Point initialization in a CFD run

[jforthofer]

This might be possible using the method in Patankar p.145. It is similar to the way WindNinja does it, as I remember. It's the "bombing the diagonal" type technique, except in Patankar it uses the volumetric source terms, but ends in the same result. Might be easy to do in OpenFOAM? Not sure what to do with the normal boundary conditions...? Set them somehow like normal, or all to pressure type boundaries?

[nwagenbrenner]

We also recently came across this paper which describes an approach we could try. It describes a point-initialization type run with a CFD model. Just like us, they are trying to generate a wind field using a CFD model using multiple point measurements (trying to match the measurements). It looks like they do some different kinds of spatial interpolations and using the idea of correlating the input speed (setting inlet boundary condition) to each simulated wind at the point measurement locations, combining things in some type of least-squares way.

Radunz_et_al_2020.pdf