Failure of convergence in the simple 1D flow model
Dear Sir or Madam,
I just built a simple 1D flow model (EOS9). I found it could not converge with my experimental parameters, however, it worked when I changed the porosity and permeability, (i.e., porosity n=0.5748 changed to n=0.1; permeability k=2.111E-19 to k=2.111E-13). I do not know why it could not work with my original experimental parameters. Is there anybody can help to revise my model?
Best regards,
Lang
2 replies
-
Lang,
The first question is: What is it you want to simulate?
You inject water in the top element, but the top element has a large volume, so it is held at 1 bar regardless of what you inject. What boundary conditions do you want at the top, Dirichlet or Neumann?
If you set the volume to *E-6 instead of *E20, and increase (or remove) DELTMX, it runs (reaching steady state very quickly without developing an unsaturated zone at the bottom).
Your capillary pressure function is unrealistic for a clay. Set CP(2) at least tow orders of magnitude higher (it's in Pa!). Consider using one of the nonlinear functions for both relative permeabilities and capillary pressures with realistic parameter values.
Set RE1=1E-8, and then it runs even if you keep the top fixed at 1 bar (which is a very delicate run to do).
At any rate: answer the first question first!
Hope this helps,
Stefan
-
Dear Dr. Stefan Finsterle,
Thank you for your prompt and helpful reply.
Actually, I wanted to simulate the contaminant transport using TOUGHREACT, and this was just the flow.inp for the TOUGHREACT. And I wanted to simulate the constant boundary concentration, so I set a big volume in the first cell. The boundary condition should be the Dirichlet boundary.
I checked the other input files for TOUGHREACT, and I thought they were okay, since it worked when I just changed the flow.inp file, which means only the flow.inp was incorrect. So I doubted there was something wrong on my flow.inp.
Best regards,
Lang