tfel
tfel copied to clipboard
thelfer
Hyperelastic material model (Ogden)
Is the implementation of the Ogden material model in MFront limited to 1st order (p=1)?
@mrgprasad The implementation presented here is limited to first order. But it is easy to modify it. You shall use it as a basis to implement your own version.
@thelfer Is there a validation test case or an example for the Ogden model implemented in MFront. I am currently trying to run the model with CodeAster, and need some help with the values for K, mu and alpha. A working example will help a lot. Thanks
@mrgprasad Sorry, I can't find any.. alpha must be strictly greater than 3 according to the code. Any positive value for K and mu shall work...
@thelfer I have set up an example (simple cube being pulled from one side) to test the Ogden model. Ran the simulation, but I can't make sense of the results obtained. Check the results file attached. Even for a small applied displacement the model shows very high element distortion (1st time-step in simulation result). In addition, the model displaces along X significantly. (It acts as if the model is not constrained)
Note: The same model with Signorini (MFront implementation) works as expected.
ogdenTest.zip
@mrgprasad Please find attach a simple unixial tensile test based on MTest. To make it work, I had to change the eigen solver. Please try it and let me know if everything seems ok in MTest. Then we'll have a look at code_aster.
@thelfer Yes, still having problems
Starting model
This is after the 1st time step
@thelfer Looks like I made a mistake. Let me explain the situation:
I was comparing the Signorini model of codeAster with MFront implementation. For the comparison, I compute the reaction forces on the constrained surface. The curves overlapped and I assumed that the MFront model works (my mistake, I should have checked the resulting deformation in paraview). Here are the results:
Code aster Signorini result:
MFront Signorini result:
Clearly the results are not the same. Looks like the cube is moving like a rigid body, although its constrained. Here are the files for comparison. compare.zip
Am I doing something wrong here? Do you have example validation setup for Signorini that works?
Once we rectify this we can discuss about the Ogden model.
Do you have example validation setup for Signorini that works?
Maybe you should have a look at the official tests (which include MFront+aster+GROT_GDEP): https://www.code-aster.org/V2/doc/default/en/man_v/v6/v6.04.187.pdf
@thelfer I used the official test example and both Signorini and Ogden models seem to work. Still have no idea where the error is in my simulation setup. Thanks for the tip!!!
Anyways, regarding the Ogden model:
- Why must the alpha value be strictly greater than 3?
- If one extends the model to higher order, then all the alpha values must be greater than 3 also?
@mrgprasad Great !
To be honest, I only had a rapid view on your files. As I am not a regular code_aster user, I am not of great help here, unfortunately.
Why must the alpha value be strictly greater than 3?
I have seen in my implementation that I used x^(alpha-2). Probably to compute the consistent tangent operator, not sure. Anyway, if x is 0, it could lead to a trouble with alpha lower or equalt to 2 (Of course, if x can't be zero, then my remark is just pure non sense). I let you check !