What could lead to the loss of symmetry in the simulation of the 2D shock-bubble interaction test case with the compressible Euler system discretized within a FV framework by an implicit time scheme?
The answer I could think of: It's the error accumulation in the solution of the block matrix by any linear solver (block matrix is manifestation of the coupled nature - due to the implicit time discretization - of the discretized equations). For example error propagates from top to bottom in the matrix.
More answers are most welcome.