Re: Non linear diffusion and prescribed flux BC


In article <44035327$0$5990$4fafbaef@xxxxxxxxxxxxxxxxxxx>,
nelson <nelson1977@xxxxxxxxx> writes:
Peter Spellucci wrote:



> nelson
>
>
>
method of fictitious points?
I assume that "0" means the x-position:
u(0)=(1/2)*(u(delta_x)+u(-delta_x))
u(0)_x = (1/(2*delta_x))*(u(delta_x)-u(-delta_x))
use the differential equation also at the boundary x=0
this gives you the necessary additional nonlinear equation .
your system is nonlinear anyway.
hth
peter

Do you mean to compute u(-deltax) from boundary conditions:

u(-deltax) = u(deltax) -(2 * dx/q0)/K(u(0))

discretizing the equation and substituting u(-deltax)?

thanks,
nelson
no
you must replace u(0) inside K also , since u(0) is unknown:
q0=(1/(2*delta_x))*(u(delta_x)-u(-delta_x))*K((1/2)*(u(delta_x)+u(-delta_x)))
hth
peter

.