On May 13, 2013, at 4:40 PM, Chuck Holbert <[email protected]> wrote:
> Yes, now I see this. So, maybe mathematically I have setup the problem > incorrectly. I would like to allow the concentrations at the interface to > increase if the rate of back diffusion becomes greater than the rate of decay. Do you also want the concentration at the interface to decrease due to "forward" diffusion? If so, then it seems your left boundary represents a finite reservoir and I think it needs to be modeled explicitly (If not, I'm not clear what physical process you are modeling). A Dirichlet condition represents an infinite reservoir, so diffusion out of the boundary doesn't deplete the reservoir and diffusion into the boundary doesn't add to it. To make a finite reservoir, you should be able to have one or more cells at the left side that start with some amount of stuff in them and then have a (default) no-flux boundary condition at the left. > The rate of decay at the interface is dc/dt = -kt. > My mistake. The decay rate is first order and given by dc/dt = -kC. With an explicit, finite reservoir, I would be inclined to model the decay with an implicit source term that's only active in the reservoir cells. _______________________________________________ fipy mailing list [email protected] http://www.ctcms.nist.gov/fipy [ NIST internal ONLY: https://email.nist.gov/mailman/listinfo/fipy ]
