Modelling a Material /MAT/LAW28 (HONEYCOMB) or /MAT/LAW36 (PLAS_TAB)

Albert Grima

Hello,

I'm currently working on characterizing a honeycomb material, and I've conducted a compression test to obtain a stress-strain curve. Now, I'm considering the appropriate method for characterizing this material in Radioss. Below is the curve:

Given that it's a honeycomb material, my initial inclination is to characterize it using /MAT/LAW28 (HONEYCOMB). Accordingly, I believe I should extract the Young's Modulus (E) from the initial portion of the curve, and the yellow region should be incorporated into the YFUNCxx section, starting from 0 strain. Am I correct in this understanding? Given that it is compression, strain must be negative.

I've also been contemplating characterizing it as /MAT/LAW36 (PLAS_TAB). Would the rationale for the graph be similar? 

So, what I'm trying to grasp is how the program handles the data. My assumption is that it simplifies the initial slope by providing the Young's Modulus, and then proceeds to the "yellow part of the picture," continuing from there. Am I correct in this understanding?

Thanks in advance

Polyvios Romanidis

Hi Albert,

First and foremost, honeycomb materials are orthotropic and so if you are simulating them with solid elements (macroscopic approach) you should use an orthotropic material law (/LAW28).  If you use shell elements and the exact honeycomb geometry, you can also consider the /LAW36 because microscopically the material is isotropic.  In that case you will have more elements and lower time-step because of its size.

The way you insert your data (and its signs) are determined by the Iflag formulation you use for your material.

The default value is the volumetric strain!

Your material will follow the elastic behavior based on the young modulus you have inserted for every direction and when it meets the provided curve will follow the data you inserted! 

You have to be careful with how you insert the strain data and you have to insert some values for tension for sure.  Most of the times we have a small region parallel to the volumetric strain axis where compression and tension behavior met as you can see in the picture and you can follow that.  It is crucial to define tension values, because otherwise the curve will extrapolate based on the slope your last values set and probably cause inaccuracies in the modelling.

Polyvios