I am simulating the daylight autonomy for the internal wall surfaces in a few rooms. When looking at the local illuminance values over the year, I have one point which gets much higher illuminances than the surrounding points. This happens on several days in summer, typically around 10 am: 5'000-15'000 lux compared to the surrounding points at about 300 lux. I tried running the simulations with 4 and 6 atmospherical bounces, but get the same pattern.
For me, the most logical reason would be that I have somewhere a gap in the ceiling/roof, though I cannot find any.
Any tips on what else to check or try?
I am assuming that it is not a bug, as the pattern is too consistent when simulating with slightly different parameters, with and without dynamical shading etc.
Its a little tough to say without seeing more of the model details. Can you share your files with us, either on the forum or via firstname.lastname@example.org?
Thanks for your reply.
It is a room with 3 rows of 7 roof windows. I have narrowed it down to the daylight passing through one row of 7 windows with NE-orientation. Each of these windows causes one or two "points" on the east-facing wall where the local illuminance peaks.
- The peaks occur 63 to 72 days per year.
- Each peak occurs always at the same time of day, either hour 9.5 or 10.5.
- The peak internal illuminances are at 20-45 % of the exterior horizontal illuminance. Max. direct transmission allowed by geometry (shades+glazing) is 52 %.
- The timing of the peaks corresponds with the orientation of the windows and the solar position.
I think the logical conclusion is that we have direct sunlight passing through the windows at specific solar positions, hitting these points. I am running an extra simulation with a finer grid (0.1 m instead of 0.4 m) in this region to check.
We are still not entirely sure what is happening here. I would like to send you the files, with the problem narrowed down to one small nodegroup. The model is quite large, I hope it is ok for you to take a look?