Find answers, ask questions, and connect with our <br>community around the world.

Home Forums OpenFOAM Forum Why Pressure iterates much faster than other fields?

  • Why Pressure iterates much faster than other fields?

    Posted by María Manuela Rosales on September 9, 2023 at 8:13 am

    Good day dear Prof. Baris and collegues,

    I’ve done some steady simulations of exteral wind flow and all the simulations and when I monitor the residuals, I see that pressure-residual-curves runs like twice faster (# of iterations) than the rest of the fields….

    I have also increased from 0 to 2 the nonOrthogonalCorrectors in SIMPLE algorithm, looking for improving accuracy. Is this the reason why pressure is being solved with more iterations than other fields?

    Thank you,

    María Manuela Rosales replied 9 months, 1 week ago 2 Members · 2 Replies
  • 2 Replies
  • Barış Bicer

    Moderator
    September 12, 2023 at 12:00 pm

    Hi Maria,

    Firstn, I dont know what kind of solver setup did you use in your case? Maybe you assigned a higher tolerance where pressure can converge faster.

    I have also one question related to nonOrtg.Correctors. Did you already check the related lecture about this? Why did you increase this correctors? Has your mesh higher non-orthogonality?

    Let me know with your setup.

    • María Manuela Rosales

      Member
      September 12, 2023 at 12:40 pm

      Good day, thanks for the prompt reply,

      Time ago I found a recommendation when running algorithm SIMPLEC (when we set ‘true’ the parameter ‘consistent’) it was required at least 1 value for nonOrthogonalCorrectos, since it helped the case to achieve faster convergence and gain acurracy as it added more iterations on pressure.

      My case is external incompressible flow and begins with an empty domain, practically all orthogonal, then it passes to have included an obstacle (extracted volume using snappyHexMesh) and in that case, the mesh gets high values of non-orthogonality and skewness. I set up both cases thinking on the worst scenario (with obstacle) …. So my new questions would be:

      1) is it not always necessary to use nonOrthogonalCorrections when using SIMPLEC?

      2) Just to confirm, as we increased from 0 to n value the non-OrthogonalCorrectors, is it expected that pressure-residuals goes 2*n faster?

      3) I expect that my simulation stops when the residuals reaches all the ‘residualControl’ values for all those fields…. could you confirm this set-up for stop simulation is well assumed? and simulation will not stop following the tolerances defined in linear solvers… just to check I’m structure the file as I was looking for…

      Thank you!

Log in to reply.

error: Content is protected !!