Pressure Gradient at Velocity Nodes

[diksanm7]

As mentioned in the paper, xcompact3d employ zero gradient pressure at direction when the velocity is Dirichlet condition. The subroutines are in derive.f90, e.g., deryvp subroutine, I want to employ p = 0 condition or maybe fixed gradient on y direction,

I just want to confirm that I "just" need to modify the expression in deryvp, and also its constant cfi6y, csi6y, and cwi6y.

Regards,
Diksan

[mathrack]

Hello,

The code assumes consistency between the divergence and the gradient operators. Thus, changing one should impact the other.

In addition to the staggered derivation / interpolation operators, you would have to update the Poisson solver as well : the current solver assumes a zero derivative at the boundary and performs a cosine transform.

Therefore, switching to non-zero derivative at the wall (or to an imposed value) for the pressure is a huge amount of development / validation.

Best regards,
Mathrack