Daylight results different when less spaces
In a "Full Model" case, I modeled 12 daylit spaces with window sizes from 2X3 to 5X6 and window to wall ratios between 10% and 40%. The daylight sensor is behind a partition so doesn't see direct sunlight.All windows are south-facing and "uncontrolled". The "20% WWR 3X4, 3' Ribbon" case is the "Full model" with all other spaces deleted (i.e. everthing is same including furniture for both). For the two spaces, all dimensions are the same and the window areas are very close (108 sf, 120 sf).
The OpenStudio Radiance Measure settings are "Fine" which seems to set some pretty good parameters except ambient resolution is defaulted.
Model results are:
Full Model case daylight sensor mean annual illuminance:
3' Ribbon: 10.4 lux
3X4: 4.6 lux
20% WWR 3X4, 3' Ribbon case daylight sensor mean annual illuminance:
3' Ribbon: 10.8 lux
3X4: 20.6 lux
Other annual daylight sensor results follow the same pattern and Illuminance maps look OK for both simulations and cases.
My question are: why are the 3X4 results different between the two simulations? Is it somehow being shaded in the "Full Model" case? Are there some parameters I can mess with to improve whichever results seem off? Or are the illuminance levels so low that the difference between the two simulations is within the expected simulation error? Is there another reason? Do I have too many surfaces in the "Full Model" case? I expected results more like the "20% WWR 3X4, 3' Ribbon" case because the 3X4 windows are taller and reach over the partitions. Is this an unreasonable expectation?
