thermal-stress analysis.

[babaye]

Hi everybody.
help please!!!!
It is my second message for the same problem. At the time of the passage from a thermal analysis to a structural analysis, I use ETCHG,TTS et LDREAD,TEMP,,time,1,file,rth.
In the structural analysis, I need to change the properties of the material of the elements (mat1 to mat2).
With MPCHG or EMODIF, this change is carried out correctly except that ANSYS does not hold of it account at the time of the calculation of the constraints for example; it continuous to use material 1 witch I declared in the thermal analysis.
Can somebody help me please?
thank's in advance!
Barha

 

[Hup]

Hi,
you can assign 1 material with both thermal and structural props.
Than no MPCHG or EMODIF is needed.
Or is there anything against?

H-up

 

[babaye]

hi,

I need two materials for the structural analysis. At the beginning, I assign MAT,1 for all the elements. Then in EACH STEP of the structural analysis, I change mat1 to mat2 for all the elements which check a certain condition (according to the temperatures calculate in the thermal analysis). Thus, I cannot declare both in advance.
If you want to better understand my problem, look please to my other question: how to use MPCHG.
Thank you for your help H-up!!
Braha.

 

[babaye]

I am disappointed!!!
I see that you work on the residual stresses due to welding (you asked lot about that) and now you would like to have an idea on my program. According to your answer I feel that you are not with the height to answer my questions, or those of the others as it is the case for Drej or Alex. You should simply have asked me help for your work !! One should preserve this forum of bad intentions!!!
Barha

 

[meshparts]

Hi Barha,

don't despair... I think about your problem but at the moment I have no other ideas.

Regards,
Alex

 

[babaye]

Hi Alex,
You were right to specify what is noted in ANSYS help: Between load steps in SOLUTION, material properties cannot be changed from linear to nonlinear, or from one nonlinear option to another. Now, what it’s sure is that it is not right just in the /SOLU but also in the /PREP7. Also, I checked that from a linear material to another, the problem disappeared. Now I’m trying to use EMODIF.... I keep you informed.
Thank’s Alex
Regrads,
Barha

 

[FArias]

Hello

have you thought about using EKILL/EALIVE? You could define 2 meshes with coincident nodes, one with mat1 and another with mat2 properties, and decide which element is alive through your temperature conditions. From the pair of coincident elements, one should always be alive and the other "killed".

Just an idea...

Regards
Fernando

 

[babaye]

Hi Fernando!!
It’s a very very very interesting idea!!!!!!!!!!!!!!! It can be my solution!! But quite simply I had never used this method. Is there any specific command or I have just to create two model: ET,1,.. and ET,2,..? and after meshing both i let some elements alive and kill the others using EKILL and EALIVE?
I have to create thermal model to calculate temperatures in thermal analysis then apply them as loads in structural analysis, where I have to check the temperature condition. Can you give me more details about this method.
Thank’s in advance.
Barha.

 

[FArias]

Hello again,

you have to create 2 coincident meshes, with coincident nodes and i.e. have mesh 2 with an initial state of EKILL, something like

cmsel,s,mesh2
ekill,all

already in the solution processor.

Then you would have to transfer the temperature distribution TO BOTH MESHES, that is where it can be tricky: I am not sure, but with LDREAD I think it reads the temperatures for the nodes, that would be OK, if it were on the elements, you have to make sure that you have the temperature distribution on both meshes.

Once you have the temperature distribution, you would have to check the temperature on each ELEMENT and decide whether it should be alive or dead, so that no pair of coincident elements is alive at the same time.

And then, only solve!

By the way, make sure that your elements support the ekill/ealive functionality.

Hope it works,
Fernando

 

[Hup]

Dear babaye,

after several days of being offline I read your response on my 'bad intention'. I am shocked. What a little information made you speculating about the issue of my work (you are wrong) and my qualities.

Good luck to you and your project.

 

[paypes]

Dear H-UP,
It’s me: BABAYE. I can’t log in with this handle, so I opened a new one (paypes).
It is not for you that I address that message, but for LAYTH. Its message was between my two messages of 8 h 20 and 12 H 32. Apparently it was excluded; I do not find his handle anywhere in this Forum. In any case, even with him I shouldn’t have judged his intentions or his qualities with so little information. Now I regret it. As for you, I thank you for your help and your reserve in your message.
Also, I thank you Fernando and Alex for your help. I found where the problem with MPCHG is. At the beginning, I declare: SECDATA, t, 1. I have to delete this number so I can after change my material number. Anyway, I’m trying now to use your method Fernando, perhaps it gives me more opportunities.
My regards for all.
Barha.
NB: I use web translator because I don’t speak English, so sorry!!

 

[Hup]

Hi, Barha,
I see the misunderstanding now. Thanks for this post, I was really curious about the problem in your coupled-field simulation.

With regards H-up