Hi PolymerFEM community,
I'm a beginner on this matter but recently I'm trying to develop a constitute material model for a Thermoplastic Elastomer using MCalibration. I'm an ANSYS user, but I realised that the material models options available in ANSYS, "are poor" for my problem. So, I decide to use ABAQUS, and PRF model could be a solution.
Firstly, I have started with the uniaxial tests (tension and compression), after that I will include de viscous tests. I have obtained a good fit using PRF-2Net-Yeoh-Power, but when I go to ABAQUS I realise that parameter a1 (pressure correction factor) was negative, and this factor must be positive. However, it a1 positive (or 0) the fit for compression side is very bad. See below the two cases:Â
Do you have some idea or any tips to solve this problem? Thanks in advance
Kind regards,
AlexandreÂ
Hello Alexandre, As you noticed, for some reason Abaqus does not allow the parameter a1 to be negative. MCalibration does not have the limitation. In order to simulate your experimental data accurately you cannot use the PRF model. Fortunately, the PolyUMod library contains models that can predict the experimental results. Also note that the PolyUMod library is available for both Ansys and Abaqus.
Finally, as you likely noticed, your experimental data has lower yield stress in compression that in tension. That is unusual, but not necessarily incorrect.
/Jorgen
@jbrektor thanks for your reply!
I am able to fit the experimental data quite well using a PN model from PolyUMod. Am I able to use it in my FEM using the trial license?Â
Regards,
Alexandre
@jbrektor Hi Dr Jorgen...I have noticed a similar issue in using JC model in ABAQUS. I have calibrated a JC model for metal strain rate sensitivity study using MCalibration. The parameter C in JC model should not be negative in ABAQUS. But when Calibrating with MCalibration it perfectly fit with negative value.
As a remedy I have set the lower limit for C as Zero in MCalibration and ran the test. Fortunately The model was not effected by just a slight change in C. So my problem got solved for that problem.
I just posted this for others for awareness.
Thanks.
Yes, that should work fine. Note, however, that the PN model due to its modular nature runs a bit slow. I typically only use that model when other models do not work.
/Jorgen
@jorgen thanks for the tip!
Do you have some suggestion about a constitutive material model to fit my experimental data before to use a more robust model (PNM)? Thanks in advance.
Regards,
Alexandre
Since that behavior is so unusual, your best bet might be the PNM.
/Jorgen
@jorgen, the previous data was wrong (compression test). Now, with the correct data, I'm able to fit the model properly:
Thanks for your help.