ABAQUS vs ANSYS

[jingyuan]

Hello,

I am trying to model fluid-solid interaction (FSI). It looks like that ABAQUS is not good on this function (even though it is said that ABAQUS is cooperating with Fluent).

I found ANSYS says it's ANSYS-CFX is a intimate integration of ANSYS and CFX, and has full function in FSI.

However, I am worrying about ANSYS in modeling complicated constitutive laws and contact behavior. Does it has the capability similar as UMAT in ABAQUS? Also, how about ANSYS' contact simulation performance compared with ABAQUS?

Before I spend time on ANSYS, I hope to get some hints about above questions. Thanks a lot.

[Jorgen]

Interesting question. I agree that ABAQUS does not have strong FSI support. As an option to ANSYS, I have heard good things about the FSI capabilities in LS-DYNA and Adina.

ANSYS does have strong non-linear material modeling capabilities. You can, for example, write similar user-material subroutines as the UMAT for ABAQUS.

ANSYS is also pretty good at contact simulations. I am sure which package has the better contact implementation.

- Jorgen

[jingyuan]

Thanks Jorgan. After read your reply, I am interested in LS-DYNA.

I have a few questions about LS-DYNA. Does LS-DYNA also have user subroutine function? Does it has a user subroutine simillar as UMAT? I checked LS-DYNA website and can not find these information.

In the LS-DYNA theory manual, I found they provide lots of material model. Looks like they provide some models related with soft tissue, such as : model 91-soft tissue (Mooney-Rivlin), model 127-Arruda-Boyce model......... All these are for passive deformation.

I found there is also another a active contraction model 156-muscle.

Can I combine model 156 with any of above passive deformation material? How this is implemented in LS-DYNA?

I never used LS-DYNA and just want to get some basic idea before I decide whether spend time on LS-DYNA. Thanks.

-Jingyuan

[Jorgen]

Here are a few comments:

:arrow: Yes, LS-DYNA allows for user-defined material models (Materials 41-50). It is done very similarly to the ABAQUS UMAT/VUMAT subroutines.

:arrow: Mat156 is only for truss elements. I have not used that material model so I am not sure how well it works. I suggest that you read the manuals.

- Jorgen

[jingyuan]

So Jorgen, have you ever used UMAT to implement muscle active contraction? Do you have suggestion about this?

PS: Thanks Jorgen for provide such an excellent FEM resource. I think it is the most professional one.

[matsgd]

Interesting question. I agree that ABAQUS does not have strong FSI support. As an option to ANSYS, I have heard good things about the FSI capabilities in LS-DYNA and Adina.

ANSYS does have strong non-linear material modeling capabilities. You can, for example, write similar user-material subroutines as the UMAT for ABAQUS.

ANSYS is also pretty good at contact simulations. I am sure which package has the better contact implementation.

- Jorgen

ADINA FSI is pretty good, and the contact algorithms have been improved significantly for the next release 8.4 (coming soon). However, the constitutive modeling capabilities are limited. I can provide some insight into this if I know more specifically which type of constitutive behavior is to be modeled.

[hhspiny]

Interesting question. I agree that ABAQUS does not have strong FSI support. As an option to ANSYS, I have heard good things about the FSI capabilities in LS-DYNA and Adina.

ANSYS does have strong non-linear material modeling capabilities. You can, for example, write similar user-material subroutines as the UMAT for ABAQUS.

ANSYS is also pretty good at contact simulations. I am sure which package has the better contact implementation.

- Jorgen

It is my understanding that one has to write an user element in ansys to implement user material, not like in Abaqus, one can write a general material model which could be then used with any element. That is probably why most people use Abaqus dealing with customized material models.

Abaqus claims it has better contact convergence and oscillation control. In generaly, my project confirms this, but it is hard to generalize any experience. I blieve both would be sufficient for most applications.

We also evaluated FSI capacity recently. It looks to me the weak coupling of Abaqus/Fluent makes more sense in more general industrial cases, as most coupling are not strong. Now Ansys purchased fluent, who knows what will happen next.

[Jorgen]

I just want to clarify that the ANSYS USERMAT subroutine can be used to define a user-material model quite similarly to the ABAQUS UMAT. It is not necessary to write a coupled user-material and user-element with ANSYS.

Jorgen