No convergence on simple Fluent simulation (2D Axisymmetrical flow around a Sphere)

- May 23, 2021 at 11:10 pm
  
  Touf_CFD
  Subscriber
  
  Hello ! Recently I've been trying to simulate a flow around a sphere in order to calculate all the aerodynamic coefficients. But my simulation isn't converging at all because my residuals aren't decreasing (picture below), and I can't seem to find the problem, so I was wondering if anyone can help me out.
- May 24, 2021 at 8:21 am
  
  Rob
  Forum Moderator
  
  For Mach 0.9 I'd use the pressure based solver as it's much easier to converge. How does the velocity field look?
- May 24, 2021 at 12:16 pm
  
  Touf_CFD
  Subscriber
  
  Thank you so much for answering my question, using the density based solver the velocity field doesn't look like much, so this morning I took your advice and ran a quick simulation using the pressure based solver the convergence is better but I don't believe it is sufficient , you will find below the screenshots of the residuals and the velocity field.
- May 24, 2021 at 1:23 pm
  
  Touf_CFD
  Subscriber
  
  Another issue I forgot to mention is that the next step is to start supersonic simulations, is it possible to do so using the pressure based solver ? Thank you
- May 24, 2021 at 1:34 pm
  
  Rob
  Forum Moderator
  
  That looks supersonic as it is: check as speed of sound in air is about 330m/s from memory. You will want a longer domain downstream to let the wake develop. In answer to the question, the pressure based solver is good to Mach 2-3 and using the pressure based coupled solver will go higher than that.
- May 24, 2021 at 5:46 pm
  
  Touf_CFD
  Subscriber
  
  Perfect, ok I will try from now on using only the pressure based solver. Thank you for your help !
- May 25, 2021 at 7:06 am
  
  aitor.amatriain
  Subscriber
  
  It seems strange that the problem is axisymmetric. You have a flow in the negative radial direction? If not, the problem is not axisymmetric
- May 25, 2021 at 9:48 am
  
  Rob
  Forum Moderator
  
  Looks OK, for axi-symmetric the axis is on y=0 The only bit that'll be missed is if the wake is unsteady as the oscillations won't be picked up as that's a 3d phenomena.
- May 25, 2021 at 2:07 pm
  
  aitor.amatriain
  Subscriber
  
  Velocity contour looks okay, but that is not an axisymmetric problem. It is a symmetric problem.
  I also agree with the comment about the 3D model. In this particular problem 2D approach is reasonable for Re<10, specially taking into account the presence of shock waves.
  In any case, I would increase the outer radius of the spherical shell even in the 2D model.