Learning Forum

[Sahil Gupta]

Hello all,

I am trying to do a transient simulation of a scenario of a valve where in the flow comes in through the inlet, applies a pressure on the wall (marked in green) and in turn pushes the wall upwards and leaves through the outlet, as seen in figure below. 



The mass flow at the inlet is a function of time step (s). 

I am using dynamic mesh with 6DOF and meshing method is layering. When selecting 6DOF, I am using constraints i,e the wall can only move from 0 to 1,38mm in the normal direction (translational movement). 

I am using a spring constant of 5e6 N/m. When I dont use a preload, the poppet moves to its highest location described in the constraints (1,38mm) in the very first time step which constitutes to a very small mass flow rate. It seems as if the spring constant is not working at all and not opposing the motion of the poppet at all. 

When I use a preload of lets say -1N, as soon as the force in the solution goes beyond 1N, the fluent hangs up and doesnt calculate further. I dont get any error but it just doesnt go beyond that timestep. 

Can anyone help out with this or have any suggestions as to how can I fix this without a UDF for defining the motion of the poppet? 

Kind regards,
Sahil

[Rob]

Are you expecting the whole green wall to move upwards, or the wall to deform? I'd start by reviewing the mesh: cell quality is one metric, another is resolution. 

[Sahil Gupta]

Hello, I am expecting the whole wall to move upwards. The mesh is coarse at the moment but I doubt that wall suddenly moving up is due to mesh quality. 

[Rob]

And if you calculate the force on the wall, what is it? 

[Sahil Gupta]

As I mentioned, if I give a preload of -1N, as soon as the force crosses 1N, the simulation hangs or rather doesnt go ahead or give any error. If I dont give any preload, then the wall move to its highest position all of a sudden ending up in negative cell volume in the very first time step. 

[Rob]

Given the mesh and narrow region any motion is going to require remeshing. How large is the domain? 1.38mm could be small or very large relative to the cell size.