I'm sure he has, and he also has probably forgot more about mathematics than most folks will ever know, but that doesn't change the reality that your statement was simply inaccurate.

Others reading it may have been misled.
You could have said any number of things, including "my clear day output curve would look like a haystack were it not for the clipping" and have been as descriptive (and probably more accurate) about it's shape.
Once you bring more technical terms into the description or conversation, more responsibility to get it right is required.
Such clear day output curves may also bear some cursory resemblance to a normal distribution, but they are not.
The more technical the terms used to describe a condition, the more specific and accurate becomes the need to use them accurately.
I merely pointed out the error with some explanation of why I think it was an error.

If you're going to attempt to get semi technical, get it right or expect to get challenged. Accuracy counts.

This would be a better forum if you were less of a loose cannon and stuck to what you know and can back up.
To do less is discourteous and rude.

[Snipely]

I believe I'm clipping based on every chart available to me - i.e. the three I've already mentioned, plus the fact that my inverter is supposed to clip at 5400 W and I'm close to that but not quite there. Before I increased my array size, there was no flat top. I'm free and clear of any shading at 180 azimuth. clipping_yep.png

[Snipely]

Just to be clear, I'm not particularly concerned that I'm clipping, nor about the loss in production. That said, if I'm clipping below the inverter's published maximum (5200 vs 5400), I'd prefer to clip at the published limit and not below it. I thought I read there was a firmware tweak that could bump up the limit a bit. I'm cool with the readings being within sampling error - they are all fairly close anyway, within 200 W or so of the maximum 5400 W when I look at the maximum output within the flat top region. I just want to "keep the inverter honest" and make sure it's performing at its best and not limiting below spec vs at spec. Thanks for your input, everyone!

[Ampster]

Yes, you are clipping based on the flat portion of that graph. We can speculate about how much that actually is. You will never be able to get a precise number because of the sampling intervals. More importantly it will vary based on weather and time of year. The two most prevalent opinions about clipping are that it is bad or that it doesn't matter in the long run. I am of the opinion that it doesn't matter in the long run. Often the amount is insignificant and over time, panel degradation will result in less clipping. Another viewpoint is that your inverter is too small. However inverter sizes don't come in convenient increments and the next size inverter would leave you with unused inverter capacity. It all depends on how you want to view it.

I hope by now you also understand that the curve is not truly sinusoidal and that you have not made any important decisions based on my use of the term. Your curve looks more like the haystacks I used to see in the countryside after the cattle nibbled on a portion.

NOTE: My comments were composed before I saw your last post. I am glad your upgrade is getting you the results you hoped for.

[Ampster]

........
This would be a better forum if you were less of a loose cannon and stuck to what you know and can back up.
To do less is discourteous and rude.

It appears someone is trying to win the Sunkist look alike contest. Despite the value he brought in terms of adding to the stickies about the fundementals, I believe this is a better forum without his name calling and negativity. I trust the moderators to be the arbiters of whether information is dangerous or outside the guidelines of this forum. They have commented more than once about the snippiness of these type of conversations and have gone so far as removing irrelevant posts which I encourage.

We are talking about a term I used in a descriptive way to describe the shape of a curve. I realize that technically that term can be interpreted to be a mathematical curve that does have a similar looking shape. I would be happy to change my description if you can demonstrate how someone could make a serious calculation error to their detriment particularly since we were only talking about the area described as clipped. This is a similar tactic Sunkist used to intimidate people and prove in some way that he was superior. To me that is a sign of an inferiority complex.That is as far as I want to go into name calling.

[J.P.M.]

It appears someone is trying to win the Sunkist look alike contest. Despite the value he brought in terms of adding to the stickies about the fundementals, I believe this is a better forum without his name calling and negativity. I trust the moderators to be the arbiters of whether information is dangerous or outside the guidelines of this forum. They have commented more than once about the snippiness of these type of conversations and have gone so far as removing irrelevant posts which I encourage.

Your information was incorrect. Leave it at that and leave it alone.

[Ampster]

............. I'd prefer to clip at the published limit and not below it. I thought I read there was a firmware tweak that could bump up the limit a bit. I'm cool with the readings being within sampling error - they are all fairly close anyway, within 200 W or so of the maximum 5400 W when I look at the maximum output within the flat top region. I just want to "keep the inverter honest" and make sure it's performing at its best and not limiting below spec vs at spec. Thanks for your input, everyone!

I understand you want to get the most out of your system. Let us know if you get the upgraded firmware. I have a newer HD Wave SolarEdge 3.8 kW and it clips somewhat consistently at 3.9 kW. Yesterday my curve looked like a tall Fez hat.

Before any readers rely on that shape to do any serious calculations about clipping, I was referring to my grandfathers Fez which was for the Shriners club and had a different shape than today's Fez hats.

[Snipely]

Update at the solstice: Still at 5.15 kW max clipped AC output, not the published 5.4 kW, and no satisfying explanation from SolarEdge as to why the HD and A inverter models have different max specs (with the A showing 400W over nameplate power). The inverter is simply "working as expected". Not too worried about it, just puzzled.