Unformatted text preview:

Lecture 25: 8 November 2010Lecture 25: 8 November 2010 Discretization of the (1d scalar) wave equation, simplifying for now to an infinite domain (no boundaries) and constant coefficients (c=1). This corresponds to the equations ∂u/∂t=∂v/∂x and ∂v/∂t=∂u/∂x. The obvious strategy is to make everything a center difference. First concentrating on the spatial discretization, showed that this means that u and v should be discretized on different grids: for integers m, we should discretize u(mΔx)≈um and v([m+0.5]Δx)≈vm+0.5. That is, the u and v spatial grids are offset, or staggered, by Δx/2. For discretizing in time, one strategy is to discretize u and v at the same timesteps nΔt. Center-differencing then leads to a Crank-Nicolson scheme, which you analyze in homework and show to be unconditionally stable (albeit implicit) for anti-Hermitian spatial discretizations. Alternatively, we can ues an explicit leap-frog scheme in which u is discretized at times nΔt and v is discretized at times [n-0.5]Δt. Discretization of the (1d scalar) wave equation: staggered grids and leap-frog schemes. Von Neumann and CFL analysis. Dispersion relation. Further reading: Strang book, section 6.4 on the leapfrog scheme for the wave equation.MIT OpenCourseWarehttp://ocw.mit.edu 18.303 Linear Partial Differential Equations: Analysis and NumericsFall 2010 For information about citing these materials or our Terms of Use, visit:


View Full Document
Download Lecture Notes
Our administrator received your request to download this document. We will send you the file to your email shortly.
Loading Unlocking...
Login

Join to view Lecture Notes and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view Lecture Notes 2 2 and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?