Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.

importing *.sat file

Please login with a confirmed email address before reporting spam

Hello to everyone,

I am trying to import CAD module *.sat file into the Comsol Multiphysics v3.5a. But while loading the model I get a following warning:
"The input entity-list is either too large or too small. Consider applying appropriate length unit."

As I am new in working with COMSOL Multiphysics, I don't get any idea how to solve this problem.

Please let me know your suggestions,

Thank you very much in advance

7 Replies Last Post Jan 21, 2010, 4:54 a.m. EST
Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Nov 3, 2009, 2:56 a.m. EST
Hi

CAD formats are tricky issues, I can only suggest to get the volume/geometry in another format, try parasolid (.x_t or .b_t) or step; or IGES or DXF if in 2D (the latter should be generated by a recent Autocad software, as earlier DXF as well as many dxf like programmes are rather poor on how the close "polylines")

Good luck
Ivar
Hi CAD formats are tricky issues, I can only suggest to get the volume/geometry in another format, try parasolid (.x_t or .b_t) or step; or IGES or DXF if in 2D (the latter should be generated by a recent Autocad software, as earlier DXF as well as many dxf like programmes are rather poor on how the close "polylines") Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Jan 19, 2010, 7:25 a.m. EST
Hi

Hm.. so you're saing that .dxf I can use only for 2d images???

But I can create 3d .dxf and CADviewers see it as 3d image... but comsol didn't understend it as 3d and somehow converts into 2d...

Do I have any chances to use 3d .dxf for COMSOL?

thx
Hi Hm.. so you're saing that .dxf I can use only for 2d images??? But I can create 3d .dxf and CADviewers see it as 3d image... but comsol didn't understend it as 3d and somehow converts into 2d... Do I have any chances to use 3d .dxf for COMSOL? thx

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Jan 19, 2010, 2:03 p.m. EST
Hi

in the newer version of COMSOL V3.5a yes now you can use 3D dxf and get the 2D, that is already a great improvement from earlier. But no for 3D you cannot use dxf, this is because the dxf format is only spare lines and not higher level 3D cad information such as boxes, spheres etc.

The CAD information in a file is of different levels, and you need to connect lines to build boxes etc (this is the combine and geomanalyze functions of comsol, but based on comsol graphics)

I do not know any CAD software that can transfor 3D DXF to 3D parasolid or another COMSOL compatible format as such, but there might be, as I'm NOT an Autocad user. My preferred CAD software is Solidworks which is based on the "Parasolid" 3D core and therefore is straight compatible with 3D comsol CAD.

Your best chance is too get your 3D dxf from the CAD software used to generate it (which probably has the true 3G geometrical representation inside) and translate it into another higher level format, i.e STEP, IGES, best is Parasolid, and then, depending on the options you have bought on the CAD import utilities, read it into COMSOL.

I believe COSOL has also a some further Autocad bridge, but for that I would suggest that you discuss with your local rep.

Good luck
Ivar
Hi in the newer version of COMSOL V3.5a yes now you can use 3D dxf and get the 2D, that is already a great improvement from earlier. But no for 3D you cannot use dxf, this is because the dxf format is only spare lines and not higher level 3D cad information such as boxes, spheres etc. The CAD information in a file is of different levels, and you need to connect lines to build boxes etc (this is the combine and geomanalyze functions of comsol, but based on comsol graphics) I do not know any CAD software that can transfor 3D DXF to 3D parasolid or another COMSOL compatible format as such, but there might be, as I'm NOT an Autocad user. My preferred CAD software is Solidworks which is based on the "Parasolid" 3D core and therefore is straight compatible with 3D comsol CAD. Your best chance is too get your 3D dxf from the CAD software used to generate it (which probably has the true 3G geometrical representation inside) and translate it into another higher level format, i.e STEP, IGES, best is Parasolid, and then, depending on the options you have bought on the CAD import utilities, read it into COMSOL. I believe COSOL has also a some further Autocad bridge, but for that I would suggest that you discuss with your local rep. Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Jan 20, 2010, 6:01 a.m. EST
Hi Ivar,

So can I be sure that I can build my geometry In SolidWorks?

I have only an implicit equation of 3d pereodic surface:
cos(x)*sin(y)+cos(y)*sin(z)+cos(z)*sin(x)=0

thanks?
Alex
Hi Ivar, So can I be sure that I can build my geometry In SolidWorks? I have only an implicit equation of 3d pereodic surface: cos(x)*sin(y)+cos(y)*sin(z)+cos(z)*sin(x)=0 thanks? Alex

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Jan 20, 2010, 10:00 a.m. EST
Hi again

To try to be more clearer about dxf and other format, as geometrical topology representation is also an important part of COMSOL here is a little more:

The COMSOL geometrical primitives points, edges=lines, boundary=surfaces, sub-domain=volumes are the full 3D representation of any object. Its starts with points and their 3 coordinates in 3D, then the lines which are a staight link between two points, or a higher order Bezier curve between three or more points. Then surfaces that are bounded by a closed line pattern, and finally volumes that are closed by surfaces.

From lines hereon you have additional information needed, the direction of the line (COMSOL s:0..1), and obviously there are two choices, selecting one "positive direction" is by convention. Surfaces have normals and up and down directions new convention, as well as tangeant directions, further conventions, and a surface is defined by a "list" of lines connected together in a "logical" way. Volumes require these lists of surfaces, and in-out normal directions (new convention) for the surface "normal".

Original DXF files just as HGL and other 2D, and later pseudo 3D file formats records only points and there cooridnates and lines as a collection of two points. This is enough to draw a view of the 3D object, but not to represent it and recompose it in 3D. Other formats like IGES, STEP etc added the lists of lines for surfaces, lists of surfces for volumes, higher order lines etc, this helps. Larger CD codes based on ACIS, Parasolid etc store far more data, also object primitives, bolean operations, repetition etc, allowing for a more detailed 3D reconstruction.

In a "plane" 2D is is rather straighforward to rebuild flat surfaces from a series of connected lines (the connectivity of end points is assured by the "repair tolerance" "snap" value (dxf does not ensure connectivity, a problem all CAM users know when milling). And COMSOL is doing this 2D reconstructon OK. But, in 3D its far more complex, and COMSOL, nor any other programme, can fully guess a solution, except for trivial cases.

COMSOL as other FEM programmes has further the difficulties that physical properties are linked to the interfaces, so if the topology changes i.e. volumes/surfaces/lines dissapears or appears, then how to guess the reordering of these boundary conditions ?
There the "human" have to help a little.

Hope this helps on the way
Ivar
Hi again To try to be more clearer about dxf and other format, as geometrical topology representation is also an important part of COMSOL here is a little more: The COMSOL geometrical primitives points, edges=lines, boundary=surfaces, sub-domain=volumes are the full 3D representation of any object. Its starts with points and their 3 coordinates in 3D, then the lines which are a staight link between two points, or a higher order Bezier curve between three or more points. Then surfaces that are bounded by a closed line pattern, and finally volumes that are closed by surfaces. From lines hereon you have additional information needed, the direction of the line (COMSOL s:0..1), and obviously there are two choices, selecting one "positive direction" is by convention. Surfaces have normals and up and down directions new convention, as well as tangeant directions, further conventions, and a surface is defined by a "list" of lines connected together in a "logical" way. Volumes require these lists of surfaces, and in-out normal directions (new convention) for the surface "normal". Original DXF files just as HGL and other 2D, and later pseudo 3D file formats records only points and there cooridnates and lines as a collection of two points. This is enough to draw a view of the 3D object, but not to represent it and recompose it in 3D. Other formats like IGES, STEP etc added the lists of lines for surfaces, lists of surfces for volumes, higher order lines etc, this helps. Larger CD codes based on ACIS, Parasolid etc store far more data, also object primitives, bolean operations, repetition etc, allowing for a more detailed 3D reconstruction. In a "plane" 2D is is rather straighforward to rebuild flat surfaces from a series of connected lines (the connectivity of end points is assured by the "repair tolerance" "snap" value (dxf does not ensure connectivity, a problem all CAM users know when milling). And COMSOL is doing this 2D reconstructon OK. But, in 3D its far more complex, and COMSOL, nor any other programme, can fully guess a solution, except for trivial cases. COMSOL as other FEM programmes has further the difficulties that physical properties are linked to the interfaces, so if the topology changes i.e. volumes/surfaces/lines dissapears or appears, then how to guess the reordering of these boundary conditions ? There the "human" have to help a little. Hope this helps on the way Ivar

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Jan 20, 2010, 10:19 a.m. EST
Hi

Well if you have only a surface, how to get a volume for 3D ?

And I havnt tried to parametrised such surfaces in Solidworks

You need to define the rest of the 3D, if this is a face of a block then I can then propose a trick from COMSOL, if I can remember it: it was documented in one of the docs, or was it one of the courses ?
You can define your function as a forced displacement and have COMSOL to deform a regular block to the desired shape, then save the final meshed volume into a new geometry.

I'll have to search for the example, but probably it is in my office and I'm not back before next week.

Will have to wait for further replies
So long
Ivar
Hi Well if you have only a surface, how to get a volume for 3D ? And I havnt tried to parametrised such surfaces in Solidworks You need to define the rest of the 3D, if this is a face of a block then I can then propose a trick from COMSOL, if I can remember it: it was documented in one of the docs, or was it one of the courses ? You can define your function as a forced displacement and have COMSOL to deform a regular block to the desired shape, then save the final meshed volume into a new geometry. I'll have to search for the example, but probably it is in my office and I'm not back before next week. Will have to wait for further replies So long Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Jan 21, 2010, 4:54 a.m. EST
Hi

Well my surfase is "closed"(except boundaries of course... but it's periodic) so I just need to choose wich side of the surface will be the matter...

It't some tubular periodic surface called Gyroid:
10*(cos(x)*sin(y)+cos(y)*sin(z)+cos(z)*sin(x))-0.5*(cos(2*x)*sin(2*y)+cos(2*y)*sin(2*z)+cos(2*z)*sin(2*x) = 14

And how I can define my function as a dicplacement if equation is implicit? I don't understand...

Thanks,
Alex
Hi Well my surfase is "closed"(except boundaries of course... but it's periodic) so I just need to choose wich side of the surface will be the matter... It't some tubular periodic surface called Gyroid: 10*(cos(x)*sin(y)+cos(y)*sin(z)+cos(z)*sin(x))-0.5*(cos(2*x)*sin(2*y)+cos(2*y)*sin(2*z)+cos(2*z)*sin(2*x) = 14 And how I can define my function as a dicplacement if equation is implicit? I don't understand... Thanks, Alex

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.