Ask Your Question
1

Radiance measure fails at end of simulation

I have a very large Radiance simulation (19k calculation points) that I'm running via OpenStudio V2.3.0. I have modified the measure script to allow for more points, but the simulation fails towards the end of the run. I have the stdout file attached with the specific error language but it appears I'm getting an error occurring during the creation of the radout.json file. Many of the Radiance output files have been created by the time it fails but I'm still missing the zone .ill files and the sql file (which are primarily what I need unfortunately). Does anyone have any experience or insight into this issue? Thanks.

stdout file link

FYR, I have another unresolved issue that may be related in someway but I'm not sure: Illuminance Map coordinates and lengths shift after reloading

jt1234's avatar
331
jt1234
asked 2017-10-19 18:55:55 -0500
edit flag offensive 0 remove flag close merge delete

Comments

I'm now thinking this may be a size issue. I ran a lesser quality run and the simulation gets further along (creates most zone illuminance maps) but still fails with the same error. The illuminance maps and zone reports still aren't produced in the output file (no ill or sql files). I'm not fully informed on the inner workings of Radiance, but would this make sense or is there another underlying issue?

jt1234's avatar jt1234 (2017-10-20 06:34:15 -0500) edit
more
  • delete
add a comment see more comments

1 Answer

1

This is basically why we added the failsafe. You are likely hitting memory or disk space errors ahead of -- or during -- the final merge to the .ill file(s). Can you dd a link to a failed run directory, so I could see all the files being created and the output?

rpg777's avatar
7k
rpg777
answered 2017-10-20 10:30:53 -0500
edit flag offensive 0 remove flag delete link
more

Comments

The partial run file directory is >12GB so I don't think a transfer is an option. However, I think you're absolutely correct and that I just hit the limit of the software and my system because I just successfully ran a 'dumbed-down' version of my model with less zones, fewer calc points, and lower quality. Thanks for the help. Much appreciated.

jt1234's avatar jt1234 (2017-10-20 11:07:44 -0500) edit
more
  • delete

Yeah, this is why most folks doing typical daylighting analysis with Radiance will look at specific spaces or collections of interrelated spaces, and apply the findings to similar spaces for the purposes of design and compliance demonstration. There are a few things that could be added to OpenStudio that would support this workflow, and still more optimizations that could be made to the Radiance Measure -- were there funding to do so. IOW, OpenStudio is targeted at whole building models but the Radiance measure has trouble with this currently.

rpg777's avatar rpg777 (2017-10-20 13:43:12 -0500) edit
more
  • delete
add a comment see more comments