Hi all,
I'm simulating a constant are duct (scramjet isolator) at M=2.0 in fluent with k-omega SST turbulence model.
In steady RANS simulation a shock wave moving over iteration i.e the shock location moving upstream and converge after several 1000 iteration. once it is converged switching to transient case and giving a small back pressure (sinusoidal frequency) resulting in constant upward movement of shock and it moved out of domain. but i should get upstream and downstream movement with respect to back pressure variation.
Anyone has any idea why this kind of situation occurs in CFD. Though the grid resolution is refined to get very clear shock structure in steady simulation.
Have you tried to start from the transient formulation and see if that still happens? Is the lenght of the duct suitable to describe the case?
Did you conduct a sensitivity study of your timestep in relation to the back-pressure oscillation frequency. Based on the time period of your oscillation, you could experiment with multiple timesteps and see if the problem persists.
A shockwave position is very sensitive to a pressure amplitude, especially when you are rising the pressure.
Try to decrease an average pressure, so your maximum would not exceed a value used for the stationary case.
Check the dependency on the time-steps and mesh configurations. by reducing the absolute size of the cells and lowering the time-step check the sensation of the problem on the aforsaid factors.
- Badges
- Science topic