[tommynego95]

Hi all,

I'm currently working on my final thesis job that involves a flow through a motorbike radiator.
I'm simulating a 2D section of the motorbike in exam via pimpleFoam solver but i encountered a problem.
Looking at the residuals they are actually pretty high but constant, especially the pressure one, which i'm scared of since it could compromise the physics accuracy of the solutions.
I put a probe in the geometry wake zone to control the pressure and the velocity and those values seems to be converged too since they do not oscillate neither vary in time.
As schemes, i started the simulation using
ddtscheme: Crank-Nicholson 0.5
grad scheme: cellLimited Gauss linear 1.0 for all the variables
div scheme: upwind discretization

and then, when the oscillation stopped i transitioned to
ddt: Crank-Nicolson 1.0
grad scheme: cellLimited Gauss lienar 0.5
div scheme: Gauss limitedLinear 1

that's to improve my final solution with a 2nd order discretization.
I'd like to specify that i defined also all the relaxation factors to 0.1.
As mesh, i've used snappyHexMesh that i then extruded via extrudeMesh to a 2D mesh to force a 2D simulation.
Also, the mesh present a non-orthogonality of about 46 so it should be pretty good.

Boundary conditions are pretty standard.
I' ve an inlet on the far left with a velocity Ux of 1m/s
Outlet is the far-right face with a fixed pressure of 0
XZ plane at y=0 is a symmetry plane
XZ plane at the top of the y-domain the same as the outlet
All the faces on XY plane are of "empty" type

Can anyone help me on lowering my residuals, especially p, or have any hint i can apply at my case?

[tommynego95]

Anyone can help me please?