Permeability (I) for a 2D cylindrical model
I have a query regarding initializing anistropic permeability in ROCKS for a 2D cylindrical geometry model. In the 3D case it is quite clear that PERM(I); i =1,2,3 stand for permeability in x,y,z directions. I would believe for a cylindrical model that perm1 and 2 can be set to be equal to the radial permeability and that perm 3 would be the vertical permeability. But some clarification on this would be really useful.
The short answer is "yes", you can do what you propose. However, fundamental to TOUGH, the three PERM(i) are just that, i.e., three values that can be assigned arbitrarily to any connection, i.e., there is no reference to a coordinate system (another fundamental TOUGH concept that is important to understand). It is only by convention that the three values refer to x, y, and z (a convention also followed the MESHMaker's XYZ grid generator). For a radial mesh, you can choose to do what you propose, and it is a sensible choice, even though I don't know on what basis you select i=1 vs. i=2 , as a radial mesh is just a 2D mesh. Check what MESHMaker's RZ2D does!
You can make good use of this flexibility, e.g., to specify special permeabilities for special connections (such as fracture-matrix connections) regardless of their spatial orientation.
Hope this is indeed clarifying,