Hot days usually have less clouds, have you though about how this is affecting the results?

Looking at the maximum power output per temperature graph I'm quite confident those maximal values are reached in sunny conditions so that pattern shouldn't be affected. It could be that hot, cloudy conditions have a different effect on efficiency than hot sunny conditions but the analysis does not account for that.

I was drawn to this article when searching for reasons for my new solar installation max power variations. I am in England, UK, around 52 latitude. I have noticed that on consecutive days, just after the spring equinox when I expect power levels to be increasing rapidly, that my maximum recorded sustained (5 minutes or more) power at the same time of day (noon, BST, is my array max point) fluctuates. Our weather is variable at this time of year, with some hot, wall-to-wall sunshine, and days of overcast drizzle, and days of sunshine-and-showers. (We love England! 😁). Over several days I noticed peaks of 7.9kW (hot, sunny), then 6.8kW (sun+showers), then another 7.9kW (hot, sunny). The ratio is 86%.
This is consistent with your temperature variation observations, and the theory that on hot days the panels cannot cool, maybe reaching 45 deg C, whereas on showery days the panels are cooled by showers (even hail) and then immediately exposed to peak sunshine. However, over all, the energy generation per day for hot+sunny is much greater than that of sun+showers (66kWh vs 55kWh in my example) because of cloud shading. So, in England (😉) like Belgium, I am not particularly concerned about this.

An interesting experiment to try, would be to spray my panels with water from the hose on a hot day and see what difference it makes.

Thank you for your work, I hope you find my comments helpful.

Dave

Shouldn't the .408% be to the power of 22 instead of multiplied? I know this depends on how the thing is defined and, therefore, this is really a question. If it would be like that then 1.00408^22-1=9.37%. Thanks for the wonderful work here, Mr. Boeye!

The manufacturer claims the trend is linear (probably to keep things simple). I agree there are better functions to describe what is happening.

Really interesting read, thanks for sharing!
I was wondering how do you measure the temperature on your PV installation? Do you have a special temperature sensor installed? I would love to get temperature data from my installation.

Hi tme,

This is a default metric for my PV installation. No special sensor was installed.

Photovoltaic power is a prodcut of photo-current and photo-voltage.

The current in linear proportional to light intensity, and the voltage changes as the log of the photo-current.

The main temperature effect is on voltage, and not current. It would be better to correct your power output by estimated available light, and THEN look at temperature effects. You should see a linear relationship between temperature and 'voltage'

Solar panels do no suddenly get worse above 25 deg C. In some lattitudes, the best days for photovoltaic power output are bright cold winter days.

http://pveducation.org/pvcd...

Hi Chris,
Thanks for clarifying this! What is interesting is that my raw panel dataset also gives me the actual current and voltage per 10 mins so I might play around with those numbers a bit. Not sure if I can get a good estimate on available light though.

Since current is directly proportional to light, you can take advantage of that.

http://pveducation.org/pvcd...

By the way, that is not my website that I'm promoting, just a helpful tool for people interested in learning about photovoltaics.

that link is dead

this one should work