Hi there

I have managed (eventually) to produce a working 2D engine simulation (just as an excercise to figure out how to do the moving meshes and valve events).

I'm trying to add a bowl cutout to the piston, but am having layering remeshing problems at the moment. FLUENT doesn't seem to like having a meshed region (quad or tri) under the dynamic layering region.

I am still using the built in piston movement functions (**piston-full** & **piston-limit**)in FLUENT 6.1.

I defined the edge between the moving bowl grid and the layer zone as an "internal boundary" and defined its movement using the **piston-full** UDF. The flat part of the piston and the bowl spline edge are defined as a wall and also use the **piston-full** UDF. The movement of the edge between the remeshing and layered zones is defined using the **piston-limit** UDF as usual. Done this way the mesh preview crashes at the first layering event, and if I use a quad mesh for the bowl and no "internal boundary", the mesh in the bowl doesn't move and thus crashes into the moving edge above it!

Has anyone had experience in this area?

Gareth

I did almost exactly the same things as you did. In the case of pistons with bowl/cavity, i think that we cannot use quadrilateral cells and use dynamic layering. I simply generated unstructured grid and it worked. I also used the piston - full udf for the piston.

Hi AMV Thanks for your response. I'll definately try that. How did the computational expense and accuracy compare when doing only dynamic remeshing as opposed to dynamic layering? Gareth

as i told you, we cannot use dynamic layering in the axi-symmetric case of pistons with bowl. we can definitely do it for flat pistons but not with bowls. So I directly generated un-str grid and used remeshing technique without any dynamic layering. We cannot use dynamic layering with un-stru grids as you know.