Euler-lagrangian-coupling 1

[Kailar]

I am modelling a wedge shape in an eulerian mesh.
Next, i have to define some contact procedures. Common is used 'the general contact'. In papers, they refer to the penalty contact so i have defined some parameters but i alywas have the same error message in the monitor 'Abaqus/Explicit Packager exited with an error - Please see the status file for possible error messages if the file exists'

Im working in abaqus explicit so the defined parameters are scarce. I can only take the default option.
Next I take the scale factor for the pressure-overclosure option. What do i have to fill in for the subjoined parameters? Maybe that's the cause of my problem...?

 

[Noirom84]

What kind of information do you get form sta. file?

 

[guru64]

As Noirom has mentioned, without looking at the sta file, you cannot make any assumptions about the cause of problem.

And you dont need to define any parameters for penalty for general contact. By default it takes some parameters. Only if you want to define non-default parameters you have to write them down explicitly.

Do let us know what was in the sta file.

Guru

http://www.abaqusguru.blogspot.com

 

[Kailar]

The sta file is empty :s :s

 

[guru64]

Alrite!! I overlooked something. The Explicit Packager itself is giving problems. So the solver itself hasnt started yet. It is unlikely that the sta file exists.

Check for error in dat file.

Guru

http://www.abaqusguru.blogspot.com

 

[Kailar]

I have attached the dat-file. please take a look.
I only have some warnings. I just have one error I described above..

 

[Kailar]

Maybe it is usefull to give some more information so you can better answer my question.

The aim of my work is to let fall a wegde in a calm watersurface. the wedge has an initial velocity that I have defined with the help of a predefined field. But you have to know that we use general contact for euler-lagrangian coupling. But I think I am stuck in this part.
Which parameters do I need to describe the general contact?
Stiffness parameters between the master en slave nodes??

 

[Noirom84]

Hmmm....I did such type of analysis some time ago. What can I say, it won't be easy :-D

As I remember in contact parameters you must define general contatt. Set contact properties as default. Can you attach input? This may be helpful. Try to set also hourglass at "enchanced" or turn off reduced integration. Most difficulty in such analysis is to achive convergent solution. Also beware of "extremly disorted elements".

Typically that kind of analysis is good example of co-simulation (Fluent+MPCCI+Abaqus).

Regards

 

[guru64]

yes..This FSI stuff is difficult.
If possible can the input file be uploaded? I understand that the file can be very big as the number of variables is around 2M. Also can you specify the version of abaqus you are using, just in case you manage to upload the file, so that we can test it?

Guru

http://www.abaqusguru.blogspot.com

 

[Kailar]

I am using Abaqus 6.8.2. It's the first version which is capable to simulate the slamming with an euler-lagrangian coupling.

Do you mean the cae file? its about 17MB. I've tried to upload the file but i didn't succeed.

Maybe there is another way?

 

[guru64]

6.8-2 definetly has very impressive CEL capabilities. You can use the following site to upload the CAE file.

http://www.yousendit.com

Once you have uploaded, you will get a link the file, and post the link back to this thread. Probably tomorrow we can find a solution.

Regards
Guru

http://www.abaqusguru.blogspot.com

 

[Kailar]

Here is the link where you can download the file.

http://www.yousendit.com/download/Y2ovV0oxaTFRWUt4dnc9PQ

I really appreciate your help.

I have attached a paper that refers to my work. Maybe it helps you a bit more..I have read it 10 times but it cannot help me either.:S Especially the last pages are important.

Thanks anyway for your help!

 

[guru64]

Hi Kailar
Thanks for the file. I just checked the CAE file.
Nothing is really wrong with the Model. I ran the model with almost 7G of memory and it ran without hiccups ( I manually stopped the calculation after some increments ). The status file showed that it used a bit more than 1G of memory for the simulation. I saw in your CAE that you have specified 50% of the memory for your simulation. Probaly that is the bottleneck. Try changing it, and set about 1.5G or more, and see if it runs.

Good luck!!

Regards
Guru

http://www.abaqusguru.blogspot.com

 

[Kailar]

hmm I have changed the value in the abaqus.ENV file to 100% but im still not surviving the analysis..Still the same error.

Maybe i have to change the memory usage in another file?

 

[Kailar]

Hm I forgot to say that i am using a laptop of 4 giga of ram so normaly it has to work I think...

 

[guru64]

hmm...

try setting in the memory as 90% of physical memory inJob Editor.

But now im not convinced that this is a memory problem, as you have 4GB. Probably this is a OS related problem. I ran it on a linux box. I havent tested it on windows machine. Tomorow i will try and will let u know.

Can you tell me ur machine config?
Windows or Linux?
32 or 64 bit? etc?

Regards
Guru

http://www.abaqusguru.blogspot.com

 

[Kailar]

The file is a bit long to put it in a message.

See the attachment..

thanks!!

 

[Kailar]

Im sorry , i did send a dutch one here you have the english one!

Im working on a 32 bit system.

regards

 

[guru64]

It seems like a OS related problem rather than a memory related one.

The Model runs in my linux machine seamlesly in for all memory settings. When I set the memory less than 1G, it informs that it requires more memory and gracefully exits.

Seems to me like a bug. You can inform Abaqus Support about it.

Unfortunately I dont have acces to Windows box to test it in that. Probably you can catch hold of a linux machine somewhere.


Good luck and regards
Guru

http://www.abaqusguru.blogspot.com

 

[Kailar]

allright,

indeed, it's a memory problem. I did a mesh refinement and it seems he started with the Abaqus explicit package.
I've seen the iterations so it's ok for now I think

anyway, thanks!