Hi all. I am rather new to julia and totally new to Gridap. Before starting a new time and energy consuming project I would like to know if Gridap is suitable for solving “highly non linear problem” (the Richards equation, Richards equation - Wikipedia). I would solve this equation first in 1D and then in 2D but having multi-material (different soil horizon with different hydraulic properties). For now I am just writing down the 1D-one material problem (weak forms, jacobians etc) on a paper sheet and before starting to code I would like some indications by you…
1 Like
I believe it is suitable. Checkout the tutorial examples
https://gridap.github.io/Tutorials/stable/
In particular checkout the 4th example and the 7th.
2 Likes
Yes it most definitely is suitable: @aerappa spent part of his PhD thesis doing this. See for example his thesis, especially chapter 3 and the papers referenced therein:
2 Likes
Hi @Paulo_Jabardo , as @ffevotte mentioned I was able to solve some benchmark tests of the Richards equation using Gridap.jl. However, I basically wrote the nonlinear solver and time integration myself because I needed certain intermediate quantities between nonlinear steps to compute a posteriori error estimators. In addition, there was no high level API at the time in Gridap.jl for nonlinear transient problems.
However, it appears that such an interface now exists as per Tutorial 18. I think this would probably be a good place to start, but if you have some specific questions don’t hesitate to send me a message.
2 Likes