SteamyTea

I am hoping that it pushed north, may get a decent August then.

Sunnier is an odd term, especially in the UK. We can mean sunnier compared to yesterday, sunnier because the sun is high in the sky, or not cloudy. When looking at power generation, we could mean just longer hours of daylight, or very powerful daylight. Why we use terms like radiation, irradiance, insolation, but much easier to stick to power and energy, as that is really what we are interested in.

Don't tell Trump, he will blame it all on the Rising Sun.

Right, I stayed up late (for me) and got up as normal, and looked at this problem a bit more. If we are trying to see what sort of difference there is between days and if a sunny day can 'influence' the next day, then the answer down here in Cornwall, is no. Only 21.5% of the time are the conditions matched (by my criteria, the no rain days), which is close to my earlier statement that tomorrows weather will be within 20% of today's. Not really. The average difference in solar power is 5.2%. There are 42.5% days that are sunnier than the day before, and 57.5% when it is not as sunny. I normalised the data to reduce the influence of seasons. Interesting that we ended up in the same place, eventhough we used different methods. I would call that a good result. Because of the rough difference of 40% (more sunny) to 60% (less sunny), this may be why we remember a few good sunny days in a row, they are not normal. The absence of normal. In the past, I looked at the influence of wind direction, I seem to remember that was a better indicator of solar power, and temperature.

Quite fun to look at days with similar hours of daylight. Then look at the temperatures and outputs.

Quite interesting as @ragg987 is pretty central in the England landmass, and has a large plain to the west. I am quite the opposite, surrounded by sea and a large ridge that splits West Cornwall quite nicely between North and South. Perranporth is on the North side. When I get home I shall have a look at the proportion of sunny and rainy days. From that it should be able to get an idea of standard error. If the error is close to the standard deviation, then we can say there is no correlation. I may normalise my data to see if it makes it clearer. Bugger that work gets in the way, it ruins free thinking.

Right, after struggling with a new scaffold tower for 3 hours, and finding out that I do need the extension to get to my top windows, I decided to do a simpler thing and chart the relationship between successive rain free days and solar power. As @Ed Davies mentioned. There is less power in winter than summer, so this possibly skews the data. Normally I would segment the year into weeks to get a clearer picture, but as there is no guarantee that Sunday really is sunny, or Monday for that matter, this cannot be done. So what I have done is to count the incidents that the sun is delivering power into 100W.m-2 bins. So 0 Up to 100, 100 up to 200 etc. Then plot the percentage count for 1, 2 and 3 extra days that are rain free. I have chosen rain as a key indicator as you can have cloud without rain, but you cannot have rain without cloud. The percentage count is of all the valid datapoints between 20/01/2010 and 16/09/2014. So about 4 and half years worth. Basically, my interpretation is that the more rain free days you get (and I get no more than 4 in a row), the more solar power you get. This may not translate directly to solar energy.

Is it going to be Python?

Let's see if I can upload my data. It is a zipped file, so extension needs changing from txt to zip Format is csv. It fats out to about 32mb Solar_Perranporth.txt

I took a slightly different approach. I looked for a rain free day during the hours of daylight, then looked to see if the next two days were rain free (left axis, blue diamond). If they were, I looked at the mean and standard deviation of the solar power, during the hours of daylight, in watts (right axis red square and purple cross) for the full 3 days. There is not a full years worth of data (got to drag some more out), but it shows a rise in mean power as summer approaches. The 'gap' in the data during July and August is not missing data, it is the rainy season down here, better known as school holidays. I have not done a proper correlation yet, truth is, it is late, and I can't think of an easy way to show it. But a quick visual scan shows more power during times of rain free days.

Can't agree with that. The Police should have been prosecuted for negligence and abandonment of duty.

They drive Saabs, if only they knew it is a Vauxhall in a special frock. They get their poloneck jumpers from Steve Jobs' bins, and glasses from SpecSavers. Pencils from Poundland and sketch books from The Works. Then add a few hundred percent markup.

There used to be a saying that the weather tomorrow will be within 20% of the weather today. You can take your pick of any aspect to measure and predict. For a laugh, over at the other place, I predicted the sunlight for a day, a year in advance. I just squeaked in, once standard error was taken into account. My first Masters was meant to be 'The stochastic nature of clouds on PV generation'. As usual for a Master's program, that idea was crushed. May still have the data, so could revisit it.

You only have to draw half of it in CAD, then mirror it. Just make sure you draw the top half, not the left hand side. Rotating it 90° anti-clock, it looks more phallic than gothic.

I went to the same school as Tony Martin. He was considered odd, but he was from Norfolk.

Dudley Moore, Francis Rossi (and his Dad's ice cream), Sarah Pascoe (though a Cornish name), A127 and A13. I still have a soft spot for the place, the mud at Old Leigh.

Just had a quick look at my usage for the 10th Sept. Mean of 243W, total 6.080 kWh It was washing day, so a bit higher than normal. Your 300W is pretty good considering that you run your water systems as well.

Were you charging the car as well? Or was it just a very dull day.

TOWIE

Guilty

An old girlfriend of mine bought a Cornish Unit, got the local builder, who was an expert on them, to do some work. Everything was a surprise to him, he had never seen one like it before. Strange as most were identical copies. So maybe this is something to do with it, hard to get work done on them.

The marginal difference between the cost of lease cars is pretty low though. Not as if these people go out and buy them with earnings they have generated on the last few sales. They may have been getting the blinds from the same company. Reminds of this:

This may be worth a listen to understand things a bit more. https://www.bbc.co.uk/programmes/m00088mc