[OliverB]

Hello,


I spent some weeks on a grid sensitivity study for a baffled stirred tank (MRF, simpleFoam). The problem was that the results were not monotonic.
I was using mean velocity magnitude in the whole tank as monitored value of interest.
After studying some publications I performed the study on axial velocity on 10 points located on a radial plane downstream the stirrer and plot it. The results were monotonic as expected. Good agreement was found with points of high velocity (1% error) and a bit less with points of low velocity ( up to 10% error).


I'm still a beginner so my questions :
Why error is more important on low velocities ?

How to evaluate conformity of test with points with low and bigger errors ?


Thanks