There were two articles in the GB Magazine (Vol 18/4 and 19/2) from Jerry Clark about the output from his PV installation in Cornwall. In 2009 we had an installation here in Norfolk. What has surprised me is the difference between the two installations month by month, looking at the percentage of the annual total produced each month. For the months Jan – Dec, Cornwall produced 3, 5, 10, 12, 13, 13, 11, 10, 10, 7, 3, 3% each month, whereas Norfolk has produced 2, 3, 8, 14, 15, 15, 14, 10, 8, 6, 3, 2% each month. Both installations used Fronius inverters – Jerry Clark used Sanyo panels and we used Mitsubishi. What results do other people have?

Also, we have recently installed a second system, facing SE – the first one faces SW. Allowing for the difference in the number of panels, we are getting 4% more from the SE system than the SW system. So far, that is only over six weeks, but again, does anyone else have any detailed output figures?

Damon, thanks for that. Your figures (on the same basis as before) are 2, 2, 7, 12, 15, 17, 16, 12, 9, 4, 2, 1%, so not dissimilar to mine. In fact, your ratio of summer:winter is slightly higher, presumably due to your orientation being east and west rather than south east and south west.

Wouldn't a little dirt make more than 4% difference?

A system we fitted in start of February has been averaging about 18% above PVGIS , which itself gives an output of approx 10% above SAP2009 the preferred MCS estimation systems.
I know we've had a great month for weather but it's was also beating PVGIS estimated outputs before April

I considered it wise to be conservative with estimated outputs
especially with the doubtful figurers being thrown around in green tech advertising ( Solar thermal , heat pump ,etc)

Are we doing PV systems a injustice by under estimating the potential return? or do you think it just been a good spring and very location dependent.

Not actually sure what the OP has in mind here, but there are some tests you can do to see if the output is 'normal'
Work out the variance and see if it is equal (very data collection method dependant), work out the standard error (will show you how far out your means can be without changing anything significant) and do a 2 tailed t-test (checks if there is significant difference between datasets). If everything is between limits then you either know your set up is good or PVGIS is good, or one or the other is bad, or they are both bad (that bit takes more research).

As for East facing, there is some research that there is less pollutants in the air in the mornings, so that can account for a bit of greater output (towards East is best for ST, but that is more to do with usage). Geographic location can make a huge difference because of local weather patterns and finally each individual panel has manufacturing tolerances on it so two 'identical' panels, side by side could show a 20% difference, this is usually masked by the control system/s.

What a lot of thoughts! Thanks to everyone. To address some of the points:
Dirt etc. The SE panels went up in March, so they are clean. The SW ones were snowed on in December, and I cleared the snow using a soft brush on a long handle, so they are clean. Both roofs are at the same angle.
None of the panels are shaded at all - we're 150 yards or more from any neighbours, and no trees are close enough/large enough to be ever be an issue.
The map in Green Building Bible Vol 2, p 16 shows that we are in the 9.5-10 solar irradiation band, so we expect to beat the salesman's predictions (based on about 9?) - which we are. My query related to the difference between SE and SW. Obviously, a sunny morning and a cloudy afternoon (or vice versa) will affect the balance, but over six weeks I reckon that will even itself out.
I do have the percentage figures to one place of decimals, if required, but I reckoned that (a) that wasn't significant and (b) it would be easier to read showing whole numbers.
Both installations have about the same about of clearance between the panels and the roof.
I'll keep monitoring, and see what happens over a full year.
Tony

Posted By: inglesideObviously, a sunny morning and a cloudy afternoon (or vice versa) will affect the balance, but over six weeks I reckon that will even itself out.

Not always, depends if you are coastal, urban or rural. Flatlander or Hillfarmer.
Cloud formation is pretty unpredictable (though you can possible spot the global 5 day pattern). What you will notice is that the variance is smaller in the winter, which can give odd results when comparing monthly to yearly means. Means are not the best gauge really, but hard to use modes as there may not be any. What you have to do is compare it to clear sky data (the theoretical maximum with no cloud) and then model the clouds from your data. Hope you have a powerful PC as your really out the limits of Excel because of the amount of data.

I don't think people realise how much sunshine levels can vary from place to place, not only in terms of averages, but also on any particular day. There are all sorts of wrinkles if you are only measuring over a short period. Here in the SE, December was one of the dullest months ever recorded (10 hrs in the whole month at Gatwick), Jan and Feb were pretty grim too, whilst April has broken all records in the other direction.

Then there are daily variations: in spring you often get E winds which can give cloudy mornings and sunny afternoons, most notably in E half of the country. On the other hand in summer cloud tends to build up during the day, so that mornings can be sunnier than afternoons. Coastal locations are always much sunnier. Moral - your difference from the SW results may be entirely due to the weather, and nothing to do with the equipment.

There are a number of things that change the output, the amount of sunlight, the hours of sunlight, the azimuth, height of the sun, humidity, turbidity, air pressure, cloud type and finally the temperature.
Your 3.6 kWh(p) system has produced around 13 kWh/day
The chart below shows the full spectrum irradiation for the first week of April down here, Monday and Saturday had the best resource with Thursday and Friday being the lowest. The really odd one is Tuesday, it peaks at 600 W.m^-2 but not for very long.
13 kWh/day is about half my usage (all electric here) so not that bad at all.

From a weather station, currently working on energy monitoring but it is getting costly, just need some customers really. Odd that people will put up PV but not monitor the house first, but then I am a reductionist.
If you want to whisper your location I may be able to get some useful data for your region.

Posted By: Jeff BWe have only had our PV panels for about 1 month so I cannot present much meaningful data as yet. We have 20 x 180 watt panels. The roof faces SSE and the angle is about 40 degrees. The best day we had generated 23 kWh and the worst 2 kWh. What has been surprising to me is the large day-to-day variation.

Our WSW facing 40 degree slope 3.9kWp system at 57 degree latitude had a best April day of 25kWh and a worst of 0.4kWh. After you have had the system a while you will likely find that big variations are not uncommon. Back in January the range was 7.2kWh to 0.04kWh.

The total output for April was almost 400 kWh. Does this sound good?

You might like to compare it to the estimates produced by plugging your system details into:
http://re.jrc.ec.europa.eu/pvgis/apps4/pvest.php
&/or you might like to look on
http://www.bdpv.fr/index_en.php
to see if anyone has provided figures for a system anywhere near you for comparison. You might also consider adding your own system to their database.

.

Posted By: skyewright

So, in an ideal world, what data would you like to see, i.e. which values and at what intervals? No promises, but maybe I can provide something better than just the daily summary of PV data?

Just remembered I prepared a report to produce CSV data for a student who asked on the Navitron Forum for data to help test a model he'd produced as a design project. That report produced CSV with 8 columns giving hourly values for:

Year
Month Number
Day Number
Hour Number
PV output in watts
Average irradiance in W/m2
Average ambient temp in C
Average windspeed in m/s

Sample attached for w/c 2011-02-22 (as a .txt as GBF doesn't see to like .csv). If that is of interest I can readily provide equivalent data for a wider period. If you'd like a different combination, let me know.

Posted By: jamesingramJust tohught I'd make people aware of this , if they weren't already

Similarly:
http://www.bdpv.fr/index_en.php

My experience of the weather