Lowering the LT to 0.47, the note on line 4 is removed.
I assume that even with blinds down there are times of year when the LT 0.6 simulation results in too large an area of the room being above 1000 lux.
Do you happen to know if this is a common result for sDA ASE analysis? I assume that the typical solution is to use a lower LT blind, or is there something else I’ve missed?
Side note - I appreciate with the internal rooms in this model and some larger floor depths that achieving any credits may not be possible no matter the combination of glass and blind LT.
I think it looks reasonable but it is difficult to tell without knowing the model. Those internal rooms are really not in your favour.
It is not always easy to find the right balance. The goal here is to reduce the number of hours where the shading is applied since this will increase the sDA – while still complying with the hourly 2% rule. Here are some of the things that I experienced when working on the recipe.
Decrease the LT of the glazing.
Decrease the shade_transmittance.
Add exterior shading. This is really counterintuitive since adding shading surely will decrease the amount of daylight. While that is true for a static model, exterior shading can help reduce the number of hours where shading is applied, which in return can increase the sDA.
Subdivide the aperture groups, e.g., two groups of four apertures instead of one group of eight apertures. The difference is that in the latter case, everything is shaded when the 2% rule says so, but in the former case maybe one shaded group is enough, while the other is unshaded.
How are the apertures grouped? Do you have one group per room (and two groups in the corner rooms)?
Another key piece of information is the north is roughly at 90deg CW (north arrow pointing right) and the location is London.
The thought behind this was a bit rushed, the model is set up with each room having a single aperture group.
I’ll definitely adjust the set up so that at least for the corner rooms have a minimum of two groups, one per orientation. I can see how multiple groups per orientation could also be a benefit, particularly when the sun is around SE/SW (there are some vertical fins to somewhat benefit).
Radiance parameters wise both simulations were performed on low settings except -ab 5, I’m intending to test at higher settings next week.
I’m still very curious about the times when over 2% of the room area is above 1000 lux. Going purely on anecdotal good practice daylight design I would have expected I should be pushing for an LT of 0.6-0.7. I would have also thought that a 0.05 LT for a blind would be a good balance of transmittance and glare reduction. In IES LM 83 23 the default blind has a 5% transmittance that is fully diffuse (no specular transmission). Is this how the recipe treats the blinds?
Another thing I can think of that should be improved in our model is the lack of border shades / reveals, which I hope will help to limit the direct illuminance to the sensors.
I think that -ab 5 is good enough to show what is possible to achieve with the model. I asked because the default parameters are likely not good enough to increase the daylight autonomy values to a point where you can achieve an sDA value that rewards credits. However, increasing the parameters (-ab) will not influence the direct results, so you can run a faster -ab 1 simulation just to get some quicker feedback on the 2% rule / ASE.
This is not how is it set up in the recipe. The “blinds” are not diffuse since there is no concept of real blinds in the recipe. The shade_transmittance value that you can change is basically multiplied by the illuminance values of the default state. This is also why the users do not have to set up any states for the aperture groups.
If this proves to be a general problem, the approach can be changed. That would require the addition and simulation of another state used for shading. The base recipe, two-phase-daylight-coefficient, is set up to do this. I imagine another approach could be to keep the shade_transmittance input, and then the recipe will modify the model and add the shading state using a generic fully diffuse blind modifier based on the input.
Thanks @mikkel - please correct me if I’m wrong (very much aware I’m a beginner when it comes to Radiance), but would modelling the blinds as diffuse help with the “problem” of the ASE 2% hourly limit being exceeded?
I need to learn more Radiance so I can understand what I should be doing to test these kinds of ideas myself.
Happy to privately share the model if it’s of any use/interest to you.
Hi @mikkel, hope you’re doing well - it’s been a while!
I’m currently reviewing some analysis from the Pollination recipe and the section below of IES LM-83-23 is still bothering me, ie default blinds having 5% diffuse transmittance and no specular component.
Have you thought any more about making this change to the recipe? Personally I think it would be great to have.
I still haven’t done a test to compare, but the few projects we have applied this on, all of them we’ve had to use a lower blind transmittance value to comply with the ASE criteria - and I’m still suspecting the way the blind is applied (ie not fully diffuse) could be due to this.
Thanks for following up. You pushed me to add a command (on my own branch) to automate the modified model with a generic blind modifier so I can easily test it in the recipe.
Just to re-answer this question, if they are modeled as 5% diffuse transmittance, and no specular transmittance, then it will likely help a lot with the 2% “problem”.
EDIT: Moved to a new topic to separate this discussion from the recipe announcement topic.
