Re: [ESPResSo-users] Fluid Dynamics & Boundaries.

[Rohit Goswami]

Thanks for the suggestion..
Is there any loss of physical data using this method as opposed to using a more traditional cfd package? (openfoam, Elmer etc.)
I thought of espresso because of the core correlator..
Also, how would the units be set here without particles for the lennard Jones sigma and other parameters?

On Fri, Jun 26, 2015 at 5:58 PM, Georg Rempfer <address@hidden> wrote:

You can simply build your geometry by combining (a possibly large number) of the primitives we have implemented. Esprecially if you run a pure LB fluid without particles, there won't be a speed penalty, since the boundary flag field is only evaluated once upon initialization of the boundaries.

If that doesn't suffice, you will have to implement a new constraint (work along the lines of one of the simpler constraints, such as sphere or wall), and write respective lbboundary wrapper functions. The files are named constraint* and lbboundary*.

On Mon, Jun 22, 2015 at 11:53 AM, Rohit Goswami <address@hidden> wrote:

The lattice boltzmann wall constraints currently don't have much geometrical shapes to allow for a flow cell construction; how are these extended (as they have been to produce the stomatocyte and other complex geometries)?
Is there any documentation on adding complex wall constraints?
I know Espresso is mostly used for soft matter, but I hope to use the lbfluid and related constraints for a CFD analysis of a flow cell, so no particles; just hard boundaries need to be defined.

Thanks.