|Nov30-09, 01:48 AM||#1|
Galerkin method for wave equation
I want to use Galerkin method to solve 3-D wave equation \nabla^2 u+k^2 u=0, with the following boundary conditions: at z=z_1 plane, u=g, and when x,y,z go to the infinity, u becomes 0.
My question is how to choose the basis function \phi_n for u: u=\sum \lambda_n \phi_n. As my boudary condition is a little different from the usual setting discussed in many books, I am confused of selecting basis function.
|Nov30-09, 06:09 PM||#2|
As long as it forms a basis, I dont think it matters. So the basis functions youre used to should work here as well. I suggest using a weak form if you have Neumann boundaries.
|Nov30-09, 10:38 PM||#3|
I indeed want to solve it in the weak form. Actually, my problem is to solve wave equation (homogenuous equation) with non-homogenuous boundary conditions, maybe it is Dirichlet problem. Of course, it can be switched to non-homogeneous equation with homogenuous boundary condition.
The main difficulty I encountered is this problem is of 3-D problem, where weak form is difficult to solve: calculating gradient of 3-D function, and then calculating the inner product in the space defined are not so easy. So I need to choose simple basis function that can be easily calculated with respect to gradient and inner product.
Anyone has similar experience? Thanks a lot!
|Similar Threads for: Galerkin method for wave equation|
|Method of showing the wave nature of EMR?||General Physics||5|
|The 2D wave-equation: d'Alemberts method||Calculus & Beyond Homework||0|
|Wave on a rope - question concerns the maths of the wave equation||Introductory Physics Homework||3|
|wave equation and graphing wave instance||Advanced Physics Homework||0|
|Can someone tell me if i'm using the right method? More wave fun!||Introductory Physics Homework||9|