Learning Forum

[Valeria Guadagno]

Hi!

I'm having a problem I am not able to solve. I'm performing a mesh convergence analysis, i.e. I'm performing the same simulation with same conditions on 4 different meshes from the finest one, the second-finest, the optimal mesh to the coarse mesh.

The simulation is pressure-based, transient, performed using 6 inlets set in inlet velocity, and 2 outlets initialized with pressure.

I plotted the results on a graph (here below) to check the errors between the different meshes, plotting the computed flow rate and the computed velocity vs Numer of elements in the mesh (i.e. type of mesh). However, it is clear that some error occurred, since flow rate and velocity at the same outlet should have the same trend, scaled of a factor ro*A (since Q=ro*v*A). However, it clearly isn't so. It's like if the ro*A is changing for each mesh (as computing Q/v gives different result for each mesh), which isabsurde since they are both fixed parameters.

Does someone have some idea of what the error may be? Why is it happening?

Thank you in advance

 

[Rob]

I'd start by reviewing the mesh, model settings and convergence. Then examine the results, ie contours, vectors etc. 

From the graphs the variation looks to be lower for the finer two meshes, but as I have no idea what changed in the mesh and what you're modelling I can't really comment. 

[Federico Alzamora Previtali]

Are you using an incompressible fluid? You might want to plot the density at this location and make sure you are getting constant rho.