Mesh sensitivity study for large setups
 

Hello,

I am doing a large centrifugal pump simulation with a suction chamber, impeller (complete) and volute. I also wanted to carry out a mesh sensitivity analysis.
But, I was wondering, if it is acceptable to use the pervious result (for example the result of the course mesh) as initial conditions for the "medium mesh".
Or, should I start every simulation from zero, which would result in a larger computational time.
Can anyone explain if this would make a big difference in the numerical results ?

Thank you and kind regards
Daniel

For a steady simulation (with a unique solution), the converged solution is independent of the initial conditions; therefore, you should be able to start from any known previous reasonable results. This is standard practice when dealing with complex physics.

Be certain the interpolator step is included such it can interpolate from the previous to the new mesh.

If the solution is reasonably close (converged), the computational time should decrease though that is not always the case since with mesh refinement other flow features missing in the initial values must also be converged.

Perfect, thank you for your response and explanation.
Thats a good point, to activate the interpolator for the mesh - didn't thought about it.