Erroneous Inlet Behaviours for Multilayer Model?

[Godunov CIR]

Hello all,

I encountered a seemingly erroneous behaviour of the inlet in the multilayer model. My application is to simulate the steadily-propagating surge on a dry inclined slope. The setup is essentially similar to the propagating front of the granular flow in the sandbox: granular front of M1EMN:, but my simulation uses the multilayer model. I expect to see a steadily propagating surge on the dry inclined bed, but I got zero depth everywhere, which is quite weird.

I used hydro.h for the hydrostatic multilayer code. 32 layers are used here. The initial condition is the dry bed:

foreach() {
    foreach_layer()
      h[] = 0.0;
    foreach_layer()
      u.x[] = 0.0;
  }

And the boundary conditions are steady-uniform flow at the inlet and free-outflow outlet:

  u.n[left] = Qparab (point);
  eta[left] = dirichlet(1.0);

  u.n[right] = neumann(0.);
  eta[right] = dirichlet(1.);

I also attached the source code "surge.c", which was modified from the sandbox code C. Robert's Kapitza wave, but the inlet perturbation, surface tension and non-hydrostatic features were removed for simplicity. Running the code results in zero velocity and depth everywhere, and the fluid does not enter the domain. This erroneous behaviour is unexpected (I expect to see the steadily propagating surge).

I do not know the reason leading to the strange results. Thank you in advance for any inspiring comment.

Jeffery Y.