Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Green's Function for Helmholtz Eqn in Cube

  1. Nov 28, 2011 #1
    1. The problem statement, all variables and given/known data
    Find the Green's Function for the Helmholtz Eqn in the cube 0≤x,y,z≤L by solving the equation:
    [itex]\nabla[/itex] 2 u+k 2 u=δ(x-x')
    with u=0 on the surface of the cube
    This is problem 9-4 in Mathews and Walker Mathematical Methods of Physics

    2. Relevant equations
    Sines, they have the properties we're looking for.

    3. The attempt at a solution
    So, if the solution is in one dimension we obviously have
    G=asinkx x<x' and
    G=bsink(x-L) for x>x'
    And these are obtained by solving [itex]\nabla2[/itex]u+k2u=0 and looking for a and b based on the matching condition at x'. I'm wondering if I can do the same thing for this three dimensional case, using separation of variables and solving for k^2 as the eigenvalue. Then enforcing orthogonality on the functions I get. My problem with this is just that it seems very awkward to have eight sets of equation for the Green's function, based on whether x<x' while y>y' and z<z' for example.

    Any help appreciated.
  2. jcsd
  3. Nov 29, 2011 #2
    I guess you could do that. You can always use Fourier transform to obtain the Green's function without BC and add a particular solution of the homogeneous Helmholtz equation that satisfies the BC instead, which I imagine would be a less awkward approach.
  4. Nov 29, 2011 #3
    So using the Fourier transform would just give you the green's function for helmholtz in 3d space which is something like an exponential divided by 4∏r followed by simply any function satisfying the BC? So in a 1-d case:
    Helmholtz Green's function + sinkx for x<x' or
    Helmholtz Green's function +sink(x-L) for x>x'

    Thanks for the help.
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook