a banerjee
09-02-2008, 01:29 AM
Hi,
I am trying to solve a nonlinear equation similar to eq.19.6.44 of NR (2nd edition, pg.) , with the nonlinear term equal to exp(u) instead of u^2, with mixed boundary conditions: inhomogeneous Dirichlet conditions (at x=0,y=0) and homogeneous Neumann's conditions (x=0,y=0).
I guess the mgfas code in NR is written for homogeneous Dirichlet's boundary conditions. Is that correct?
I read section 19.4 (eqns. 19.4.16 - 19.4.27).
Should I use a cosine expansion for u = u_j(x)cos(j*pi*y/L) (summed over j) as this section suggests to take care of the homogeneous neumann's b.c's?
Is it practicable in this case? Is there an easier way? Please let me know.
Thanks,
Arunima.
I am trying to solve a nonlinear equation similar to eq.19.6.44 of NR (2nd edition, pg.) , with the nonlinear term equal to exp(u) instead of u^2, with mixed boundary conditions: inhomogeneous Dirichlet conditions (at x=0,y=0) and homogeneous Neumann's conditions (x=0,y=0).
I guess the mgfas code in NR is written for homogeneous Dirichlet's boundary conditions. Is that correct?
I read section 19.4 (eqns. 19.4.16 - 19.4.27).
Should I use a cosine expansion for u = u_j(x)cos(j*pi*y/L) (summed over j) as this section suggests to take care of the homogeneous neumann's b.c's?
Is it practicable in this case? Is there an easier way? Please let me know.
Thanks,
Arunima.