I have been running on an urban context. I have successfully generated a number of results on this scale. However, when I start to run the analysis on a number of the boundaries ( i have to extrude them to run as I ran into another issue where I cannot run them as surfaces or meshes)…
The latest issues is that a number of the outlines cause the daylight analysis to throw and error saying: “solution exception: index out of range: 0”
It does happen for every curve ( have been able to isolate at least one group that does not cause the error). All curves are close, periodic and planar. Im working in meters and my min grid size is 10 meters (i know this is large but when you are working on a city it is not very big)…
I have updated to the latest version so i know it is not a version issue. I’ve also tried reducing the grid size to 1 m.
I’ve internalized a sample that does not work and a sample that does work. does anyone see anything i’ve missed?
I’ve also reviewed similar posts but none of them seem to help:
http://www.grasshopper3d.com/group/ladybug/forum/topics/questions-f…
http://www.grasshopper3d.com/group/ladybug/forum/topics/daylight-an…
The GH file with the internalized data is too large to upload here (88mb). Is i’ve added it with a box link below:
https://perkinswill.box.com/s/pf3yodsmf51i20xeb2t5trmwuzlm8di1
This second gh file has only limited context internalized (still 10mb so can’t attache directly)
https://perkinswill.box.com/s/lfezlb1pmiqqg97n3dz1u81xd0qtnkaj
Hi Jeremy,
I have checked your file and it should work fine.
Firstly, I suggest that you update your LB version, the last one is 0.0.64. (Feb).
There are two things you should to do when you use LB sunlightHours:
1) check the direction of analysis surfaces, sometimes it doesn’t work because they are flipped.
2) when the model is huge like this one, try to simplify it. If it is not possible, try with a part of it and when you are sure and all it is fine, run the complete model. Moreover, for the first tests of your model I suggest that you use a big gridSize, e.g. 10 meters.
Let me know if it works
LB_sunLightHours_DNA.gh (435 KB)
update: the project was based on gis data which was too far off origin. moving the base data to the origin seems to be producing better results. more to follow…
if not then there is possibly a limit to how much can be processed: i am currently running approximately 100K+ context meshes on either a large base surface (which worked) or around 400 smaller parcels (which is not working)…
Thanks Antonello. I didn’t know there was an update. I’ll update my graph. i did check the direction of the surface. which is what baffled my as to why this didn’t work. i had successfully tested with a test file but did not test with only a portion of the model (this is really good advice)…
update: moving the geometry closer to the file origin resolved the issue
Hi Jeremy! Thank you for posting the issue and also the update. Maybe we should calculate the boundingbox and give a warning for cases far from the origin.
that might actually be a good idea… this seems to be the bane of many urban projects…
