Dear All,
I am modelling a simple D_PLAN element subjected to a sine cyclic change of capillary pressure on one side (Code Aster v16.4.11). I have noticed that both porosity and stress on the exposed side show an increasing trend which seems unphysical. After many trials, I think I have identified the problem on the integration strategy: in R7.01.11, equations 4.1.1-1 and 4.1.7-1 clearly show that applying pc1-pc2 is not the same as applying pc2-pc1 since there is a term related to the final state only (Slq+). This translates into the trend that you can see in the attached plots.
Of course decreasing a lot the time step this problem is mitigated (but never removed), but at expenses of unreasonably large computational time. For instance with daily cycles (frequency = 1 cycle per day) for external input, 10 time step per hours are not sufficient to decrease the average stress increase to more than 0.4 MPa in 60 days. I have tried playing with PARM_THETA parameter (from 0 to 1), but this seems to have no effect.
Can anyone suggest any solution to this problem?
I have attached the case with time step equal to 1 hour and frequency = 1 cycle per day 
and same frequency but time step 0.1 hours 
This is the input:
Thank you,
Corrado
