Compare MGIS and MTEST (python)

[NsmKsm]

Hello,

I am implementing an elasto-visco-plastic problem (for polycrystals) in python and I want to use Mfront as an integration function (for behaviour) through MGIS. In order the make sure that everything is correct, I picked the "Meric Cailletaud" model as an example and I compared the results generated by MGIS and MTEST.

My example is found in this link: https://github.com/NsmKsm/compare_mgis_mtest,

The generated results from MGIS and MTEST are different and, sadly, I could not figure out what's happening.

Thanks in advance,
Best,

[thelfer]

Ok, an easy one :) You set the time increment to 0 in the MGIS case. With the following correction, everything is fine:

    dt = time[tn] - time[tn - 1]
    mgisbv.integrate(m, mit, dt, 0, m.n)

[NsmKsm]

Thank you for fast response.
So the third argument is the time increment. What about the fourth ?
Also, is there a way to impose mixed conditions (e.g. imposing strains and stresses to simulate tensile tests) ?

[thelfer]

So the third argument is the time increment. What about the fourth ?

Fourth argument is the number of the first integration point. This way, jobs can be easily r parallelized.

Also, is there a way to impose mixed conditions (e.g. imposing strains and stresses to simulate tensile tests) ?

In MTest, certainly.

In MGIS, this is not straightforward. MGIS is meant to be called by a FEA/FFT/MPM solver: it does not handle equilibrium out of the box and is not meant to.

[thelfer]

BTW, you may want to read the C++ documentation which details the meaning of each argument :)

[NsmKsm]

Thank you.