The behaviour of the PV system

Some interesting and curious phenomena are coming to light (if you are a PV geek).

Yesterday was the first time in a week that we didn't get close to 8 kWh. It was pretty cloudy all day, with occasional light drizzle. However, we still managed to generate 4.7 kWh - which is as near as anything to what we hope to see as an average for the year.

Phenomenon 1: When it's cloudy, the PV system comes on earlier.

On sunny days, it can be quite late by the time the inverter switches on - even now, around 9.00 (which, with daylight saving, is close to 0800 in "solar time", where you base noon on the time the sun is closest to overhead). It seems to be the case that if it's cloudy (but reasonably light) it comes on earlier - today it was on a good hour earlier than that. I think this is because one of the effects of the clouds is to diffuse the sunlight across the sky. As long as the cloud cover is not too thick, the brightness of a cloudy sky overall tends to be greater than the brightness of a blue sky.

Obviously, without direct sunlight, the panels won't generate as much power as in sunshine. However, this brings me to ...

Phenomenon 2: The highest output peaks come not on the sunniest days, but on sunny intervals on cloudy days.

In theory, the output of our inverter should not exceed 2000 W. The running output we have been led to expect is around 1800 W. Look at the peak output from the last few days:

Date	Peak output/W	Total output/kWh
06 Apr	1741	8957
07 Apr	1609	8750
08 Apr	1584	8454
...
12 Apr	1992	7912
13 Apr	1930	4701

6-8 April saw us in settled anticyclonic weather. I suspect that the reason for the gradual decline in peak output (and total output) was that this sort of high pressure results in a haze layer gradually building from ground level up. Barely perceptible, particularly to begin with, it does however lead to the rise in hayfever conditions at this time of year! However, it may be thick enough to gradually result in more of the sunlight not arriving at the solar panels.

12 and 13 were both cloudy; 12 intermittently cloudy, and 13 as I said earlier on, fairly steady grey. However, on both days, peak output was higher than in the anticyclonic weather. There are a couple of possibilities for this. The first is that it was no more than a momentary spike, and it wouldn't have maintained it for any length of time. The second is that the haze layer is significantly less defined outside the anticyclonic conditions, and when the sun was able to break through the cloud, the solar system was able to generate a higher power.

This requires a little more research, which will be easier when the promised PC interface card arrives - we can see just how much time is spent at high power on the less sunny days. It is certainly satisfying to see the LED on the solar output meter blinking away every two seconds as another Wh is generated in bright weather.