in/out boundary conditions

walterd

I know that In/out boundary conditions are not available. Are they going to be included in the next vsn ? If not, does anybody have any suggestion about implementing these features ? How to proceed , ideas , in short anything that might be helpful for someone willing to undertake this task.

SLongshaw

Hi,

I believe the next release of DualSPHysics (3.0) is likely to include such periodic boundary conditions and past cases have shown that the software framework is certainly capable of supporting this.

In the meantime, I suggest you have a quick Google for topics related to SPH periodic boundary conditions as it is a topic quite a few have considered in the past, it should really be a case of defining a new type of boundary particle and ensuring its interaction with fluid particles does what you need.

Apologies I can't be specific s this not really my area of expertise within DualSPHysics, hopefully one of the guys who has past experience of this (and is in the process of implementing it) will see this post too and come back with something more for you!

walterd

Thank You for your reply, maybe I wasn't clear in my post. I was looking into inflow / outflow boundary conditions. I have experience in software development in Eulerian frameworks, here it's lagrangian so I guess the mass conservation should be implemented in a way that if a particle is going out of the domain, another should come in. Also it should be possible to mark the inlet/outlet surfaces ad find a way to understand when a particle is crossing an outlet boundary, of general shape, thus some kind of intersection particle/triangulated surface should be implemented. That is what I can forsee in extreme summary. Maybe the people that are already working on this feature can tell me more, Please ?
Thank You

SLongshaw

Hello, sorry yes perhaps I wasn't too clear either.

What you say is absolutely correct, one particle out, one in is what has been implemented before (and I suspect will be what gets released with the next DualSPHysics) as this ensures mass is conserved.

In terms of how the boundary will be defined, technically it will likely be a special case of boundary particle, however the current design framework allows for different sets of particles to be defined using very complex geometries, this is the beauty of a particle based method, rather than needing specialist collision detection for various shapes you can simply build the shapes out of particles and the geometry is intrinsically handled.

The method has definitely been used successfully before, I believe it just needed tidying up in terms of how it is able to be used (i.e. how you are able to specify which particles are defined as inlet and which as outlet).

If you can't wait for the next release then implementing it yourself within DualSPHysics 2.0 will be challenging but certainly possible.

mashy

If you mean open inflow/outflow boundary conditions I would suggest looking at a paper by Federico et al. 2012 "Simulating 2D open-channel flows through an SPH model".

I would assume however implementing something like this into DualSPHysics would not be a small undertaking as the boundary condition method is drastically different from the Dynamic Particle method currently available, though with the very modular framework of the DualSPHysics code it should definitely be possible.