A simplified Kelvin Helmholtz instability was simulated. Dry air filled up the upper half of the domain, and humid air the lower half. A 4% difference of density was assumed between the two gas. 2 m/s difference in velocity across the interface was fixed. The inlet velocity of the dry air was perturbed with a sinusoidal signal in order to generate the instability.
Files to reproduce the case with Code_Saturne 2.0-rc1 :
usiniv-f90.gz (2.91 KB)
kelvin-med.gz (1.18 MB)
instabilities-xml.gz (1.88 KB)
Hello Jean,
This is a really nice example here!
However, you might have better results with symmetries farther from the middle (haven’t try though…). It seems you have some “blocking effect” due to the presence of the symmetries at bottom and top.
David
I redo the computation with a wider inlet. Indeed, it changes the flow a little. I redo also the set up without usini1.f90: everything was done by the GUI. I upload the new files below.
kelvin-xml.gz (1.93 KB)
kelvin-helmholtz-med.gz (3.56 MB)
