best open source

[Vmax]

I also share my experience: I used both openfoam (which I still use) and this turn.
better the first for rans problems, better the second for les.

snappyhexmesh is a meshatore octree that never convinced me, engrid is another story, just started, for now well.

in any case doing cfd with opensource tool is not easy and to date, although openfoam is an excellent solutor, the pre-post tools (especially pre) are not yet in my opinion mature for use on geometries and problems of industrial-grade complexity.

 

[Ing Italy]

thank you vmax...I used in your opinion in addition to the other valid already had; It's been a long time since I read your comments from which I try to learn and read between the lines... I have to say I'm interested a lot, but good weekend

 

[PierArg]

Hi, I'll take advantage of it to ask for some things.
I'm using elmer for a mechanical piece fem simulation. since the piece is not really simple I meshato it via netgen. the problem is I can't export it.
I better explain the steps:
- 3d modeling software is creo
- Export the file in step format
- loading the step in netgen
- I make the mesh
- imposed the format to export to "elmer format"
- click the "export" button netgen crashes

Could you give me directions to solve the problem?

Thank you.

 

[Gionata]

What do you think of quarrel? is free up to 1500 knots, over costs 64 euros.
is a multiphysicist and seems very well done.

 

[PierArg]

What do you think of quarrel? is free up to 1500 knots, over costs 64 euros.
is a multiphysicist and seems very well done.

I've never heard of it.

is this the official website?http://www.lisa-fet.com/while we wait for the "guru" of the fems to give their opinion:

 

[Gionata]

you the program is that of the link.
looks very powerful, works on 2d curves, on 3d surface and solid.
the beauty is that it works on the knots and not on the faces of the surface/solid.
you can create curved/surface base objects within the software.

 

[Matteo]

you the program is that of the link.
looks very powerful, works on 2d curves, on 3d surface and solid.

as I told you in the other discussion, great software, but if the modeling of axiemi/multibody is a limit not recently respect for example to the open source of ef.
... and then lisa is not open source!

the beauty is that it works on the knots and not on the faces of the surface/solid.
you can create curved/surface base objects within the software.

What happens if I have to remeshare? I lose loads and constraints?

 

[Onda]

trying to exemplify:
fea programs are divided into two large families:
the first is that of the programs intimately linked to the programs cad:
pro-mecanic
solid work
etc.
etc.

in these programs the loads are linked to the surfaces, and remain such to the modifications of the part. are very strong in the resulution of contact issues. very useful in mechanical design as modeling and analysis become two intimately linked moments.
they all have a series of simplifications for the load of bolts, for the pins, and for other typical situations in the mechanical construction that make the analysis extremely quick.
Normally (but individual programs test what I say) are software where composite materials are not analyzed and where they are closely linked to cad geometry. with the predictions (easy execution) and its defects. (mainly difficult to simplify complex models, which perhaps require shell and beam surfaces.) lack of composite analysis. restrictions in nonlinear analysis and orthotropic materials.

The other big family is that of stand alone programs, like:
Patran-nastran
classic ansys
abacus
etc.

in these programs it tends to work on knots and surfaces and solids often (not always) serve only to create mesh.
They are extremely powerful, but modeling is much harder and requires more skill and more time.

I use them both:
models of metal parts, with contacts I solve them in solid works simulation.
complex models of composite parts (always shell) solve them with patran nastran.

I believe, but I do not know, that lisa is a program that works on the knots and elements, with all the merits but also all the defects of this system.
It's up to you to see if he does your thing. I don't think he's got a composite module, so seeing what you're trying to do doesn't seem like the best.
I realize that open source or low price composites are not found, so it could probably be a good compromise.
from a quick look, it seems to me that it can not do buckling analysis on shell elements.
Then we should evaluate the validity of the formulation of the elements, but we enter a very complex area, where I can't help you, basic I'm not a fea theorist.

to finish the disquisition, the last trend is to insert stand alone programs, inside very powerful cad, thus extending the potential of the product in an incredible way.
It's the case of abacus inside catia, nastran inside nx and the same ansys workbench, which is not really a cad, but it binds very well to any advanced cad program.
Unfortunately such a system costs such a figure that only a very advanced company can afford.

 

[Gionata]

hi wave we also resent here, from what I could see lisa handles composites.
it is possible to insert different characteristics of the material on the three axes, and even at different angles than x,y and z.

 

[Onda]

hi wave we also resent here, from what I could see lisa handles composites.
it is possible to insert different characteristics of the material on the three axes, and even at different angles than x,y and z.

Bye-bye.
means that it manages orthoper materials. lack of a module that transforms the laminate into the orthotrope 2d (a little what makes laminator).
from what I understand about the program, you have to calculate for each laminate its apparent properties and then insert them, which for a simple piece is feasible, but for a complex model leads to huge times and possibility of very high error.
Besides, he can't post processing ply by ply. means that once you calculate the deformations, nuon you can go to see which ply breaks and in which direction.
a little limiting to do a composite analysis.

 

[Gionata]

If I understood well for composites it is better to work on a shell, that is, a surface without thickness.
at this point from you the mechanical characteristics obtained according to the orientation and thickness of the laminate (with the laminator for example).
at this point, since the fems working on solids are simpler, is it possible to use them however creating a solid with 1mm thick?
otherwise there is an economic or opensource working with surfaces?

 

[Onda]

If I understood well for composites it is better to work on a shell, that is, a surface without thickness.
at this point from you the mechanical characteristics obtained according to the orientation and thickness of the laminate (with the laminator for example).
at this point, since the fems working on solids are simpler, is it possible to use them however creating a solid with 1mm thick?
otherwise there is an economic or opensource working with surfaces?

normally, in a fea program the layers are inserted, not the apparent properties. It is then the task of the program to derive the apparent properties, to conduct the analyses, and, starting from the deformations obtained, to recalculate the stresses for each layer.
You can work with solids and properties, but it's a lot harder.
think of an intersection area of two skins. How are you doing?
open source working with composites I don't know about it.
with surfaces. Mah, I'd say calculix, or salome. I never actually used them. .

 

[Maurowater]

Hi.

It's my first message as you see, I read some information about your forum, and I found it very useful.
in the company where work has arisen the need to model fluidly some machines that we use for reclamation operations, (we operate in the environmental field).
It would be a matter of determining the behaviour of a hypothetical fluid, a kind of pudding (both of large tanks) that is taken from a kind of axial pump.
the robot has already been designed with inventor, is there any tools for inventor making fluid dynamic modeling?
in a company where I worked previously used fluent, otherwise I heard about saturn code also with graphic interface and use not too complicated

Mature