[ripperjack]

Hi all,

I am coding a finite-difference CFD solver and I am using staggered grid. My question is: how do we calculate <u'v'> for post processing?

The reason I ask is because, for the staggered grid configuration, u and v are stored at the east and north face (not at a same point). It will be problematic if we simply do <u(i,j,k)*v(i,j,k)>. My solution now is to interpolate u and v to the cell center, and then do <u(i,j,k)*v(i,j,k)>. I am wandering if we can have a better solution. Thanks in advance.

Best regards,

[FMDenaro]

Originally Posted by ripperjack

Hi all,

I am coding a finite-difference CFD solver and I am using staggered grid. My question is: how do we calculate <u'v'> for post processing?

The reason I ask is because, for the staggered grid configuration, u and v are stored at the east and north face (not at a same point). It will be problematic if we simply do <u(i,j,k)*v(i,j,k)>. My solution now is to interpolate u and v to the cell center, and then do <u(i,j,k)*v(i,j,k)>. I am wandering if we can have a better solution. Thanks in advance.

Best regards,

yes you wrote, you need to interpolate... if you have a second order code, a linear interpolation is congruent to the local truncation error of the FD scheme