Hello all,

I am trying to model (in 2D) a very simple geometry of a solid piston traveling down a cylinder.

The 2D model is the cross section of the piston/cylinder arrangement taken along it's axial length. The problem is modeled in Gambit as follows. It is actually split up into three sections which proceed horizantally from left to right as follows:

The first third has a pressure inlet (High pressure 18 bar) as the left hand boundary followed by a rectangular grid for the fluid region following the slug.

The right hand boundary for this section (lower pressure 1 bar) is a wall which is actually the left hand (trailing) edge of the moving slug.

The middle third is actually a space which represents the body of the piston. The third section has a wall for it's left hand boundary which represents the right hand edge of the piston then a rectangular mesh for the fluid ahead of the piston, the right hand boundary of this section is a pressure outlet. I have a udf which compiles and will move the piston at a variable speed.

I'm using a inviscid, unsteady implicit coupled model and I could not lunch my computation because Fluent do not accept this configuration.

Does anyone have any suggestions on what I might be missing? Thanks so much!

Alain,

Sounds like a simple problem. Dynamic mesh motion does not depend on your viscosity model or your solver. Are you using layering technique for remeshing? I would check if it is set up right. If you get negative volumes then something is not moving as it should. Does your piston movement depend on solution variables (pressure perhaps)? If not you can use a profile function instead of an udf

Dear Vivek,

Thank you for your message.

Yes I'm using a layering technique for remeshing and I build a UDF enable to calculate the piston velocity using Newton law's (the piston movement depend on solution variablesressure). Could you tell me what bondary conditions I should put in wall cylinder and in piston wall?

Thank you