[ZNoel]

Hello all,

I am trying to model the broadband noise spectra of a slightly underexpanded supersonic jet at Mach 1.5. I am using a realizable k-e turbulence model, ideal gas and sutherland viscosity for the gas. I set the reference acoustic power to 4E-10 in the acoustic model with 200 realizations and 50 Fourier modes. The energy equation was turned on and the solver methods were all second order upwind. Under the boundary conditions, I selected the non-reflecting acoustic model to specify the pressure at infinity (atmospheric), which makes sense to me because this is a steady state simulation. I set the Courant number to 1 with default under-relaxation factors (0.8). The simulation ran for about 50 iterations, and then the x/y velocities, continuity, and energy equations received a divide by 0 error.
I tried reducing the Courant number and the under-relaxation factors to 0.1 each, and the same type of error occurred.

I ran the simulation without the acoustic model and I got the shocks and Mach diamonds in the right place and was in good agreement with my theoretical calculations.

Any ideas on how I can get the acoustic model to work?