The outflow boundary conditions currently do not quite give physical results. This is due to several factors and is related to the "trapping" of excess pressure in the computational domain that cannot escape (through the boundaries). A partial fix is to include:
do k = 1,nz
do j = 1,ny
where(Qxhi_e(j,k,:,mx) > 0.0) Qxhi_e(j,k,:,mx) = -Qxhi_i(j,k,:,mx)
enddo
enddo
The other thing to try is to set the pressure to 0--being careful to acknowledge the pressure floor--through the energy equation (by setting the energy to the fluid kinetic energy).
The outflow boundary conditions currently do not quite give physical results. This is due to several factors and is related to the "trapping" of excess pressure in the computational domain that cannot escape (through the boundaries). A partial fix is to include:
The other thing to try is to set the pressure to 0--being careful to acknowledge the pressure floor--through the energy equation (by setting the energy to the fluid kinetic energy).