Hello,

If anyone has experience in using a partially staggered grid arrangement (velocities at grid nodes and pressure at center) for finite difference schemes, how do you discretize the convection term ? (incompressible flow)

Thank You.

you may refer to a paper: L. Huang, Numerical solution of the unsteady incompressible N-S equations on the curvilinear half-staggered mesh, Journal of Computational Mathematics, Vol.18, No.5, 2000, 521-540.

One scheme was worked out at Los Alamos 30 years ago. The first code was named YAQUI, the more general version was called SALE.

SALE was published in 2D and 3D versions. Check Journal of Computational Physics in the 70's and 80's. There'll be references there to the LANL reports.

These were precursors to the KIVA family of codes.

That's the grid configuration that Harlow and Welch used for the MAC method in 1965. There is a discreteization of it on the web in my thesis http://www.cc.gatech.edu/~carlson/pa...son-thesis.pdf

In case you don't have access to the J.of Comp. Phys.