[LauraSumner]

Hi everyone

I have encountered a very strange problem.

I am simulating a 2D rectangular box, length L, with an oscillating wall running from 0 to L-dL located halfway up the channel. The gap at the end of the channel allows exchange of fluid between the top and bottom channel.

I move this wall on every timestep such that it oscillates according to some predetermined function.

I also have two small moving windows, which should conserve the volume, needed due to the difference in area between the top and bottom channel as the wave oscillates (its not a symmetric wave!)

I implement this middle wall through defining it as a baffle, where the slave and master patch apply to the top and bottom faces.

I have realised that at the tip of the wall, in the exchange gap, the field is oscillating, and the entire U(x) in the top and bottom channels is also moving. The velocity is significant, although its time average is zero.
I have made sure that the end of my middle, osciallting wall has a blunt edge, as i thought maybe the sharp edge was causing the problem. This field now remains.
I am sure my calculations for the window volume conservation are correct.

Does anyone know why this should happen? What other numerical issues can cause a field to appear at a different part of the domain?

My Courant number is very low and so I am sure my timestep and mesh sizes are ok...

Any insight would be very valuable!

Many Thanks in advance!

L