Valve action

Hrvoje Jasak (Hjasak)

Hi All,

I have been asked to provide an update on valve action work I have been doing over the last year - it seems some people are keen to use it and it isn't quite finished yet. If anyone is interested to work on this, here's the state of the work:

- the basic building block is the topological change engine. This is done, well tested and robust.
- on top of this, there is a set of "mesh modifiers" - they implement the topological changes you need in terms of primitive changes. Currently available are layer addition/removal, attach/detach boundary and sliding interface. All three are tested and robust
- the information on valve motion and similar needs to be pulled into the code. For this purpose, I have added the engive valve class, defining various patches and lift curves etc and I'm quite happy with it. the information, of course, needs to be provided in the foam format, so straight conversion from kiva or star has not been considered or implemented
- topological stuff needs moving mesh, which is also implemented and robust. to my knowledge, the solvers are already in a decent shape with mesh motion and even Lagrangian fracking works on moving meshes + in parallel (?). Mesh motion in parallel is OK, but topological changes are not.
- definition of valve action now reduces to specifying appropriate mesh modifiers, including sliding interfaces and layering. This can be done in a number of ways and I have based my further work on a Kiva example. This will be varied, depending on the type of mesh and action you want and is probably quite case-dependent.
- I have got a class for Kiva meshes which automatically locates all mesh modifiers and runs well for my sample 2-D geometries. for 3-D with inclined valves, I had to add the deformation of the valve curtain + the bil below the valve.

The current state of affairs is that the combination of the motion and topo changes leaves a bit of mess in the mesh and I need to clean this up - this should be oth quick and easy, but currently there's a few things that are much more important.

Finally, you will need some solvers which can deal with topological changes. I have built in the data mapping and that's good and robust, but the only flow solver that actually works is icoTopoFoam. I have no intention of updating other solvers before I really need them. When I get some time together, I will carry on working on my Kiva meshes and show some results on the web site in due course.

So, there you are: all the tools are there and ready to be used but I haven't finished the packing yet.

Hrv