Elastic modulus as a function of relative density

[liammo29]

Hi,

I am totally new to writing USDFLRs. I know I need an input file and a subroutine. Do I write the input from scratch or can I obtain that from my model. I am then trying to make the elastic modulus as a function of the relative density. I am assuming this is similar to the examples given for modulus as a function of strain? Say for example E = 2*RD + 2. If anyone could write an example that would be amazing. My case is just a plate with a porous metal plasticity relative density of .9. Once this is done, if I run a job with this would the resulting stress/strain curve have a new slope based on my relation?

Thanks so much.

 

[Michael Koopman]

Why not just use the Gurson material model that is available within Abaqus? There should be a simple way to modify the pore size within this material model that can meet your goal directly?

 

[liammo29]

Because as far as I can see the Gurson model only account for a change in yield stress and does not actually change the slope of the stress strain curve due to a porosity. I have tested a simple 3d sound plate and applied to porous plasticity model with RD .9, q1 1.5, q2 1, q3 2.25 and got the same modulus as if there was not porosity.

 

[Michael Koopman]

I misunderstood your question to regard the onset of plasticity which is the general interest in a material above the density ratio you identified. There is actually a change in Young's Modulus but the amount is hardly appreciable. Total strain energy data and lower q3 values are what I expect to identify better candidates for Gurson's model. I suspect that Cam Clay is too loose bound and Gurson is too ductile for a aggregate type with resin or ceramet binder?
The Cap model should allow you to get where you want but the corner behavior is difficult with no assurance of local convexity. A Cap with a hyperelastic user routine sounds like a difficult problem to tackle. Specific preform test models are the best approach to avoid winging it on general problems. DCT was the forerunner on this approach, so thanks for the recollection. You will plough through it, keep plugging.
I believe the Cam routine allows for a one-sidedness and this may also lend a resolution that can be satisfactory, especially if you are modeling plies where the bidirectional wetting may poorly represents the adhesion behavior.

 

[liammo29]

Thanks for the response. Definitely some interesting things for me to look at. The change in modulus due to porosity below 10% can actually be almost a 20% reduction. Right now my goal is to find a way to define the elastic modulus so that it is a function of the inputted relative density. I am thinking a user defined field variable is the only way but I am really not sure at all. Say I said E = 2*RD+2. How could I write a USDFLD to achieve this?