New 4kW solar install - understanding the data

oldskoo1 · 21 May 2015 at 8:47PM

3 days ago i had my 4kW array installed.

16x 250w PhonoSolar Onyx
4000TL-D Samil SolarRiver inverter

Today i generated 24kWh and i managed to capture plenty of data that i can begin looking at.

At this time i don't fully understand how to read the data and i would like to understand a few things.

First i would like to understand the dips. Unfortunately i was not home today so i couldn't put this down to cloud cover. Today we did have cloudy conditions, mainly light cloud and haze.

I would just like to understand, can these dips be the result of overheating and a reduction in efficiency, as just before the dips were peaks.

Also is there any more observations we can draw from this graph. Peak sustained power look ok?

Grid current just for reference

The next observation was shading morning and evening on my south facing installation which was at the time a real concern and drove the position of my panels. Early, before 8:45am the right side of the roof gets a minor shading from the chimney at the top. Late after 7:20pm i get a shade on the lower left part of the roof, gradually going upwards. Expecting more generation later on in the day i decided to mount them high an in the future potentially add a 3rd row for a 6kW array.

Judging by the AC Power and string power graph, there doesn't seem to be any significant influence from this shading of generation figures. So does this back up what i was advised here, that the perfect no shade generation at these times have such poor solar energy naturally from low sun passing through the atmosphere that shading makes very little to no difference? I'm assuming this is the case for a south facing install. A west facing partial shade might have more of an issue?

If you look at the string power, there is no difference between the two strings at 8:45am (and before) or 7:15pm and after. I know just after 7 it drops a lot, but i was home and cloud came in after 7pm. But curiously, you can see in the photo that the left array got shading and the right was not yet, but seemingly no difference between strings. So unless i'm configured wrong, it seems irrelevant.

Here is the shading at 7:15pm

And the string voltages - assume these look ok?

p.s.
The strings do show a higher power at all times (pretty much) for string 1. This is apparently a common issue with Samil inverters, reading 0.5amps higher on string 1. It is over-reading, string 2 is accurate.

I also confirmed that my generation meter is showing 0.5% higher generating than my inverter.

Raw data for string power/voltage & grid data.
https://docs.google.com/spreadsheets/d/1rAlMLJ61RCbbZd-cix5gzZ2hm6CEHtpWpeMMZvsnkpg/edit?usp=sharing

Thanks for reading this geeky post!

nigelpm · 21 May 2015 at 9:17PM

Highly likely the dips are cloud.

I'd wait for a no cloud day and then do the analysis.

oldskoo1 · 21 May 2015 at 9:20PM

nigelpm wrote: »

Highly likely the dips are cloud.

:T very pleased with that as filling in those dips will give me a good idea of my maximum potential generation on a perfectly clear day with a not too warm temp to keep the panels cool.

rogerblack · 21 May 2015 at 9:22PM

The eye is a truly terrible light meter.
A halving of light is not very perceptible to the eye.
Just before a cloud obscures the sun, it gets very bright.
This reflected light adds to the solar light, causing a small peak before the cloud obscures the sun.
Overheating will never cause effects like that, they are absolutely typical of cloud.

http://pvoutput.org/intraday.jsp?id=8689&sid=7055

random local (to me) site to compare to for today.

Click 'prev day' a few times.

oldskoo1 · 21 May 2015 at 9:25PM

rogerblack wrote: »

The eye is a truly terrible light meter.
A halving of light is not very perceptible to the eye.
Just before a cloud obscures the sun, it gets very bright.
This reflected light adds to the solar light, causing a small peak before the cloud obscures the sun.
Overheating will never cause effects like that, they are absolutely typical of cloud.

http://pvoutput.org/intraday.jsp?id=8689&sid=7055

random local (to me) site to compare to for today.

Click 'prev day' a few times.

thanks! thats entirely the knowledge i was hoping to learn after sharing this data

:beer: