Why there is no filtering option for DNS case

[Madhurjya814]

Hi, in the line below

Incompact3d/src/xcompact3d.f90

Line 41 in abb010e

if ((itype.eq.itype_abl.or.iturbine.ne.0).and.(ifilter.ne.0).and.(ilesmod.ne.0)) then

...Why for DNS, no filtering is used?
Thanking you.

[slaizet]

Because we do not need filtering in a DNS as all the scales are resolved.

[Madhurjya814]

I used first and second derivatives as 6th order compact scheme and ilesmod=0 (DNS). And the simulations did not diverge and I could successfully validate. Is it ok to run DNS along with both derivatives as 6th order compact scheme with AB3 as time integration scheme?
Thanking you.

[slaizet]

Yes that's absolutely fine and a good choice for a DNS.

[Madhurjya814]

I wanted to ask one question. For pipe flow in this study https://www.sciencedirect.com/science/article/pii/S0142727X24003655 it has been mentioned as Direct Numerical Simulation. Due to uniform grid points the deltar+ was 1.8 in radial direction. Is it ok to report as DNS if I can do simulation with the same resolution they have used in that paper?

[slaizet]

yes absolutely

[Madhurjya814]

Thank you Sir.