Re: Non linear diffusion and prescribed flux BC


In article <1141122678.582041.146990@xxxxxxxxxxxxxxxxxxxxxxxxxxxx>,
"nelson" <nelson1977@xxxxxxxxx> writes:
but how can i use the expression of q0 you wrote in solving my
equation? I use Picard iteration in order to solve it (and i can't
change the solver.). The euation is a Richards equation and in none of
the paper i read they explicity implement prescribed flux boundary
conditions... so sorry for my question..

u(0) is unknown but i campute K(u(0)) using the value at the past
nonlinear iteration level.

nelson


but you know the function K(.)! so this adds one more nonlinear equation
to your discretized system which is a nonlinear anyway. if you would use
the u(0) from the previous timestep you would introduce a first order
error in the boundary conditions. there is no reason to do this
hth
peter
.