Free Finite Element Program 3

[steveyeung]

Hi,

Does anybody know a good but free finite element program for

1. Elastic Analysis
2. Creep Analysis

for bridge engineering.

The important point is the program has a strong meshing capability...

 

[johnhors]

Jeff

First, thank you for a very good port of a Linux application into windows, well done!

"port the application from UNIX unchanged" that explains why the input files must be in unix format and not DOS?

I found that the solver is faster than Abaqus but more memory hungry. Also it produces an error when trying to solve models with more than about 60000 nodes, whilst I see from the web site that the Linux version does not have this limitation, is this a limitation of the windows SPOOLES solver? Also the frd results files have slight differences in the format to the specification from Femsys Ltd, which are easily overcome anyway.

Finally imho, it's easily the best freeware FEA available.

 

[jbaylor]

johnhors,

I am glad that the Windows port of CalculiX has been useful. Guido and Klaus deserve the appreciation for making their applications open source.

I thought I had fixed the UNIX-DOS file format problem in CCX. Are you finding this problem in version 1.4? Is this also a problem with CGX? I am currently working on a bug regarding control characters showing up at the end of set names when generated by CGX via picking from the screen. It would not surprise me if there are a few more issues like that.

There was a post to the CalculiX user group on Yahoo (

http://groups.yahoo.com/group/calculix/)

a few weeks ago with a bug report and fix to a memory problem for larger models (the user had a 65k node model). I do not know if that will solve the problem you are referring to, but it may. The fix will probably make it into the next version.

Regards,
Jeff

 

[steveyeung]

Johnhors:

1. I found the file “cygwin1.dll” was in directory “C:\Program Files\CalculiX\ccx_1.4” I don’t know this directory is taken to house the solver because I haven’t gone through all documentation,


2. I have “glut32.dll” in C:\Program Files\CalculiX\cgx_1.4

3. Sorry for misleading……There is no problem with viewing the documentation

However, there is a problem with the installation of ActivePerl…. Since I installed this program on my machine in my office and I was not the “system administrator”, I should have some kinds of problem in the installation. However, there was not error message at all and ActivePerl seemed to have set up on my machine successfully


I have just installed ActivePerl and run calculiX_1_4_win_005.exe? Are there any more programs I have to install or set up I have to do to any *.bat files??

 

[steveyeung]

Jeff,

Do you have any idea why the following message popped up when I run a command “”cgx result.frd”” shown in the manual -- C:\Program Files\CalculiX\cgx_1.4\doc\html\node5.html


Error Message
---------------------------------------------------------------------------------------------------------------
GLUT: Warning in C:\Program Files\CalculiX\cgx_1.4\cgx_1.4.exe: glutReshapeWindow: non-positive width or height not allowed

In addition, my cmdStartup.bat has already been set to:

@echo off

rem Modify this file to change your startup settings for the CalculiX command window.

set HOME=\\SGBCRA1001\home$\chun7981\My Documents\Analysis_Data
set CALCULIX_ROOT=C:\Program Files\CalculiX
set PATH=%PATH%;C:\Program Files\CalculiX\bin;C:\Program Files\CalculiX\ccx_1.4;C:\Program Files\CalculiX\cgx_1.4
rem (the doc paths should be URLs in quotes and have a trailing question mark)
set CCXDOCS="file:%CALCULIX_ROOT%/ccx_1.4/doc/html/ccx.html?"
set CGXDOCS="file:%CALCULIX_ROOT%/cgx_1.4/doc/html/index.html?"
set BROWSER=start c:/progra~1/intern~1/iexplore.exe

cd %HOME%

cls
echo.
echo CalculiX command window
echo usage: ccx jobname
echo (no extension)
echo.
echo usage: cgx [-b^|-g^|-c^|-duns2d^|-duns3d] filename [ccxfile]
echo (extension required)
echo -a auto-mode, geometry file derived from a cad-file must be provided
echo -b build-mode, geometry file must be provided
echo -c read an solver input file (ccx)
echo -duns2d read duns result files (2D)
echo -duns3d read duns result files (3D)
echo -g use element-group-numbers from the result-file
echo -stl read stl triangles
echo -v (default) read a result file in frd-format and optional a solver
echo input file (ccx) which provides the sets and loads used in the
echo calculation.
echo.
echo usage: acis2fbd ^< infile.sat ^> outfile.fbd
echo.
echo usage: vda2fbd infile.vda ^> outfile.fbd
echo.

I also found that in order to run result.frd I have to save result.frd in this directory “” C:\Program Files\CalculiX\common\site””. As you can see above, I have set the path as “”set HOME=\\SGBCRA1001\home$\chun7981\My Documents\Analysis_Data””. In theory, the program is able to read any data saved in …..\Analysis_Data.

 

[johnhors]

Steve

Perl is only required for the installation test, which you get to from the start up model, everything else should run okay.

Jeff

Yes, the unix-dos issue was in the previous version, I hadn't checked the latest delivery for this. And thanks for the link to the memory issue.

 

[jbaylor]

steveyeung,

The environment variable %HOME% is currently only used as the start up location when you open a new command window. When you run the "CalculiX Command" short cut in your start menu, does it start in your analysis data folder? If not, open a DOS shell and type:
cd \\SGBCRA1001\home$\chun7981\My Documents\Analysis_Data
Does this result in changing your current directory to the analysis data folder? If not, try mapping this file system as a network drive.

CGX will only look in the directory in which it is running for files. If you execute "cgx result.frd", you must be in the same folder as "result.frd".

I have not seen the glut error before. What operating system are you using? Does the CGX window still appear and function?

Regards,
Jeff

 

[Steelforbrains]

I don't mean to sound pessimistic but is this Calculix program really the best freeware FEA program available? I haven't tried it yet, but it seems to have a lot of bugs, which scares me if I'm building a huge model. I have spent hours before trying to load FEA software only to find out that the program was definitely not worth the trouble. I have to balance the fact, however, that I'm to cheap to buy a registered FEA program. Could I be wasting my time loading Calculix or is it more reliable than I believe?

 

[jbaylor]

Steelforbrains,

You are right to be pessimistic (I think all engineers should be by nature). Like any other FEA application, there is a learning curve to using CalculiX. Much of what we are talking about here are issues regarding usability that one would encounter when installing and learning to use Abaqus from the command line. There are sure to be bugs as well (the glut error probably is one), just like with any other software application. The benefit that an open source application offers is that the users who find the bugs are able to provide the fixes. You can also browse the source to see the actual formulation for an element, solver, etc.

My time contributed to the CalculiX project eventually shows up on the same balance sheet as my license costs for commercial codes (I just paid my bill for a commercial app). Before I started working on CalculiX, I spent a few months downloaded and evaluating quite a few other applications. CaluliX is the closest to meeting my finite element analysis needs, though it may not be for everyone.

One thing that CalculiX does have, that I think may be unique among open source finite element analysis codes, is a support mechanism. The user group is more responsive than many commercial support systems. If you want to get a feel for the usability, I would check out the archives on the Yahoo group. There are some very serious users. I will admit though that I am not using it exclusively for my analysis work. My customers often expect specific commercial applications, either ANSYS or Abaqus. I also work on a commercial FEA application (not one of the two preceding).

All that said, I guess I need to go back to bugs. Any application (including FEA) will have bugs. I have found them in every commercial FEA app I have used and I find them in CalculiX. Use all of them with care. The thing I like about CalculiX is that I can fix the problems that I find.

Regards,
Jeff

 

[GregLocock]

sfb

try lisa, it's been around forever.

But I bet you'll prefer using calculix. Seems to work just like it says on the box.




Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[steveyeung]

Jeff,

The program runs on Windows 2000

Steve

 

[GregLocock]

Jeff I d/led your install to XP home sp2 and it worked with no fudging. Thanks. Oh, sfb, try Z88 as well.



Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[jbaylor]

Thanks for the feedback Steve and Greg. I hope it will be useful for you. Don't hesitate to email about bugs.

It is tough to tell how much serious use the Windows version of CalculiX is getting, but there have been more than six downloads per day since 1.4 was released. There have been quite a few engineers who have given feedback that suggest that they have been using it successfully. The Linux version has been around for a long time and has had a lot of folks pounding on it with serious problems, while the Windows port has had only a few months now. There are minimal code changes for the Windows port and none in the FEA application CCX.

If it will be useful, I can post some of my notes regarding the other open source applications I have evaluated.

Regards,
Jeff

 

[PhilEduard]

To steelforbrains,

You may try Cadre Pro 5. Somebody has recommended it to me in this forum too and you can use it for 50 runs free or buy it for like less than $400.00 unlimited adn for your solution it seems a good one because it is an FEA program. Try searching with search engine. Good luck.

 

[Steelforbrains]

Thanks Phil. I have used RISA 3D and I like the way it works but you can only save a model with 40 or less nodes. I will try Cadre Pro, are there any limits on model size?

 

[PhilEduard]

I have not build real big models in my present evaluaton (I'm just using a 512 MB Laptop, 3.0 GHZ in my apartment) but I think there is no limit in size. The last thing I realized that did not suit one of my purpose is that it does not support evaluating torsion in shells. Seems excellent and quick doing beam element, even spheres, sheel plate but no twisting (I hope I was mistaken the last I tried it). Good luck.

Regards,

Phil

 

[mrMikee]

Steelforbrains

CADRE has three FEA programs, check here for comparison:

http://cadreanalytic.com/compare.htm

If you don't need plates or dynamics the Lite or evaluation versions should be ok. Also note that the evaluation version (free) doesn't have tension only or compression only members.

None of these programs handle multiple load cases well, however.

Regards,
-Mike

 

[Steelforbrains]

Most of the problems I come across are simple I-beam/angle/channel frame mezanines/support structures. I need a program that I can quickly build a frame and apply my loads.

 

[mrMikee]

I used CADRE for several projects and think it's a good program for the money ($142). Since I typically have to design for 20 or more load cases I had to find a program better suited for running multiple combinations.

-Mike

 

[bkal]

I also used Calculix - but was double careful to make as many sanity checks as possible to convince myself that results were ok. What was good about it was that it almost seamlessly integrated with FEMGV pre/post-processor (while I had a licence). I have a number of FE programs which I developed over time for my own purposes, but I would like to get hold of a free / cheap pre and postprocessor, preferably for Windows. Any suggestions?

 

[johnhors]

Bkal

Look at :-

GiD -

http://gid.cimne.upc.es/index.html

and

Roshaz -

http://www.roshaz.com

(has seamless integration to Calculix)

The Calculix pre and post is closely based on the command line language of FEMGEN and FEMVIEW when they worked together, they then developed (seperately after ending their collaboration) into Cadfix and FEMGV respectively.