field variable dependency - hyperelastic material

[IceBreakerSours]

I have been trying to assign field variable dependency for materials other than Marlow and running into issues. I have uniaxial tensile stress-strain data that I'd like to be able to scale with a field variable (from 0 to 1).


I guess I could scale it by "find and replace in place" using MATLAB or Python but I was hoping to avoid that.

It is a 3D model and the analysis is static.

http://www.eng-tips.com/faqs.cfm?fid=376

http://www.eng-tips.com/faqs.cfm?fid=1083

 

[IceBreakerSours]

Let me also try another way of asking this question.

What options do I have to scale the nonlinear stress-strain uniaxial tensile test data? Currently, I am doing the following:

*******************
*parameter
param=0.1
row1=0*param
row2=10*param
row3=20*param
..
..
..

*******************
Hyperelastic, ogden, test data input
*Uniaxial Test Data
<row1>,0
<row2>,0.5
<row3>,0.6
..
..
..

Also, just to be sure, isn't X-axis (direction 1) as defined in the orientation card assumed by ABAQUS to act as the primary axis of loading?

http://www.eng-tips.com/faqs.cfm?fid=376

http://www.eng-tips.com/faqs.cfm?fid=1083

 

[MechIrl]

What I've done for this in the past (for plastic materials) is to make some big tables with all of the stress-strain data values for each value of the field variable, which I think is what you meant by your MATLAB/Python option? Other than this maybe you could make a fit to the data then scale the coefficients using either field variables or something like UHYPER.

 

[IceBreakerSours]

Thanks MechIrL.

make some big tables with all of the stress-strain data values for each value of the field variable

Correct; I guess I am doing the same thing with a parameter definition except that in the approach you are suggesting all possible stress-strain curves are defined in the INP (intermediate ones being interpolated by ABAQUS) whereas in my current approach only one stress-strain law is defined for one value of the assigned parameter.

Other than this maybe you could make a fit to the data then scale the coefficients using either field variables or something like UHYPER.

This sound problematic to me at first glace. I may have it wrong but I can't simply scale coefficients and expect a scaled non-linear relationship in uniaxial stress and strain out of a material model like Ogden, can I? I may have to do curve fits on a few possible curves, see how the parameters vary, and then, if possible, assign an appropriate profile to field variable dependency. What do you think?

http://www.eng-tips.com/faqs.cfm?fid=376

http://www.eng-tips.com/faqs.cfm?fid=1083

 

[MechIrl]

Yes, the second approach may yield a complex relationship between the material parameters and the field variables. Determining a relationship might well be a pain and of course has no guarantee of leading to an accurate fit. Using UHYPER might be a slightly neater approach, but would still require some way of fitting the data as best possible.