Excessive distortion error

[FEMbiaatch]

I'm working on the simulation of needle penetration into a thermoplastic polyurethane foil (0.4mm thick) using cohesive interaction behavior between two parts of a phantom corresponding to the outer diameter of the cutting edges of the needle tip. As for the model, I'm using a symmetric model by considering a needle to be a rigid part with frictionless contact between the surfaces. The mesh of the phantom is relatively coarse to keep the stable time inc to be reasonable. I have performed uniaxial tensile and relaxation tests on the material to be calibrated into the model. The model runs smoothly when I consider the phantom to be Neo Hookean hyperelastic material with incompressibility, but I obtain excessive distortion error with the Ogden hyperelastic model, which would be a better fit for the material. What would be the reason for such behavior? Any input would be much appreciated. Thanks!

 

[FEA way]

Did you try using the material evaluation tool in Abaqus/CAE ? It allows you to check the performance of different hyperelastic models for your data in different conditions and can be used to find instabilities in material behavior.

 

[FEMbiaatch]

Yes, I tried using material evaluation in Abaqus CAE, which shows that it is stable for all strains. But I can see a warning which states that the hyperelastic material is unstable in the .dat file.

 

[FEA way]

You could try increasing the range of strain considered for evaluation. According to the evaluation, are there any other hyperelastic models that fit the test data well?

Damage is modeled using surface-based cohesive behavior in this case, right ? Have you considered using cohesive elements instead ?

 

[FEMbiaatch]

What would be the reasonable range of strain to consider for the evaluation? I have to check for other hyperelastic models.
I considered surface-based cohesive behavior as it is computationally less expensive than cohesive elements. Would it be better to model using cohesive elements?

 

[FEA way]

It should correspond to the range of strains that may occur during the simulation.

Various papers about needle insertion modeling in Abaqus usually describe cohesive elements being used for this purpose. They are more complicated but can potentially provide better results in such cases.

Also, take a closer look at the location with excessively distorted elements - maybe the mesh should be refined there or there's another reason for the distortion.