Concentration gradient at boundary

Log Out | Topics | Search Moderators | Register | Edit Profile

Author Message   tom johnson (tom_j) New member Username: tom_j Post Number: 1 Registered: 03-2006 Posted on Thursday, March 16, 2006 - 08:23 pm:    I am using version 5.0.9, 2d professional version. I am having difficulty in dealing with a sharp rise in a concentration variable at a boundary. It is a time dependent problem of ioinic migration in an electrically active medium. The concentration profile at this boundary does not result in a smooth ascension/descensiion. I have used mesh control parameters to refine the mesh to handle the steep gradients and use small initial time steps to control error. Do you have any ideas of how to stabilize the concentration rise at this boundary? Regards, Tom 1.3 K ionic_flux.pde "ionic pde"   Robert G. Nelson (rgnelson) Moderator Username: rgnelson Post Number: 565 Registered: 06-2003 Posted on Monday, March 20, 2006 - 02:59 pm:    1) There appears to be no X-variation in this problem, so you could save an enormous amount of time by shortening the X-length by a factor of 100 or so.... 2) Nonlinear terms like dx(C)*dx(V) or dx(V)^2 are treacherous because they lose the sign of the terms involved. When the values start out at zero, oscillations that preserve the magnitude cannot be distinguished from smooth solutions. I'm not sure this is the cause of the trouble here, but if there is another formulation that is single-valued it might be worth trying.