CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   Main CFD Forum (http://www.cfd-online.com/Forums/main/)
-   -   IGES to Gambit Problem (http://www.cfd-online.com/Forums/main/738-iges-gambit-problem.html)

A. Al-Witry April 19, 1999 20:23

IGES to Gambit Problem
 
Hello everyone

I am currently trying to model a heat exchanger plate geometry with 15 rows X 65 columns of tube banks in Fluent5 in a two dimensional manner. The outline of the plate is zig-zagged and is not single continuous wireframe but a very long zig-zagged line surrounding the tube banks.

The Problem is: when importing the geometry from IGES (from AutoCAD) into Gambit (and after deleting all un-used points+uniting all duplicates) the external wireframe does not unite into a single face (which is a necessary step if I want to deduct all the internal tube bank volumes from the flow path). Gambit gives the message that it can not connect real with virtual edges!!!.

Any clues about this one?!. Kind regards.

Stephane Baralon April 20, 1999 13:06

Re: IGES to Gambit Problem
 
Real geometries in Gambit are ACIS kernel based geometry. I know that the recent versions of autocad including solid modelling are also based on the same kernel.

A solution to your problem may be to read your autocad IGES file into a more recent release of autocad (which, I hope, is backward compatible) and be sure to save it as an ACIS file.

Then, during the import procedure Gambit should not create any virtual geometry. it should recognize the ACIS kernel.

However, this is hypothetical.

I hope I have been of help.


Alberto Tamm April 20, 1999 13:08

Re: IGES to Gambit Problem
 
Hi,

First sorry for my bad english.

I had the same problem some time ago, and finally I took only the important points in a different layer in Autocad (that defines the geometry) and export it to Gambit.

The problem with Autocad is that the faces arenīt connected and the defining point arenīt the same for that faces. So itīs very difficult to became conected faces. You can delete one face between two faces and then construct a new, but be carefull with the points.

I recomend you tho export only the geometry defining points and construct the geometry in Gambit.

If you build the geometry in Gambit begin with Faces and use boolean tools for generating the complete model, so you will be sure that the vertex,edges and faces are connected. Try to use only real vertex, edges,faces or volumens, so you can always use booleans tools.

If you can construct the geometry in Pro_E or in Microstation you donīt have this problem.

I hope it will help you.

Alberto Tamm TU-Darmstadt

Scott Gilmore May 6, 1999 22:06

Re: IGES to Gambit Problem
 
Stephane is correct that Acis .sat is the best way to move geometry from AutoCAD to Gambit.

Also note that the next release of Gambit (1.1, due later this month) will contain a new IGES reader that always produces all-real geometry. It will also contain a "healer" that can repair some types of problems in imported geometry.

Regards, Scott

Jonas Larsson May 19, 1999 07:18

Re: IGES to Gambit Problem
 
Which is the best way to import CADDS5 geometries into Gambit?

If you use IGES gambit 1.0.4 frequently fails to import all geometry (some surfaces missing etc.). Also, you get a lot of virtual geometry which is a real pain to work with. For the moment we are often forced to work with ddn/pcube instead. What improvements will the new IGES reader have? How will it handle discontinious surfaces, trimmed surfaces etc? Is there a date set for the release of 1.1?

Scott Gilmore May 27, 1999 15:14

Re: IGES to Gambit Problem
 
Jonas Larsson writes:
> Which is the best way to import CADDS5 geometries into
> Gambit?

If your model is a solid within CADDS5, export it as an IGES Manifold Solid B-rep Object (MSBO) and import it into Gambit 1.1. It should come into Gambit as a solid volume. If you have any problems, send the IGES file to your local Fluent office and we'll take a look.

Another alternative, albeit a costly one, is to use Theorem Solutions' CADDS->ACIS translator ("CADVERTER"). Go to http://www.theorem.co.uk for more info.

> If you use IGES gambit 1.0.4 frequently fails to import
> all geometry (some surfaces missing etc.). Also, you get a
> lot of virtual geometry which is a real pain to work with.
> For the moment we are often forced to work with ddn/pcube
> instead. What improvements will the new IGES reader have?
> How will it handle discontinious surfaces, trimmed
> surfaces etc?

The new IGES reader always creates "real" (ACIS) geometry, supports IGES MSBO solids, and has better support for trimmed surfaces. We have seen excellent results for numerous IGES models. The new "healer" can, in many cases, automatically stitch real ACIS volumes after importing IGES surfaces with the new translator. Results are best when the model was a solid in the original CAD system.

> Is there a date set for the release of 1.1?

Gambit 1.1.0 was released last Friday. Please contact your local Fluent office for downloading instructions or to request a CD. (Download is available already; CD's next week or the week after.)

Regards,

Scott Gilmore Fluent Inc.

Sangaria June 29, 1999 22:41

How to Convert....?
 
Hi

I want to know how to convert cad geometries to gambit format.

If I have only wireframe cad geometries ,not solid model, how to use them in gambit?

if only solid model available to gambit, how to express complex geometries?

please, send me your help in my e-mail.... thank you

Scott Gilmore June 30, 1999 07:15

Re: How to Convert....?
 
Sangaria,

You can import your wireframe geometries in IGES format, then form wireframe faces from your edges, and/or sweep and revolve edges to create faces. You can then stitch a volume from the faces. If your model contains only planar faces, you can directly stitch a volume from the wireframe.

I suggest that you contact your local Fluent office for assistance. They can look at your wireframe model and suggest the best approach.

Scott


All times are GMT -4. The time now is 21:06.