|
[Sponsors] |
December 18, 2006, 00:31 |
strongly implicit method
|
#1 |
Guest
Posts: n/a
|
sir,
I don't understand the SIP (http://en.wikipedia.org/wiki/Stone_method). what does N represent Ax = (M-N)x = (LU-N)x = b in the above equation A and b are known values , what is the need of iteration here. |
|
December 18, 2006, 00:34 |
Re: strongly implicit method
|
#2 |
Guest
Posts: n/a
|
A is known but finding Ainv is very difficult, so we prefer iterating. (to know x you need to know Ainv).
|
|
December 18, 2006, 02:33 |
Re: strongly implicit method
|
#3 |
Guest
Posts: n/a
|
could you explain breif about this algorithm.
|
|
December 18, 2006, 02:48 |
Re: strongly implicit method
|
#4 |
Guest
Posts: n/a
|
on cfd-online, in the wiki there is a section on numerical methods, there you will find the link about stone's method. Read that.
|
|
December 18, 2006, 05:30 |
Re: strongly implicit method
|
#5 |
Guest
Posts: n/a
|
A contains 5 diagonals (Poisson equation discretisation result via the 5 points scheme).
If we try to performe the incomplete LU decomposition, L and U will contain 3 diagonals each, but the multiplication will produce M with 7 diagonals, so N must contain at minimum 2 diagonals of M that correspond to zero diagonals of A. So M = L*U and M-N = A |
|
December 19, 2006, 00:50 |
Re: strongly implicit method
|
#6 |
Guest
Posts: n/a
|
yes i got 7 diagonals. with L has 3 diagonals, U 2 diagonals the additional two diagonals are stored in N matrix. Could you please expalin how to proceed after this. all are sparse matrix, so how to solve this equation.
|
|
December 19, 2006, 07:26 |
Re: strongly implicit method
|
#7 |
Guest
Posts: n/a
|
The best advice I can give you is to look for this book "Computational methods for Fluid Dynamics", Ferziger & Peric, the SIP is clearly explained and you also can download book codes.
Email me if you have difficulties to find them. |
|
December 19, 2006, 07:56 |
Re: strongly implicit method
|
#8 |
Guest
Posts: n/a
|
sir, how to download codes from peric book... could you please give your mail id
|
|
December 19, 2006, 07:57 |
Re: strongly implicit method
|
#9 |
Guest
Posts: n/a
|
this is mail id d_santhosh2@yahoo.com
|
|
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Implicit iterative Gauss-Seidel method | dearboy | Main CFD Forum | 5 | November 29, 2010 10:18 |
need paper Strongly implicit method | santhosh | Main CFD Forum | 1 | December 13, 2006 13:52 |
Implicit method of FSI | cjtune | Siemens | 0 | March 12, 2003 13:07 |
implicit method on unstructured grid | Koo | Main CFD Forum | 5 | July 29, 2001 02:43 |
Modified Strongly Implicit Procedure of a nine-point finite-difference representation. | Zhong Lei | Main CFD Forum | 1 | September 18, 1998 18:54 |