Muscle Modeling

[DaVinci]

Greetings fellow members

I have started with a project involving the modeling of muscle behavior.
I have the geometry of the specific muscle bundle in Abaqus.

I will be using a hyperelastic strain energy function, which incorporates the following:

1. large deformations
2. anisotropy (due to fibre orientations)
3. muscle activation
4. non-linearity

I don't have physical test data though. (axial/biaxial/shear tests)
I am considering writing up a Umat, and also Uel.

Is it acceptable to validate a model computationally?
I am considering starting out small, working on one complexity issue at a time.

If anyone has experience with muscle models in Abaqus, assistance with this matter would be great.

Kind Regards

[Jorgen]

Sounds like a great topic! Very interesting.

I think that your approach sounds fine. I would also break the problem down into smaller parts. In this case I assume that the validaton will involve both validation of your user subrotine implementations, and validation of the developed material models. Some of the validations can certainly be done numerically.

Best of luck and keep us updated of your progress.

- Jorgen

[uraela]

Hi all,

I am an undergrad working on modelling the pelvic floor muscle. For the material, I would like to use a model developed by J.Li et al. (A nonlinear anisotropic model for porcine aortic heart valves, 2001) but have no idea how I should go about implementing it.

The nonlinearity is in terms of a strain dependent Young's modulus, no strain energy function was mentioned in the paper. Which subroutine should I use? ORIENT, UMAT or UHYPER? I was actually even thinking of calling Abaqus from Matlab then updating the young's modulus in Matlab. Any best/simplest way to do it? Hope someone can help. THanks lots!:o

[Jorgen]

I would use an Abaqus UMAT subroutine. It should not be too difficult to write an UMAT since the response is basically linear elastic.

-Jorgen

[DaVinci]

Greetings

What are the basic (very basic) steps to creating user-defined items such as: elements, materials, hyperelastic models etc.?

Is the major work of defining sub-routines done in Fortran? (and how exactly does Fortran work? eg compilers, latest versions, availability etc)

Regards
DaVinci (YK)

[Jorgen]

Yes, the subroutines are typically written in Fortran. Abaqus presents their examples in Fortran-77. I much prefer Fortran-90 (or newer).

You should make sure that you use the compiler and version that are recommended by the FE program. For Abaqus, that is listed on their web site.

-Jorgen

[DaVinci]

Once again, thank you for the assistance.

I am using Fortran90 - version 6.5, and Abaqus CAE version 6.71. I am now on my way to creating sub-routines for Abaqus.

Best Regards
Davinci (YK)

[DaVinci]

Greetings

I am using Compaq Fortran v6.5, and Abaqus v6.7-1. (Abaqus was installed first) I am running this on Windows XP.

When trying to use a user-subroutine (umat) in Abaqus, i get the "ifort.exe not found in PATH" error. Please advise on the best way to sort this out. I've searched forums, but havent found a working solution.

I was advised to run the job from Abaqus command, after running dfvars.bat in the same window.

Also, not exactly sure on how to view results from abaqus command window, after running a job, or even that it worked.

Kind Regards
Davinci

[DaVinci]

Hi again

I ran the Abaqus verification again, and also checked the Abaqus .env file.
Both indicate that it is searching for Intel Fortran version 8 or higher.
I am currently under the impression that Compaq Visual Fortran v6.5 is not sufficient.

Regards
Davinci

[DaVinci]

Hi again

Also... I just checked the log file created after running a job from Abaqus command window. It gave the "infort.exe ..." error as well.


Regards
Davinci

[Jorgen]

The "ifort.exe" executable is the name of the Intel Fortran compiler. If I remember right, the Compaq fortran compiler is called "df.exe". It is quite possible that your Fortran compiler is simply too old :(

You might be able to solve the problem by modifying your abaqus_v6.env file. Specifically, change the "ifort" strings to "df". I am not sure if that will work or not, but it might be worth a try if you do not want to purchase the Intel compilers.

- Jorgen