CFD Online Logo CFD Online URL
www.cfd-online.com
[Sponsors]
Home > Forums > Main CFD Forum

SIP for SIMPLE code

Register Blogs Members List Search Today's Posts Mark Forums Read

Reply
 
LinkBack Thread Tools Display Modes
Old   June 5, 2002, 15:15
Default SIP for SIMPLE code
  #1
Li Yang
Guest
 
Posts: n/a
Dear all,

My current 3d unsteady pressure correction SIMPLE code is using TDMA solver. Please could you let me know whether the SIP or ICCG solver would be able to speed up my calculation ? When comparing with the TDMA, how fast they would be ?

Thanks in advance.

Li

  Reply With Quote

Old   June 5, 2002, 23:54
Default Re: SIP for SIMPLE code
  #2
mukkarum
Guest
 
Posts: n/a
r u generating code on metlab

Sir i am working on research project "revemping of Francis Turbine applying CFD". I have got book "An introduction to computational fluid dynamics (finite volume) by malalasakra" it is very good book and i have understand SIMPLE, SIMPLER AND SIMPLEC method of solving Navier Stock equation but in this book just presentation that how Navier Stock equation can be solved but there is no solve example. Do you know any book or article in which this type of example is present. I mean that for any given problem(good if turbine rotor)firstly assume pressure field and then find velocity field then check pressure correction equation and repeat this procedure still convergence occure. waiting for your prompt reply thanks
  Reply With Quote

Old   June 6, 2002, 05:27
Default Re: SIP for SIMPLE code
  #3
Li Yang
Guest
 
Posts: n/a
Dear Mukkarum,

You can refer to Ferziger and Peric's book: "Computational Methods for Fluid Dynamics". Some codes introduced by this book can be downloaded from internet. If you can not get a source code from an expert, those codes could be a very good starting point.

Regards

Li

  Reply With Quote

Old   June 6, 2002, 23:20
Default Re: SIP for SIMPLE code
  #4
mukkarum
Guest
 
Posts: n/a
Assalam Alaykum wa Rahmatullah Dear Sir I am starting to study ferziger and peric's book and contact you soon . I have study this book some days before but find malalasakra's book is good than this book. I will ask question to you when i will feel difficulty,could you help me.I tried to email you this massage but failed because there is some problem in your email box. Wasalam, MUKKARUM HUSSAIN RESEARCH ASSISTANT
  Reply With Quote

Old   June 8, 2002, 12:09
Default Re: SIP for SIMPLE code
  #5
Andrea Panizza
Guest
 
Posts: n/a
Dear sir, a valuable reference for your case is Ferziger-Peric's "Computational Methods for Fluid Dynamics", where SIP, ADI, ICCG, as well as multigrid methods, are compared against different versions of line solvers using TDMA, for elliptic and mixed type partial differential equations. Personally, I suggest you to use Schneider and Zedan's MSI instead than SIP, since the programming effort is similar, but time to convergence is in general quite smaller in 2D and slightly smaller in 3D. For what it concerns ICCG, I don't think you may apply it to the pressure correction operator in SIMPLE, since it is neither symmetric or positive definite, even if it may be close to both when the velocity field is smooth enough. You'd better go for BiCGSTAB or GMRES, instead.

Greetings, Andrea
  Reply With Quote

Old   June 9, 2002, 10:23
Default Re: SIP for SIMPLE code
  #6
Li Yang
Guest
 
Posts: n/a
Dear Mukkarum,

I think the best way to learn CFD is to find an expert nearby, for example a professor or a lecturer in CFD in your city.

Regards

Li
  Reply With Quote

Old   June 9, 2002, 10:37
Default Re: SIP for SIMPLE code
  #7
Li Yang
Guest
 
Posts: n/a
Dear andrea,

Thank you very much for your reply. I know some people have applied CG, Block ILU or GMRES to 3d compressible solver. However, for incompressible flow, I am relatively a new comer. I would be very interested in some reference papers, if you could provide any, about MSI, CG or GMRES which are applied to SIMPLE method based solver. Is your code using MSI or GMRES ? Do you need preconditioner when using GMRES ?

Thank you again.

Li
  Reply With Quote

Old   June 11, 2002, 14:09
Default Re: SIP for SIMPLE code
  #8
Andrea Panizza
Guest
 
Posts: n/a
Dear Li, 1)an introduction to MSI, which provides sufficient detail for the coding of a subroutine, is given in "Computational fluid mechanics and heat transfer" by J.C. Tannehill, D. A. Andreson and R.H. Pletcher. As for reference papers, here they are the ones I know of: -Stone, H. L., (1968), "Iterative Solution of Implicit Approximations of Multidimensional Partial Equations", SIAM Journal of Numerical Analysis, vol. 5, pp.530-558 (the original paper by Stone on his Strongly Implicit Procedure); - Zedan, M., and Schneider, G. E., (1985), " A coupled strongly implicit procedure for velocity and pressure calculations in fluid flow problems", Numerical Heat Transfer, vol.8, pp.537-557 (the original paper on MSI procedure, an improvement of SIP); - Chen, K.-H., and Pletcher, R. H., (1991), "Primitive variable, strongly implicit calculation procedure for viscous flows at all speeds", AIAA Journal, vol. 29, pp. 1241-1249 (a paper about a block version of SIP); - Chen, K.-H., and Pletcher, R. H.,(1993), "Simulation of three-dimensional liquid sloshing flows usinga a strongly implicit calculation procedure", AIAA Journal, vol. 31, pp. 901-910 (application of block SIP to 3D incompressible free-surface flows);

2)As I told you in my precedent message, CG may not be applied to the inversion of the pressure correction operator B(n) in SIMPLE, since B(n) is not symmetric positive definite. You may do a Gauss transform , e.g. apply CG to B'(n)*B(n) where B'(n) is the transpose of B(n), but this would square the condition number, with a correspondent large increase in computational times. Anyway, if you should find any reference about successful applications of CG to the inversion of B(n), you would do me a great favour by telling me. GMRES, CGS, BiCGSTAB may all be used, instead: in the "Handbook of computational fluid dynamics" edited by R. Peyret, pp.53-65 you will find a comprehensive list of reference papers, along with a brief comparison of different methods. For direct coding of such methods, I point you to http://www.netlib.org/linalg/html_te...Templates.html, where you will find templates for iterative methods, by Dongarra et al.

3)Preconditioning is often required for Krylov subspace methods, e.g. GMRES, CG, BiCGSTAB and the like, but unfortunately I only found efficent preconditioners for Stokes flow, which don't work for Navier-Stokes. Howewer, I heard that Howard C. Elman at the University of Maryland studied the problem.

4)As for my still-a-long-way-from-completion SIMPLE code, I plan to use V-cycle multigrid with point Gauss-Seidel or zebra Gauss-Seidel as a smoother, since I found it to be very effective for the numerical solution of general elliptic equations. MSI will be added in a second moment as an optional smoother for curvilinear grids, since it is way more robust than the formers when using non cartesian grids.

Greetings, Andrea

P.S.: have you ever encountered a strange behaviour of the residual error, like high accuracy in some zones and much lower accuracy in others, when using TDMA for a simple problem like Laplace equation on a square with Dirichlet b.c.?
  Reply With Quote

Old   June 12, 2002, 05:49
Default Re: SIP for SIMPLE code
  #9
Li Yang
Guest
 
Posts: n/a
Dear Andrea,

Thank you very much indeed. The information you have provided is very helpful.

Regards

Li
  Reply With Quote

Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are On


Similar Threads
Thread Thread Starter Forum Replies Last Post
The FOAM Documentation Project - SHUT-DOWN holger_marschall OpenFOAM 242 March 7, 2013 13:30
Debugging Unsteady 2-D Panel Method Code: Wake Modeling RajeshAero Main CFD Forum 5 November 10, 2011 06:48
Small 3-D code Zdravko Stojanovic Main CFD Forum 2 July 19, 2010 10:11
Design Integration with CFD? John C. Chien Main CFD Forum 19 May 17, 2001 15:56
own Code vs. commercial code Bernhard Mueck Main CFD Forum 10 February 16, 2000 11:07


All times are GMT -4. The time now is 21:54.