How to set fixed boundary conditions based on element status

[Hanrry]

I'd like to set fixed boundary conditions on some element, but not at the beginning of simulation, but during the analysis.

I want to impose fixed boundary conditions on these failed element during analysis,

however, the *boundary keyword has the format:

*Boundary, Fixed or *Boundary, USER

element set, degree of freedom constrained

I can't use *Boundary or DISP Subroutine because i don't know the failed element priori.

If there any way I can impose fixed boundary conditions during analysis based on element status (failed or not)

 

[Hanrry]

please help me

 

[akabarten]

Hi,

If there any way I can impose fixed boundary conditions during analysis based on element status (failed or not)

There is no such option.

Boundary condition activation/deactivation can be done only between steps (OP=NEW option).
It means you need to know time when you want to turn on/off boundary condition a priori.

Regards,
Bartosz

 

[Hanrry]

Thanks for your reply. I know there's big limitation on *Boundary keyword, I'm thinking of use subroutine to get around, for example, can I define the failed element in the previous step in a single element set by some subroutine, and then use *Boundary, Fixed in the next step.
I know how to do it manually, but the difficulty is how to define that element set automatically during the analysis and passed it into next step *Boundary keyword

 

[akabarten]

Hi,

Just for my understanding. You want to have no BC and at specific time (related to element fail time) apply BC.
If I am right please find my comments.

1. DISP subroutine applies BC always, you can not switch on/off BC this way, you can only control BC value.
What do you want to define for U array for time when BC is not active?

2. My understanding of *BOUNARY, FIXED is a little bit different but here I might be wrong
It allow to use current value of BC for further step. But you have to first have this BC active and then
you can use it. If you have no BC at the begging than it will not work.

Correct me if I am wrong but fail element is removed from the model so what point is to apply BC for this element?
I have no experience with element failure so am asking.

Regards,
Bartosz

 

[Hanrry]

Yes, that is exactly what I want. before the time of failure, the element behaves normally, but when the stress/strain reaches a critical value, it will fail. and them I want to make the nodes of the failed element fixed from then on (for example, a minor crack is generated during the analysis and I want to the crack surface to be fixed when it come into being).
So
1) when BC is not active, all the elements behave like normal stress-strain element.
2) I know that *Boundary,Fixed/USER all ways apply BC and can't be switch on/off during the analysis, So I'm thinking is there anyway to update the element set that *Boundary applies to after every increment.But right now the only way I know is to edit element set manually in the input file. That's so sad.

Thanks,
Hanrry

 

[mrgoldthorpe]

This could be involved, but you could try connecting your intended BC nodes to the boundary via (say) simple truss elements. You then vary the elastic stiffness of each of these truss elements via (User defined) field variables.

Initially set the elastic modulus of any particular truss element to a very high value so the truss is stiff and simulates the BC.

When you wish to relax the BC then the stiffness of a particular truss element needs to be ramped down to a small, non-zero value over a reasonable period of analysis time in order to avoid convergence issues.

 

[Hanrry]

Thanks for your reply. but I don't quite understand how to connect BC to truss element, Can you please explain more in details? Because in my model, I'm using CPE4 elements with XFEM. SO when damage initiated I want to impose a fixed BC condition. So my purpose is not to relax boundary conditions during the analysis, but to impose boundary conditions based on element status. Because I don't know which element will fail at the beginning, so I can't impose BC on the nodes priori,

Regards,
Hanrry