A few days ago the discussion, prompted by Damon again, about the validity of East facing PV came up. I think I mentioned, though may have just thought it, that if Grid Balancing is to become a feature then there would need to be some East and some West facing as well as South facing.
So I did what I do, went and found some solar data and started to look at it.
The preliminary results are below:
The "Solar Resource" chart shows the comparative difference between the two locations (Norfolk and Cornwall).
The "Fraction of solar resource before and after noon" chart shows the distribution. I have time shifted by half an hour for the differences in longitude, not spot on I know but close enough.
There is also some data missing from both sets that I have not yet corrected for, but apart from Cornwall's Month 1 (January) there should not be enough data missing to make a significant difference.
A quick look at the data seems to show that there is little difference in resource distribution between Norfolk and Cornwall, Cornwall being better in the summer. Both locations seems to be better in the afternoon, or more West facing.

The pictures.

Heard about it on the wireless this morning, 600 pages of rubbish for 10 of interest.
Shall put it on the Reader I think.

East facing pv will be affected by morning mist, which is generally not something that would be repeated in the evening so west facing would be unaffected by this phenomenon. I would suspect that this could have a large variance on a very local basis though.

Lots of afternoon and evening mist here. Geography and topology greatly changes things.
St. Just never sees any sun, not ever.

Better hurry up on the PV.
Do you have any shading on the proposed side? That can kill some systems. I was amazed how much a few blades of grass covering a ground mount system caused the panel to warm up on the shaded bit.

No chance of this year, too much to do to repair the roof, which we can now do with confidence since the golf course which caused all the damage has closed down. At times, this place was like a suburb of Beirut. Tiles smashed, one of the large sash windows in the kitchen broken, vehicles damaged.

<blockquote><cite>Posted By: SteamyTea</cite>Lots of afternoon and evening mist here. Geography and topology greatly changes things.
St. Just never sees any sun, not ever.</blockquote>
Oh, you poor things! Meanwhile up in the far North I've seen more Sun in June than in April+May together (well more PV generated). And more rain possibly ...

The only sizeable bit of clear roof left I've got is slightly-S-of-E and I've been trying to work out if it would be worth putting more PV &/or ST on it. I've been wondering for a while about checking am -vs- pm generation on my slightly-W-of-S panels, suspecting that mornings were better: but I didn't think I had the data to check it.

But following a prompt from Damon, I've found I have the full minute-by-minute log of inverter stats starting Sun Mar 21 06:12:01 2010

The 31.9MB file is linked from http://www.ccandc.org/cgi-bin/pv?START=end-1d&END=now
if anyone else wants to play with it :-) The usual provisos: the data only applies to me (and probably the array in the next street N of me) ~12km from the North Sea, just out of the Tyne Valley, both of which lead to very localised weather.

Now to work out what I want to know, then how to extract it from 597,653 rows of data!

-- Chris

Looks like you get sunnier mornings.

I am still playing with your data, going to try and put a magnitude on the difference.
Does your logger change the clock to the really, really, annoying and pointless BST?
And are you 5° East or West of South?

<blockquote><cite>Posted By: SteamyTea</cite>Change of plan, have checked it out and it is the other way around (was to do with blanks and zeros, text and numbers).</blockquote>

Panels face a little west of south, so in theory I expect to get better exposure p.m. but my impression was that a.m. are often clearer than p.m. This clearly says not. Hmmm subjective impressions 0, reality 1 :-)

Looking at the program (fslurp) that talks to the inverter it does
InverterGlobals.timestamp = time(NULL);
then
timeBuf = ctime_r(&InverterGlobals.timestamp, ctimeBuf);
So yup, it uses localtime :-/

Hmm now that I've found that, I could just change that to gmtime_r() ... but I also find there's a new version out so perhaps I'll check that out first.

Thank you once again for extracting meaning from raw data.

-- Chris

<blockquote><cite>Posted By: SteamyTea</cite></blockquote>
I hope this wasn't an overnight processing run :-)

<blockquote>It is pretty marginal up your way though.</blockquote>
I suspect that PV altogether is a pretty marginal proposition up here :-/

<blockquote>Is there any shading from anything at all?</blockquote>
There is shadow from the dormer on one panel in the morning and one in the evening as the sun crosses the ridgeline, but shouldn't last long (15-30min?). On clear days I imagine I can see the clipping as a notch out of the bell-curve.
In winter there may be a lot more from buildings on the other side of the street -- I am slightly higher but the sun doesn't get very high: it rises and sets south of the ridgeline.

<blockquote>You do seem to get extra morning sun in the winter, so that may be what you notice, unless like me you are usually awake by 5:30AM.</blockquote>
I may be awake at that time, but it's the walk to the Metro ~7am (later in winter!) when I notice the weather :-)

I noticed on Saturday morning that at 8am the sun was a/ well up, b/ v.bright, c/ still behind the ridgeline, i.e. not hitting the panels :-( So could both be true? it is clearer in the mornings but the sun doesn't hit the panels until after 9am so the extra time in the afternoon/evening (from the slight W view) wins over the clearer morning.

As you say: needs more data.

All this slicing and dicing of data is making wonder what tools and techniques I missed skyving most of my A-level maths lessons!

-- Chris