Dear All,
I am going to consider rate dependency of the human cortical bone. it gets brittle ( higher elastic modulus, shorter plastic strain range up to failure).
I tired MAT_24 and MAT_124 by definite a strain rate dependent table for LCSS card, but these material models imply that the last effective plastic strain for all stress vs. effective plastic strain curves must be the same. Also, I suppose these materials do not consider the strain rate effect on the elastic modulus.
I would greatly appreciate sharing any idea describing how to model the attached graph .
Hello Mr. Khakpour,
MAT_24 does not consider strain rate effects on elastic modulus but it seems like it is best suited for your simulations.
You will need to create a table with *DEFINE_TABLE that contains the strain rate of your single curves as x and the curve ID as y. You can use a table as a LCSS by inserting "-" before your table id.
Please note that material models always use a relationship of true stress to true strain. Since your curves have a clear peak I assume you are using engineering stresses and strains which is not correct and will lead to inaccurate simulations. Generally the true stress will never decrease unless your specimen is fractured.
After you have converted your curves you will need to lengthen your shorter curves. I recommend fitting your curves with the Hockett-Sherby equation and then using the parameters to extrapolate your data.
Regards
Gino Mathlouthi
Dear Gino Mathlouthi
Thank you so much for your comprehensive answer. Actually the point is that for different strain rate, fracture happens at a different plastic strain and decreasing the stress is due to reaching failure point. Indeed, I need to consider different failure point for each strain rate ( and consequently not equal plastic strain curve length), which LS-DYNA does not accept.
BR,
Shahab
Dear Shahab Khakpour,
materials generally harden with increasing strain. The gradient of the true stress over the true plastic strain will be large at first and then decrease, but it will never become negative.
Your curves at high strain rates quickly reach a point at which the gradient is almost zero. This will lead to a small increase of load resulting in very large plastic strains, which means your material will fail much sooner on a macroscopic scale.
You should be able to model your experiment like this with a *MAT_ADD_EROSION condition containing your failure strain. You will probably have to fit that value to the data you have.
At some point your specimen will start to neck, which means that the cross-section will decrease faster than the strain will harden the specimen. This will lead to your force or stress decreasing.
If you want to perfectly model your experiments, you will have to add a damage model like GISSMO or DIEM to adjust the stress fall-off after necking.
Regards
Gino Mathlouhi
- Badges
- Science topic
- Similar topics
- Engineering
- Modeling