I vaguely recall someone recently asking about how big an increase in output you can get from "edge effect". Today during yet another mostly cloudy (but gradually clearing
) morning there was a small break in the clouds and I recorded a nice short-lived increase in ouput, 2630W from a rated 2190W. That is 1.2 times the rated output of the panels. The Outback FM-80 MPPT charging regulator is supposedly limited to 80A, but it is slow to clip to this level, and short overcurrent peaks do regularly occur. I once saw 100A out of it at ~25.5V.
The air temperature was 8.5C, and I measured the cell temps to be ~21C shortly afterwards. Larger effects would be expected at colder temps. The peak in the solar radiation at 10:33:30am was a bit too brief to show on the Pyranometer recording, which integrates over about 1 minute, but there was a peak of 896W/m^2 (on a flat surface) at 9:50am, which is ~1600W/m^2 on my tracked panels facing the sun (34deg above the horizon). The array showed a few peaks around 2300W then.
Cloud edge effect is caused by reflection of light from clouds and forward scattering of light within the cloud, which adds to the radiation coming directly from the sun. It is often said to be caused by refraction, but IMO this is incorrect. There is no giant lens in the sky with a focal lenth of a kilometre or two to concentrate the light on your panels! Certainly water droplets are always refracting incident light, but with a focal length better measured in microns than kilometres, that light is dispersed up at cloud level. Some hits other droplets and is reflected, some is refracted and further dispered, well up in the atmosphere.