On Mon, Jul 2, 2018 at 3:16 PM, Zhekai Deng <zhekaideng2...@u.northwestern.edu> wrote: > Hi Daniel, > > Thanks for the pointed sources. I think I still need some help on how to > setup my boundary at irregular triangular gmsh. Even though the locations of > my boundary conditions are simple (two straight lines inclined with fixed > degree), my mesh generated from gmsh is triangular and irregular. Thus, my > boundary location lines are not fully aligned with the face edges in the > mesh. I wonder how to setup the "where" option in case like this?
It depends if the boundary condition is defined with a sharp or diffuse formulation. If you're using a sharp boundary formulation then the mesh faces need to be aligned along the boundary. In that case I don't know any other way. Gmsh will certainly let you do that. > Right now, I am thinking that I need to mark my boundary by myself. I have > been trying a lot of statements like following but it seems doesn't work > out: > phi.faceGrad.constrain([a*local_shear_rate*(1-phi.faceValue)*phi.faceValue], > where = np.abs(yFace -(leftheight - (L + > xFace)*np.tan(25.0/180.0*np.pi)))<eps) That probably won't work. If the boundary conditions are internal to the domain then they need to be defined as part of the equations, not using constraints. > I am not sure whether It is better to formulate my boundary condition as a > source term instead of using faceGrad.constrain... Exactly, formulate the boundary condition as a source and variable diffusion coefficient. -- Daniel Wheeler _______________________________________________ fipy mailing list fipy@nist.gov http://www.ctcms.nist.gov/fipy [ NIST internal ONLY: https://email.nist.gov/mailman/listinfo/fipy ]
