Many interesting things on your graph showing the east/west split. I appreciate it's only a single day and therefore we don't know if it's a typical day or a fluke. For example, the east generation is erratic in the morning when the sun is on those panels, and perhaps is being affected by passing clouds, and smooth in the afternoon when the sun is a best glancing across the panels, or they are lit solely by sky. And the west panels show the same effect in reverse. Also, "noon" isn't exactly 12:00, and anyway as your roof isn't exactly at 90 degrees you don't expect complete symmetry. And on your time line axis possibly it's set itself to BST. Nevertheless, if you look at where the east and west generation lines cross, you could read that as noon being just a couple of minutes after 13:00.