3D multi scale tesselation

[ravikiranbollineni]

Hello Romain,

I am able to create a 2D multi scale tessellation using the code below:

neper -T -n 50::from_morpho -dim 2 -domain "square(0.3,0.3)" -morpho "diameq:1,1-sphericity:lognormal(0.145,0.03)::lamellar(w=0.005)" -reg 1:1 -statface diameq -morphooptistop "nlopt_reps=1e-6" -o n50-id11

The geometry looks fine for my application

But when I am creating 3D model with a thickness of 0.001 using the code below :

neper -T -n 50::from_morpho -dim 3 -domain "cube(0.3,0.3,0.005)" -morpho "diameq:1,1-sphericity:lognormal(0.145,0.03)::lamellar(w=0.005)" -reg 1:1 -statface diameq -morphooptistop "nlopt_reps=1e-6" -o n50-id13

I am not getting exactly the same geometry as above with thickness 0.005.
can you help in this regards?

Thank you,
Ravi.

[dikshanshs]

@ravikiranbollineni : For 3D, seed positions are different in comparison of 2D

[ravikiranbollineni]

Thanks @dikshanshs for answering question. Can you explain the process to get the 3D model that looks similar to the 2D model as above with thickness 0.005.

[dikshanshs]

@ravikiranbollineni you can use seed position and use columnar or try the following command
neper -T -n 50::from_morpho -dim 3 -domain "cube(0.3,0.3,0.005)" -morpho "columnar(z)::lamellar(w=0.003)" -reg 1:1 -morphooptistop "nlopt_reps=1e-6" -o n50-id13

[rquey]

There are several problems in your command:

• the domain is very thin compared to the average in-plane cell size, and so the 1-sphericity distribution cannot be matched (cells tend to have much smaller sphericities than imposed). diameq:1 is imposing equal-size cells, but this would require small sphericities. The structure you get is a (bad) compromise between these contradictory inputs. It is often possible to "detect" bad/contradictory/impossible inputs from the value of the objective function, which should rapidly become small with respect to 1 (as in your 2D case).

• In 2D, the lamella normals are in plane (x-y), while in 3D, they are 3D. This is why the 3D tessellation does not have in-plane lamella normals.

I assume that what you want is an extrusion of the 2D tessellation. In this case, the best probably is to proceed in 2D steps:

1. Generate a 2D structure (1st scale):

neper -T -n 50 -dim 2 -domain "square(0.3,0.3)" -morpho "diameq:1,1-sphericity:lognormal(0.145,0.03)" -o 2d

2. Generate a 3D structure from the 2D structure, and add the 2nd scale:

neper -T -n 50::from_morpho -domain "cube(0.3,0.3,0.005)" -morpho "voronoi::lamellar(w=0.003,v=random(x,y))" -morphooptiini "file(2d.tess)" -o 3d

Visualize:

neper -V 2d.tess -cameracoo 1:1:0.5 -cameralookat 0.15:0.15:0 -cameraangle 20 -cameraprojection perspective -camerasky 0:0:1 -dataedgerad 0.0003 -datacellcol id -print 2d
neper -V 3d.tess -cameracoo 1:1:0.5 -cameralookat 0.15:0.15:0 -cameraangle 20 -cameraprojection perspective -camerasky 0:0:1 -dataedgerad 0.0003 -datacellcol "scaleid(1)" -print 3d

Needs version 4.2.1-21.

[ravikiranbollineni]

Thanks for answering Romain. I am not able to find the Neper 4.2.1-21 version. Could you let me know where I can find 4.2.1 version.
Thanks.

[rquey]

It is the version you can get here: https://github.com/rquey/neper.

[ravikiranbollineni]

Thanks Romain for the help.

[xshang93]

@rquey Came across this post and I'm curious: is there a way to extrude 2d to 3d when doing a 3 scale tessellation? I didn't have any success because the seeds loaded will always be overwritten on the 2nd scale tessellation.