[lloco]

Hi;

I modeled inviscid incompressible external flow around a cylinder with symmetric BC but the velocity or pressure field are not symmetric despite of good convergence. Any idea is appriciated.

LL

[stuart23]

What did you do with the viscous terms? An inviscid (Stokes Flow) solution should be symmetric as Re = 0.

Also, define "good" convergence. Floating point error may also make solutions look different to what you might have expected, but not by much.

[lloco]

Thanks ! I used the inviscid option of Fluent. The simulation converged for residuals as small as 1.e-7 but the solution was considerably asymmetric (?) with some recirculation behind the sphere.

Later I changed the pressure discritization from Standard to PRESTO and got much better results. Still not quite symmetric but much better than standard discritization. You may take a look.

Btw, I am solving the Eulair equation (viscosity=0) rather than Stokes flow(viscosity=infinity). In the Eulair equation, the pressure and velocity components all must be symmetric (v is anti-symmetric) at the both side of the cylinder. The Same for velocity components in Stokes flow. But pressure contours must be asymmetric in the Stokes flow !