|
[Sponsors] |
January 31, 2004, 15:45 |
Error Analysis
|
#1 |
Guest
Posts: n/a
|
Hello CFD friends
I am working on an error analysis of a time marching scheme.Can anybody help me to find out how to calculate error for every time step and then the for a finite no. of time steps.By the way I am dealing with adams Bashforths time stepping method. Thank You in advance. |
|
January 31, 2004, 22:53 |
Re: Error Analysis
|
#2 |
Guest
Posts: n/a
|
Are you integrating odes or pdes? If they are odes, look up a book on odes, and read the chapter on multi-step and BDF methods. Analysing consistency is easy (and well known for Adams Bashforth) but stability (and types of stability like R-stability, alpha-stabililty) is a little more complicated than one-step methods. The best book I have come across is by E Hairer, SP Norsett and G Wanner, called "Solving ordinary differential equations".
If you are dealing with pdes, there are similar methods of analysis, but often local truncation errors scale as some non-linear product of spatial grid size and the local time step. J. W. Thomas's book called "Numerical Partial Differential Equations: Finite Difference Methods" should give you some good help. |
|
February 1, 2004, 12:51 |
Re: Error Analysis
|
#3 |
Guest
Posts: n/a
|
Thank u 'whatever' for the information I will go ahead and find these books
|
|
Thread Tools | Search this Thread |
Display Modes | |
|
|