Main Page | Report this Page
 
Loading..
   
Science Forum Index  »  Math - Numerical Analysis Forum  »  ? numerical dispersion
Page 1 of 1    
Author Message
Cheng Cosine
Posted: Fri Dec 08, 2006 12:07 am
Guest
Hi:

Using FD to approximate wave eqns, inapporpriate FD scheme could lead

to numerical dispersion. Will FEM have similar problem?

When FD is used, one can check its modified eqn to see if there is numerical

dispersion; how does one do something similar for FEM?

Thanks,
by Cheng Cosine
Dec/07/2k6 NC
Back to top
Guest
Posted: Sat Dec 09, 2006 1:20 am
Cheng Cosine wrote:
Quote:
Hi:

Using FD to approximate wave eqns, inapporpriate FD scheme could lead

to numerical dispersion. Will FEM have similar problem?

When FD is used, one can check its modified eqn to see if there is numerical

dispersion; how does one do something similar for FEM?

Thanks,
by Cheng Cosine
Dec/07/2k6 NC

There is no difference. See Sec 32.2 in

http://caswww.colorado.edu/courses.d/IFEM.d/IFEM.Ch32.d/IFEM.Ch32.pdf
Back to top
Han de Bruijn
Posted: Mon Dec 11, 2006 4:48 am
Guest
carlos@colorado.edu wrote:

Quote:
Cheng Cosine wrote:

Using FD to approximate wave eqns, inapporpriate FD scheme could lead

to numerical dispersion. Will FEM have similar problem?

When FD is used, one can check its modified eqn to see if there is numerical

dispersion; how does one do something similar for FEM?

There is no difference. See Sec 32.2 in

http://caswww.colorado.edu/courses.d/IFEM.d/IFEM.Ch32.d/IFEM.Ch32.pdf

Indeed. There is no difference. Because, basically, the two methods are
equivalent.

Han de Bruijn
Back to top
 
Page 1 of 1       All times are GMT - 5 Hours
The time now is Fri Nov 21, 2008 1:23 am
Science Forum Index

Powered by phpBB & ML | Disclaimer



Contact Us