[mhd_mrt]

Hello all,

I am trying to simulate a moving normal shock wave in a tube by specifying the pressure, temperature and velocity before and after the shock wave. For some reason my results aren.t perfect and I am wondering id it has to do with my boundary conditions or numerical dissipations or ...? The results should follow Anderson's Modern Compressible Flow, Chapter 6.

Below you can see my problem setup. I have a 2D rectangular mesh and I am running Unsteady Simulation, Dual Time Stepping, ROE scheme with Runge Kutta Explicit Time Discretization. Any comment or suggestion with regard to capturing this moving normal shock will be highly appreciated. My config and result are attached.

Thank you!
Mehdi

[talbring]

Hi Mehdi,

what exactly is your problem ? Do you get oscillations near the shock ? If so, try to reduce the value of LIMITER_COEFF in the config.

Tim

[mhd_mrt]

Hi talbring!

Thanks! Yeah I realized that LAX-FRIEDRICH is inherently very dissipative across shocks and the mesh resolution needs to be super refined in order for the solver to produce comparable results to either HLLC or ROE.

It also very well depends on the strength of the moving shock (=pressure ratio). I was running this for the extreme case of P2/P1 =10 which corresponds to even a higher P4/P1 ratio. Also the oscillations are occurring at such a short distance when considering the speed of the moving shock (W)

Thanks!