[lorenzowz]

Hello, my name is Lorenzo and I'm a PhD student from Sapienza University working on shock-fitting.
Consider the following configuration: a 2D simple duct with a moving shock-wave (left to right). It is a representation of the shock generation by an "impulsive" moving piston which is slowed down at a certain time. The problem in non-stationary. The computational domain is a rectangle with closed (eulerian) walls (up, down, right) while the left one is the piston surface. I'm simulating the piston's motion as the grid deformation (keeping the same topology) using an ALE (arbitrary langragian eulerian) approach. Note that the whole grid is going to deform even if at different rate depending on its proximity to the piston. Is that possible to simulate this configuration with the unsteady_dynamic_mesh or unsteady_static_mesh in SU2?

Regards,
Lorenzo

[economon]

Dear Lorenzo,

This type of simulation is indeed possible in SU2. You can compute this scenario as an unsteady flow on a dynamic mesh, as you mention, but some care will need to be taken when setting the velocities at the no-slip walls.

The piston on the left will be moving in the physical problem, so the fluid velocity on the piston surface should be set to the piston velocity (grid velocity). However, the upper and lower boundaries will be shrinking/expanding due to the mesh deformation, while in the physical problem, they are fixed as the piston slides past them. Therefore, the fluid velocity should still be set to zero on those walls rather than the grid velocity.

Hope this helps get you started,
Tom