I'm using the detailed dynamic shading control to compare the shading behaviour according to different orientations. My reference building has two windowed facades, in one case South and North oriented (case A), in the other East and West oriented (case B). In both cases I set up an internal sensor located 2 m from the window which is regulated through an automated thermal control with a 500-2000 lux range. In order to reduce the computational time I analyzed only the month of January.
Results from case A:
SHADING ON SOUTH
SHADING ON NORTH
The north-oriented shadings go down essentially due to the solar radiation coming from the window located on the opposite side.
However, I've found a problem when I consider the shadings located on the East and West oriented facades .
Results for case B:
SHADING ON EAST
SHADING ON WEST
It seems like the East oriented shadings schedule contains both the results for East and West.
I repeated the simulation also using different climatic files, but the results remain the same.
Thanks a lot
First, only looking at the month of January will not speed up your calculations at all. DIVA/Daysim will still calculate the total annual illuminance calculation, so feel free to apply an occupancy schedule which takes advantage of that.
It is somewhat surprising that your West shades don't get triggered. Could you post your model?
I'm surprised too! Here the model.
Thank you so much,
This looks like a limitation in Daysim to me. http://daysim.ning.com/page/daysim-header-file-keyword-advanced-dyn... suggests that Daysim does not ever calculate both shading states closed at once. "In case the blind 1 is partly closed and blind 2 is fully closed DAYSIM is generating the additional DC file on the fly without a further ray-tracing run by consecutively subtracting the effect of blind 1 and 2 compared to the base case. This will yield accurate results in case the two shading systems behave 'additive', i.e.e they either do not overlap." Your model probably does not behave well with regards to this assumption.
Deleting the East shading state does allow the West shades to be lowered, which kind of supports this thought.
honestly I'm not totally sure I understood. Why do you think that my model does not behave well with regards to the Daysim assumption? May be because the solar radiation entering e.g. from West has an influence on the sensor located nearest to the east-oriented shading? Do you think that for Daysim could be a problem manage the incoming solar radiation from west also to control the east-oriented shades?
What appears strange to me is that this problem doesn't seem to concern the case with the north and south-oriented shades.
I noticed that if I run a simulation with only north-oriented shades they remain always open because the illumination on the sensor point doesn't overcome 2000 lux. But, if I consider the shades located on the south-oriented facade too (see my first post) the north-oriented shades go down since in this case the illumination on the sensor point overcomes the treshold value.
May be there's something that escapes me.
Apologies for the late reply. What seems to be happening is that your East shading state is still closed during the hours the West shading state is triggered. This is likely due to the threshold needed to return to the previous shading state. Daysim doesn't calculate all the combinations of shading possible -- it only makes an estimate of what the illuminance should be with only one shading state closed. Unfortunately this estimate seems to remove too much light from the Western control sensors and keeps your West shades from triggering due to the assumption Daysim makes internally here. To be fully rigorous, Daysim would need to test every possible combination of shading states, but it does not do so.
This isn't your fault or your model's fault at all, but rather is tough simulation assumption / simplification in the software. If you look at the two controls independently (turn off east shading controls and only look at west) and vice-versa, you can get the proper control schedules but the combined effects of illuminance in the space become lost.