Assigning atmospheric boundary conditions
Hi all.,
I am a relatively new user in TOUGH3.
I am using TOUGH3-EOS7CA module for quantifying methane gas emissions into the atmosphere.
I am very unclear about modeling an atmospheric element in the ROCKS block or assigning an atmospheric boundary condition to a geometry that I have created.
I read about specifying a Dirichlet boundary condition with large volumes and using this free software AddBound.exe in the TOUGH3 manuals. I could not find this software or the several pdfs which were put up as links in some of the posts related to this.
Could someone help me in this regard?
4 replies
-
Hi T G,
You can find AddBound and all other small mesh-manipulation codes zipped up in Ancillary.zip on the TOUGH2 web site at https://tough.lbl.gov/licensing-download/free-software-download/ (or download slightly enhanced versions at https://www.finsterle-geoconsulting.com/download/) .
It is, of course, not necessary to use AddBound to generate atmospheric boundary conditions; you may use any other convenient way to generate large elements at the top boundary.
It seems that the links on the Forum to the Tips&Tricks files are indeed dead (I informed LBNL about it). I therefore attach some slides you may find useful (specifically, see Slides 19-23).
Best,
Stefan
-
T G,
The concept of large elements to apply Dirichlet boundary conditions is that due to its large volume, any inflow or outflow of mass or heat to that element won't change its thermodynamic state (i.e., pressure, temperature, mass fractions). That is exactly what you want in a Dirichlet boundary element. (It's not just a numerical trick, but represents actually the physics at such boundaries).
Now, to specify time-dependent Dirichlet boundaries, you not only have large volumes, you also have very large sink/source terms, so the conditions in the boundary element actually do change (making them time-dependent); but again, the inflow from or outflow to regular-domain elements do not impact the conditions in the boundary element (thus, they are indeed Dirichlet boundary conditions). Again, this is exactly how the physical systems behave (large flows of air masses on a windy day or large heat inputs lead to atmospheric pressure fluctuations, but the pressure is not affected by the comparatively very small exchanges of soil gas with the atmosphere; tidal effects lead to large water movements in the ocean, put seal levels are not affected by the comparatively small exchanges of water with rivers or the sea floor).
Finally, the fluxes between regular elements and large-volume boundary elements are correctly calculated at the interface between these two domains, driven by the gradients between the system states in the two connected elements.
Hope these explanations dispel your doubts about this approach.
Stefan