The effects of shading on solar PV

pinnks

I have not found a thread dealing with this issue, so thought I would share my experience. Apologies if my search abilities are lacking and this has already been done to death.

Anyone thinking of installing solar PV will receive advice about shading from their prospective installer but that advice will almost certainly be based on a computer model fed with data from Google maps plus minimal measurements taken on site.

I obtained several quotes for my installation but only one firm referred to shading from a neighbour’s house during the middle of winter for an hour or so each day and the impact of other houses was not mentioned by any of them. It is not until the system is installed that you begin to understand the impact of real shadows that objects will cast on your roof.

I have two solar arrays on two roofs at 90 degrees to each other, with a chimney between the two. One is 3.5 kWp; the other 1.75 kWp. Attached to the chimney was a large, high-gain TV aerial with a rear plate about 80cm x 80cm. While the shading models produced by the installation firm took account of the chimney they could not, realistically, take account of the aerial.

The effects of the aerial shading on the larger array are small given that the roof is WSW-facing, with chimney at the south end. You see a dip in production as the shadow moves round but as the roof does not get direct sunlight until late morning when the shadow has passed, overall production in the morning is only about 10% of rated maximum, so the additional impact of any shadow is minimal.

The impact on the second, SSE-facing, array is significant. In May the chimney shadow hits one panel at about 4pm but the aerial shadow races ahead of that, hitting that panel at about 1pm. Production falls almost instantly by about 1kW and stays at that level until the chimney shadow comes round, when it drops again. The rate at which production falls is the same for each shadow.

So, that aerial was costing something like 3 kWh production per day for, say 180 days per year or, at 16p per unit for the above 4kW tariff, something in the region of £85 per year. It also has a small impact on the main array of perhaps £10 per year. The result? The aerial just had to be moved to the other end of the house where its shadow no longer causes any issues. Cost me about £20 for materials and a shed load of nervous energy to work up a ladder at over 8m but worth it in my view.

And the lesson from all this? Avoid shading where possible but if the cause can be moved or eliminated, do the maths and consider making the necessary changes. Even if I had employed an aerial company to do the move it would have paid for itself in 2 years!

I would have popped in a couple of photos but can't figure out how to do - shame.


Gary

asturdy2

same here, had a metal chimney on the south side of our house, cost £250 to be moved and is worth every penny. Installers said it would have an effect but not as much as it did do before it was moved.

jimjames

Even a window open can affect the output from shading the panels!

http://uk-solarpanels.blogspot.co.uk/2012/07/effect-shade-solar-panels-output.html

pinnks

I thought I would come back to the issue of shading and the possibility of retrofitting Power Optimisers on a non-SolarEdge inverter.


You'll have to do a bit of work to read this one though as I took a bunch of pictures today, added narrative as the shading occurred and saved it as a PDF behind the link below.
https://dl.dropboxusercontent.com/u/14870257/More%20on%20the%20impact%20of%20shading.pdf


The question is, based on the impact of the shading on my SSE 1.750kWp system with SunnyBoy 1600 Inverter, would it be worth even considering a PO retrofit?


Happy reading...

tunnel

I have to say that I'm quite taken aback by the little amount of shading to the panels that affects your system to bring down the generation. Before retro fitting my SSW with solaredge it used to take at least 15-20% of the first panel going into shade before the rest were pulled down to its level. That was using an SMA inverter(now used on the WNW).


Can you not compare your generation figures to that of PVGIS to see roughly what the difference is. If the other system is performing as it should then any difference should come down to bu$$er in question, at least you'll have some semi accurate figures to see if your retro plan is feasible.

pinnks

I was equally surprised at the apparent impact of such a small amount of shade - it does seem extreme.

I had not thought of using PGVIS, so thanks for that tip. I have now done so, with some equally unexpected results.

The system went in in March last year and was projected to produce 1,621kWh. It actually produced a few kWh more than that despite the impact of the TV aerial mentioned in the first post in this thread.

Using default settings in PGVIS for Swindon, 1.75kWp, -20 orientation and 45 degree slope it returns 1,670, so a mere 50 more than I achieved with that TV aerial shading for the first 3 or so months from memory.

Not sure what to conclude from that...

tunnel

Did you check the month to month results in with that? You may find the winter months aren't that affected(my ESE has a hip roof that provides shade similar to yours only in winter months)


We did have some cracking months last year so it may be better looking month on month. Have a particular look at July last year which was way over PVGIS for most of the posters on the renewables thread and see how you compared, mine was over 20% up on PVGIS for that month, how does yours compare?

pinnks

Even more intriguing. Using Jan to Mar 2015 and the rest from 2014 gives the table below. April and May could be explained by the aerial (or the weather of course...) The other months are I think, inconclusive. Maybe I need to monitor for 10nyears and then decide based on the average - lol


Month PGVIS Actual
Jan 62.5 73.7
Feb 81.9 82.7
Mar 157.0 140.8
Apr 188.0 148.3
May 200.0 180.7
Jun 199.0 217.4
Jul 198.0 231.5
Aug 174.0 176.5
Sep 157.0 139.4
Oct 116.0 102.4
Nov 75.9 57.6
Dec 56.8 64.5

pinnks

That didn't go well. How do I insert a table of text?

EricMears

pinnks wrote: »

Even more intriguing. Using Jan to Mar 2015 and the rest from 2014 gives the table below. April and May could be explained by the aerial (or the weather of course...) The other months are I think, inconclusive. Maybe I need to monitor for 10nyears and then decide based on the average - lol


Month PGVIS Actual
Jan 62.5 73.7
Feb 81.9 82.7
Mar 157.0 140.8
Apr 188.0 148.3
May 200.0 180.7
Jun 199.0 217.4
Jul 198.0 231.5
Aug 174.0 176.5
Sep 157.0 139.4
Oct 116.0 102.4
Nov 75.9 57.6
Dec 56.8 64.5


(That didn't go well. How do I insert a table of text? )

I think we can probably all see what sort of table you were trying to achieve so pretty formatting isn't really all that important.
However, if you do want to display it as a neatly formatted table, probably the best way would be to display the table on your own computer monitor, do a screen dump and save that to an image file, store that file on another web server then use the url of that file to include the picture of your screen dump in a new posting.
(I don't think I said it was simple )


BTW, the pvgis figures are taken from average values for the last 30 years; you might need to monitor for another 30 years to get absolutely reliable figures - and even then there might be some discrepancies if (say) a couple of years were badly affected by a major volcanic eruption somewhere.

pinnks

Cheers. That version of the table was the manually post-posting edited version... I thought of doing a screen capture but thought that would be over kill, oh yey. Anyway, as you say, it is readable, so job done.

I get that data can vary and comparisons also but did the comparison following Tunnel's post. Trouble is, it seems inconclusive in terms of helping with the way the shading influences production, unless that it, I am misreading things?

I can see that production falls off a cliff as the panel(s) to into shade by looking at daily graphs but estimating the annual impact is less easy.


Back to the drawing board I guess.