[fjavi90]

Hello everyone

I am trying to implement a subsonic outlet boundary condition for the Euler equations on one side of a cartesian grid and I read this paper and this paper but I cannot get it to work as expected.

I am using a one-dimensional approach on the boundary and I am using ghost-cells. I've already tested that the simple extrapolation technique does not work. Is there any simpler appoach than that of those papers that does the trick?

At the moment I am testing a simple vortex convection case, and the vortex does not leave the domain properly and generates spurious oscillations.

Any help will be appreciated.

Thank you

[FMDenaro]

Originally Posted by fjavi90

Hello everyone

I am trying to implement a subsonic outlet boundary condition for the Euler equations on one side of a cartesian grid and I read this paper and this paper but I cannot get it to work as expected.

I am using a one-dimensional approach on the boundary and I am using ghost-cells. I've already tested that the simple extrapolation technique does not work. Is there any simpler appoach than that of those papers that does the trick?

At the moment I am testing a simple vortex convection case, and the vortex does not leave the domain properly and generates spurious oscillations.

Any help will be appreciated.

Thank you

Provide the details of the BCs. setting you are using

[fjavi90]

Originally Posted by FMDenaro

Provide the details of the BCs. setting you are using

Thank you for your quick reply. I have a mean flow with the primitive variables being:
rho_inf = 1; ux_inf = 0.5; vy_inf = 0.0; p_inf = 1;

The vortex is superimposed as a perturbation on that mean flow. Since the velocity goes to the right, and it is subsonic (being c_inf = sqrt(1.4) = 1.1832), on the right edge I need to impose one of the variables (because there is one wave that is entering the domain), and I read that the most common choice is to impose the pressure at the boundary and extrapolate the other variables from the interior, am I right?. I don't know what is the value of the pressure to impose (and the corresponding values at the ghost points) since it is not a far-field BC, so the pressure at the outlet varies over time.

I can always impose the analytical solution on the ghost points for every time step, and then the vortex leaves the domain perfectly, but this test case is to try to understand the precedure in order to apply it to other cases that don't have analytical solution.

I hope I made myself clear, sorry if it is a trivial question.

Thanks

[FMDenaro]

Originally Posted by fjavi90

Thank you for your quick reply. I have a mean flow with the primitive variables being:
rho_inf = 1; ux_inf = 0.5; vy_inf = 0.0; p_inf = 1;

The vortex is superimposed as a perturbation on that mean flow. Since the velocity goes to the right, and it is subsonic (being c_inf = sqrt(1.4) = 1.1832), on the right edge I need to impose one of the variables (because there is one wave that is entering the domain), and I read that the most common choice is to impose the pressure at the boundary and extrapolate the other variables from the interior, am I right?. I don't know what is the value of the pressure to impose (and the corresponding values at the ghost points) since it is not a far-field BC, so the pressure at the outlet varies over time.

I can always impose the analytical solution on the ghost points for every time step, and then the vortex leaves the domain perfectly, but this test case is to try to understand the precedure in order to apply it to other cases that don't have analytical solution.

I hope I made myself clear, sorry if it is a trivial question.

Thanks

I found good results fixing the total pressure at the inlet and the static pressure at the outlet