Deleting volume inside a volume

[ay8tb]

Hi,
I've been experimenting with deleting volumes and have tried to delete internal volumes (i.e. for a volume with thickness 3 mm, I'd like to delete a volume within that from 1 to 2 mm deep). However, when doing that, the code doesn't fall, but instead produces a volume inside a volume, despite its delete command.

For example, here's a line of code

volumes = mapdl.geometry.vnum[0]

CutPart = mapdl.block(5, 10, 42, 44, -1, -2)

Part = mapdl.vsbv(nv1 = volumes, nv2 = CutPart,  keep1='delete', keep2='delete')

Does this mean it's not possible to delete internal volumes? Or is there another command I should be using?

[mikerife]

@ay8tb Can you please post a whole input script so that we can try on our own and get an opportunity to dig in and troubleshoot?
Mike

[ay8tb]

Hi @mikerife,

Here's the whole input script:

from ansys.mapdl.core import launch_mapdl

mapdl = launch_mapdl(run_location=path, additional_switches='-smp')
input = mapdl.input('Sample.anf')

mapdl.prep7()
FullPart = mapdl.geometry.vnum[0]
CutPart = mapdl.block(5, 10, 42, 44, -1, -2)
Part = mapdl.vsbv(nv1=FullPart, nv2=CutPart, keep1='delete', keep2='DELETE ')
CutPart2 = mapdl.block(7, 11, 50, 60, -1, -2)
NewPart = mapdl.geometry.vnum[1]
Part2 = mapdl.vsbv(nv1=NewPart, nv2=CutPart2, keep1='delete', keep2='DELETE ')

And here's the input file
Sample.zip

[mikerife]

Hi @ay8tb The code you posted works as you intend. The result plot should be showing the areas of the voids (with results). If you list the volumes are there more than one?
Mike

[germa89]

We should also consider, that the plot commands based on Pyvista only plot the "outter face" of the volume (internally the issue mapdl.aslv and mapdl.aplot. So they do not plot all the inside elements. Therefor it can look like there is a volume inside a volume, but that is actually the "inside" faces of the inner removed volume.

You might want to check by getting the list of volumes (mapdl.vlist()) or by using mapdl.vplot(vtk=False) to using MAPDL plotting backend.

[germa89]

I'm testing your script...

If you use mapdl.lplot() you can see:

And then you can see that:

>>> print(mapdl.vlist())
LIST ALL SELECTED VOLUMES.

 *** MAPDL - ENGINEERING ANALYSIS SYSTEM  RELEASE                  24.2BETA ***
 Ansys Mechanical Enterprise                       
 00000000  VERSION=LINUX x64     08:23:51  JUN 10, 2024 CP=      4.042

  File: C:\Users\ayehi\OneDrive - University of Virginia\Desktop\ANF files in X



 ** WARNING: PRE-RELEASE VERSION OF MAPDL 24.2BETA
  ANSYS,INC TESTING IS NOT COMPLETE - CHECK RESULTS CAREFULLY **
   NO.SHEL      AREAS               VOLU      #NODES #ELEM   MAT  REAL  TYP ESYS SECN
     2  1    23    25    24    21   8783.          0     0     0     0    0    0    0
             20    22    11    14
              2    13    10    12
              8     7     6     5
              4     3     1
        2    17    18    16    15
             19     9

>>> len(mapdl.geometry.volumes)
1

Then we can be sure that there is just one volume, and the inner volumes are just empty.

[ay8tb]

Hi @mikerife and @germa89,
Thank you so much for your answers. Like @germa89 mentioned, I was more considered about it looking like a volume inside a volume (so nothing was really actually deleted, it was just added). Printing out the list helped.

[Razin-Roslan]

Hello, I am facing an issue with deleting a volume inside a volume too, so I don't think raising a new issue is necessary (do let me know otherwise).

I have generated a body created by extruding an area and would like to cut out a cuboid from it, both geometries can be seen below (the cutout geometry is just some arbitrary cuboid as an example):

However, I get this error:

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
Cell In[60], [line 41]
     [38] main_body = mapdl.vext(main_body, dx=width/1000) 
     [39] cutout = mapdl.block(x1=0.01,x2=0.015,y1=0,y2=0.1,z1=0,z2=0.001)
---> [41] mapdl.vsbv(nv1=main_body, nv2=cutout, keep1='delete', keep2='DELETE ')
     [43] mapdl.vplot(show_lines=True, quality=5, show_bounds=False)
     [45] mapdl.exit()

File c:\Users\Razin\AppData\Local\Programs\Python\Python310\lib\site-packages\ansys\mapdl\core\_commands\preproc\booleans.py:1532, in Booleans.vsbv(self, nv1, nv2, sepo, keep1, keep2, **kwargs)
   [1462] """Subtracts volumes from volumes.
   [1463 
   [1464] APDL Command: VSBV
   (...)
   [1529] NV2.
   [1530] """
   [1531] command = f"VSBV,{nv1},{nv2},{sepo},{keep1},{keep2}"
-> [1532] return self.run(command, **kwargs)

File c:\Users\Razin\AppData\Local\Programs\Python\Python310\lib\site-packages\ansys\mapdl\core\mapdl.py:3206, in _MapdlCore.run(self, command, write_to_log, mute, **kwargs)
   [3204 # check if multiline
   [3205] if "\n" in command or "\r" in command:
-> [3206]     raise ValueError("Use ``input_strings`` for multi-line commands")
   [3208] # Check kwargs
   [3209] verbose = kwargs.pop("verbose", False)

ValueError: Use ``input_strings`` for multi-line commands

I don't see how I might be using mapdl.vsbv any different than how ay8tb has, though I am very new to PyMAPDL/MAPDL so I might be unaware of something I am doing that may be completely wrong. Deleting a volume from a more basic volume works fine for me, it's just deleting it from a slightly more complex volume where it falls apart. Here is the full code:

from ansys.mapdl.core import launch_mapdl

mapdl = launch_mapdl()
mapdl.clear()
mapdl.prep7()

no_chamber = 11

width = 15
height = 19
slit_length = 2
slit_height = 12
slit_thickness = 1
top_thickness = 2
side_thickness = 2
chamber_length = 8
chamber_height = 13
end_thickness = 3
connecting_chamber_height = 2
connecting_chamber_width  = 2

length = end_thickness*2 + chamber_length*no_chamber + slit_length*(no_chamber-1) - slit_thickness*2

mb0 = mapdl.k(x=0, y=0, z=0)
mb1 = mapdl.k(x=0, y=0, z=height/1000)
mb2 = mapdl.k(x=0, y=length/1000, z=height/1000)
mb3 = mapdl.k(x=0, y=length/1000, z=0)
main_body = mapdl.a(mb0,mb1,mb2,mb3) # rectangular outline of part 0.015 * 0.112 * 0.019

for i in range(0,no_chamber-1): # cutting out slits from the part
    s0 = mapdl.k(x=0, y=(i*(chamber_length+slit_length)+end_thickness+chamber_length-slit_thickness)/1000, z=(height-slit_height)/1000)
    s1 = mapdl.k(x=0, y=(i*(chamber_length+slit_length)+end_thickness+chamber_length-slit_thickness)/1000, z=height/1000)
    s2 = mapdl.k(x=0, y=(i*(chamber_length+slit_length)+end_thickness+chamber_length-slit_thickness+slit_length)/1000, z=height/1000)
    s3 = mapdl.k(x=0, y=(i*(chamber_length+slit_length)+end_thickness+chamber_length-slit_thickness+slit_length)/1000, z=(height-slit_height)/1000)
    slit = mapdl.a(s0,s1,s2,s3)
    main_body = mapdl.asba(main_body,slit)

main_body = mapdl.vext(main_body, dx=width/1000) 
cutout = mapdl.block(x1=0.01,x2=0.015,y1=0,y2=0.1,z1=0,z2=0.001)

mapdl.vsbv(nv1=main_body, nv2=cutout, keep1='DELETE', keep2='DELETE')

mapdl.vplot(show_lines=True, quality=5, show_bounds=False)

mapdl.exit()

[germa89]

Hi @Razin-Roslan

The command mapdl.vext does return a string with multiple lines:

>>> main_body
EXTRUDE AREAS      2 TO      2 BY      1
      IN DIRECTION  0.1500000E-01,  0.000000    ,  0.000000    
      SCALED         0.000000    ,  0.000000    ,  0.000000

That is why mapdl.run (used under the hood on mapdl.vsbv) complains about it.

I'm afraid you cannot have:

main_body = mapdl.vext(main_body, dx=width/1000) 

You need to do some logic like:

• Select no volumes (mapdl.vsel("None"))
• Select the area to extrude (mapdl.asel(...)).
• Extrude (mapdl.vext)
• Get the new volume id from mapdl.geometry.vnum.

mapdl.vsel("None")
mapdl.vext(main_body, dx=width/1000) 
assert len(mapdl.geometry.vnum) == 1  # (optional) making sure there is only one volume selected.
main_body = mapdl.geometry.vnum[0] # main_body is now 1.

Regards!

[Razin-Roslan]

Hi @germa89,

Thank you very much for the reply. The changes that you provided worked for the cutout block that I provided, and exporting it as an iges file gives me exactly what I expect, a single solid body.

However, changing the cutout to a block that is completely inside the main body, i.e.
cutout = mapdl.block(x1=0.002,x2=0.013,y1=0.003,y2=0.009,z1=0.004,z2=0.017)
results in an iges file with 2 solid bodies instead of the expected 1. The cutout is its own solid body, and the main body is not cut at all, with both geometries overlapping.

Using geometry.vnum, geometry.volumes and vlist(), I get these outputs. Based on your previous answer to ay8tb, these show that there is only 1 volume, at least to my understanding:

>>> print(mapdl.geometry.vnum)
[3]

>>> print(len(mapdl.geometry.volumes))
1

>>> print(mapdl.vlist())
LIST ALL SELECTED VOLUMES.

 *** MAPDL - ENGINEERING ANALYSIS SYSTEM  RELEASE 2024 R1          24.1     ***
 Ansys Mechanical Enterprise Academic Teaching     
 00608699  VERSION=WINDOWS x64   02:56:14  JUN 13, 2024 CP=      1.875

                                                                               


   NO.SHEL      AREAS               VOLU      #NODES #ELEM   MAT  REAL  TYP ESYS SECN
     3  1     2     1     3     4  0.2746E-04      0     0     0     0    0    0    0
              5     6     7     8
              9    10    11    12
             13    14    15    16
             17    18    19    20
             21    22    23    24
             25    26    27    28
             29    30    31    32
             33    34    35    36
             37    38    39    40
             41    42    43    44
             45    46
        2    50    51    49    48
             52    47

Running the script that ay8tb provided and exporting it as an iges file with mapdl.igesout('C:\\...\\file','iges') also gives a similar issue, where the solid body is deleted if it is touching the "air" AND the main body, and is not deleted if it is fully inside the main body. It also does not actually cut the main body:

Is the exported iges file not representative of what the solid geometries actually are in the script? Or am I using the export command incorrectly?