Hello!
I’m doing a grid based illuminance analysis with HB+ and Radiance.
To generate the analysis grid, I’ve modeled a surface on Rhino, and then used it as an input in the LB_Generate Mesh Component.
I’ve found that the direction of the face normals directly affects illuminance values, as you can see in the figure below.
When normals are facing upwards: Illuminance values are higher, though the pixels are really jagged.
When normals are facing downwards: Illuminance values are lower, but the final image is way smoother.
I’ve used the same conditions of sky, date, hour and radiance parameters to test both conditions, so I believe this inconsistency is really related to the direction of the face normals.
What is the correct direction for these face normals, when performing a grid based simulation on HB+?
Upwards, because that is how we measure illuminance in the real world. Grid-based simulations are essentially a proxy for calculating work-plane illuminance. And the idea there really is to calculate the amount of light falling on the workplane.
If you turn the grid downwards, you are measuring reflected radiation from the floor.
Light reflected from a diffused surface with low reflectance (floor) will be a lot smoother and uniform than that incident from a direct source (sun and sky). When facing upwards your grid points “see” the ceiling , sun and sky so the gradient and variation in values is likely to be higher.
Interesting! I also just experienced the same thing and conclude that the surface direction/ surface normal influence the calculation results. I just attempted to verify daylight simulation in Radiance with measured data (one single room) and I found that the simulation results matches the measurement if all surface normal faces interior.
Previously I found significant gap (illuminance) between simulation and measurement. What I did is: (1) check each surface direction of the building model in rhino using “dir” command and (2) change the surface direction using 'flip" command, then (3) referenced these surfaces again in Grasshopper for defining HB surfaces.
Hi @sunarywend !
You information was very helpful!
I’ve used a similar workflow, but for the analysis surface.
The G input in the flip component allows you to insert a guide surface to match.
So: 1) I first created this guide surface, with the normals facing upwards, 2) I internalized this srf in a geometry component, and 3) connected to G
This will make sure that any analysis srf, with the wrong normals direction (downwards) will be flipped, and the ones that are correct (upwards), will remain correct.
Though I don’t think my workflow will work with vertical surfaces (e.g. walls), because each srf will need a different normal direction.
Hi, I’ve just updated my Ladybug to 1.1.0. and cannot find a way to map the illuminance apparently there is only the % of the time that daylight passed the threshold. I was wondering how you could make your model with illuminance values?
Tnx
