[Amber0922]

Hi,
I simulated a cylinder case in SU2 with a Mach number of 0.01 and set the inlet boundary condition as farfield, which I believed would remain unchanged. However, after running for some time, the Mach number increased unexpectedly. Therefore, what type of inlet boundary condition can maintain its original state? The details of the boundary are provided below.

-------------------------------------------------------------------------------------
MACH_NUMBER=0.01
AOA= 0.0
SIDESLIP_ANGLE= 0.0
FREESTREAM_TEMPERATURE= 288.15
REYNOLDS_NUMBER= 3900
REYNOLDS_LENGTH= 1.0
% Navier-Stokes wall boundary marker(s) (NONE = no marker)
MARKER_HEATFLUX= ( CYLINDER, 0.0 )
%
%
% Far-field boundary marker(s) (NONE = no marker)
MARKER_FAR= ( FARIN )
%
% Pressure outlet marker
MARKER_OUTLET= ( FAROUT, 10)
%
%%Symmetry boundary marker(s) (NONE = no marker)
% Implementation identical to MARKER_EULER.
MARKER_SYM= ( UPPERWALL, LOWERWALL, FRONT, BACK )
% Marker(s) of the surface to be plotted or designed
MARKER_PLOTTING= ( CYLINDER)
%
% Marker(s) of the surface where the functional (Cd, Cl, etc.) will be evaluated
MARKER_MONITORING= ( CYLINDER)
MARKER_ANALYZE = ( CYLINDER )

[bigfootedrockmidget]

Hi, this sounds a convergence issue, in the limit that the residuals go to zero the boundary conditions should be respected. So how is the convergence for your case?

[Amber0922]

Originally Posted by bigfootedrockmidget

Hi, this sounds a convergence issue, in the limit that the residuals go to zero the boundary conditions should be respected. So how is the convergence for your case?

Hi，
The case do not diverge, and the physical phenomena and trends agree well with the experiments, but the inlet velocity increases over time, which causes the Reynolds number to change. How can I adjust my boundary conditions and parameter Settings so that my entry conditions remain the same？

[RcktMan77]

You have indicated that you're simulating a cylinder case in SU2. What does this mean? Are you simulating internal flow through a cylinder, or are you simulating external flow around a cylinder?

If you're simulating external flow around a cylindrical solid body, then you shouldn't be specifying a static pressure outflow boundary condition. Both your FARIN and FAROUT boundaries should be set to a Riemann Invariants farfield boundary condition.

If you're simulating internal flow through a cylindrical passage, then the upstream inflow boundary should be set to a subsonic inflow boundary condition in which you prescribe total pressure and total temperature, and your downstream exit boundary should be prescribed as a subsonic outflow boundary condition in which you can specify static pressure, (or mass flow out).

Note that you're simulating a very low Mach flow condition while using the compressible flow solver. This situation introduces numerical instability and a slow convergence rate due to the large disparity between the acoustic wave speed and waves propagating at the fluid convection speed. You may want to opt for using either the incompressible flow solver in this situation, or use the Turkel pre-conditioner to scale the time derivatives to help improve robustness and convergence rate. Otherwise, obtaining a converged solution is going to likely require an extended number of iterations before you reach convergence.

[Amber0922]

Originally Posted by RcktMan77

You have indicated that you're simulating a cylinder case in SU2. What does this mean? Are you simulating internal flow through a cylinder, or are you simulating external flow around a cylinder?

If you're simulating external flow around a cylindrical solid body, then you shouldn't be specifying a static pressure outflow boundary condition. Both your FARIN and FAROUT boundaries should be set to a Riemann Invariants farfield boundary condition.

If you're simulating internal flow through a cylindrical passage, then the upstream inflow boundary should be set to a subsonic inflow boundary condition in which you prescribe total pressure and total temperature, and your downstream exit boundary should be prescribed as a subsonic outflow boundary condition in which you can specify static pressure, (or mass flow out).

Note that you're simulating a very low Mach flow condition while using the compressible flow solver. This situation introduces numerical instability and a slow convergence rate due to the large disparity between the acoustic wave speed and waves propagating at the fluid convection speed. You may want to opt for using either the incompressible flow solver in this situation, or use the Turkel pre-conditioner to scale the time derivatives to help improve robustness and convergence rate. Otherwise, obtaining a converged solution is going to likely require an extended number of iterations before you reach convergence.

Thank you for your help. Do you mean I should set the MARKER_RIEMANN=(FARIN, FAROUT) in my case and choose the imcompressible flow solver instead?

Thank you again for your unselfish help!

[RcktMan77]

I think you will be better off switching to the incompressible RANS solver for your particular flow condition. I'm not sure what the MARKER_RIEMANN boundary condition type does in SU2--I would have to look into its implementation in the source code as it isn't documented. I think the MARKER_FAR option should be appropriate for your case.

For your particular case, I'm not certain about what the boundary condition setup should be as I don't quite understand what your outer boundary topology looks like or exactly what you're trying to model (i.e., external or internal flow). I'm confused why you have multiple boundary patches set to a symmetry boundary condition type.

If your outer boundary represents a box, then all boundary patches of the box would be set to MARKER_FAR except the boundary patch that lies along the symmetry plane. If your outer boundary topology represents a hemisphere, then there would be a single hemisphere boundary patch set to a farfield boundary condition type and a single symmetry boundary condition. Similarly, if your outer boundary topology is a cylinder, then you would have 3 boundary patches of the cylinder set to farfield boundary condition types and a single symmetry boundary condition.