KaiKostack / bullet-constraints-builder

Add-on for Blender to connect rigid bodies via constraints in a physical plausible way. (You only need the ZIP file for installation in Blender. Click the filename and at the next page "Download", right click and "Save As" won't work!)
https://inachuslaurea.wordpress.com
GNU General Public License v2.0
135 stars 21 forks source link

Export displacement time history #14

Open HollowHeartNet opened 2 years ago

HollowHeartNet commented 2 years ago

Dear Kai,

is there a way to increase the time increment for the export of the displacement time history? Regardless of the steps per second I choose, the time increment for the displacement time history is always 0.04 s in the exported csv file which is in my case too inaccurate. Thanks in advance.

KaiKostack commented 2 years ago

Hi, there are two ways to do this. You can either use the time remapping functionality of Blender to scale time to a different amount of frames. That way values for additional frames between original frames are interpolated, but this only works for displacements not for forces. Do this change after simulation and before export. Time values in the output file might need correction but I haven't tested this.

blender-time-remapping

The second method is to modify the speed (time scale) setting before simulation because then the BCB will simulate with more time steps per second. If you are interested in more accurate simulation data then this is the preferable way.

blender-time-scale

HollowHeartNet commented 2 years ago

Thank you very much for the explanation, Kai. I have noticed that changing the speed parameter e.g. to 0.1 leads to a very inaccurate structural behavior under a weak seismic excitation. The expected response under the input ground motion must be within the linear range (I have checked it using FEM). I obtain linear structural response using the default speed (1.0). When changing the speed to 0.1, however, the structure collapses during the simulation. As I said, this should not happen. Maybe I am doing something wrong but that is my observation so far.

KaiKostack commented 2 years ago

While I have tried to make time scale work as good as possible it is not well tested. There can be different reasons for this extreme change in behavior. For instance the constraint solver iterations are defined per simulation step, not per second like the steps value. So you can try to lower this setting by the same factor as the time scale to make the solver settings consistent with default speed. I hope this helps. Let us know if you find a solution.