grain shape and mesh size

[shuaizhou001]

Hello!
I use this command: neper -T -n 125 -morpho "columnar(x)" to generate the columnar grain model (the following figure). In this model, there exists some small grains and I cannot mesh by large rcl. My purpose is to control 125 columnar grains and 3500-8000 Hex elements. Who has a method to control the random columnar grain (the grain size not have a big difference) and element number?

[smallrain1996]

You could try a strategy like this (with parameters that might still need to be adjusted):

neper -T -n 125 -morpho "size:normal(mean=0.008,sigma=0.002),columnar(x),gg"

For the number of elements, there is no very definite strategy and you need to keep trying

[shuaizhou001]

Thank you for your reply, This is a good method.

[rquey]

neper -T -n 125 -morpho "size:normal(mean=0.008,sigma=0.002),columnar(x),gg"

This can give some improvements, but there are 2 problems with this command:

• size and gg are redundant (gg contains a diameq distribution, which may somewhat conflict with the size distribution)
• gg contains a sphericity distribution, which does not behave well for a columnar structure (the sphericity depends on the length along the columnar direction, which is not something we want...)

[rquey]

Use -morpho to control the grain size distribution, e.g.

neper -T -n 125 -morpho "diameq:lognormal(1,0.1),columnar(x)"

for a narrow lognormal distribution.

I would generally recommend to add a sphericity distribution, but this does not behave well for elongated grains (elongated grains have a low sphericity, which depends on the length along the columnar direction...). To control the sphericity, the best is to generate a 2D tessellation and then extrude it:

neper -T -n 125 -morpho "diameq:lognormal(1,0.1),1-sphericity:lognormal(0.1,0.03)" -dim 2 -o 2d
neper -T -n 125 -morphooptiini "file(2d.tess)" -transform "rotate(0,1,0,90)" -o 3d

[shuaizhou001]

Because if the model exists a small grain and use rcl 1.2, the small grain cannot mesh. The inp file will lost this grain. For this model, rcl1.3 will mesh with a failure.

[rquey]

Il the element size is larger than a grain then, yes, some grains may not be meshed. This will happen less frequently if they are fewer small grains (small standard deviation of size or diameq).

[123fgh123]

Dear rquey:

When I tried the following commands to generate columnar grain with periodicity, I found that 2D gains maintained periodicity very well，but the 3D gains (generated from 2D gains) lost the periodicity.

neper -T -n 125 -morpho "diameq:lognormal(1,0.1),1-sphericity:lognormal(0.1,0.03)" -dim 2 -periodicity all -o 2d

neper -T -n 125 -morphooptiini "file(2d.tess)" -transform "rotate(0,1,0,90)" -o 3d

Is there a method to generate columnar grain and maintained periodicity meanwhile?

Thank you!

Bests,
Guofeng

[rquey]

neper -T -n 125 -morphooptiini "file(2d.tess)" -transform "rotate(0,1,0,90)" -o 3d -periodicity x,y

[123fgh123]

Ok, thank you very much.

Is it possibile to further control the tesrsize of vtk file which was generated from columnar grain?

I want the tesrsize to be 1:128:128, but the obtained result seemed to not be reasonable because the grain mophology was destroyed.

neper -T -n 125 -morphooptiini "file(2d.tess)" -transform "rotate(0,1,0,90)" -o 3d -periodicity x,y -format vtk -tesrsize 1:128:128

Bests,
Guofeng