[Syed Wajeeh]

Hello all. I am simulating a supersonic bell nozzle in ANSYS fluent with a pressure ratio of almost 50.32 and according to the analytical relation of steady one dimensional isentropic flow, I should get an exit Mach number of 3.21. When I run the simulation on a planar axi-symmetric 2D case, I get the desired Mach number but when I move to 3D case the Mach number increases to nearly 5.5 and I am unable to get the reason behind it. Shouldn't the Mach number be less in 3D case as compared to 2D ? I am a bit confuse in this case and I will be very thankful for any valuable reason/guidance regarding this issue. Thanks in advance to all :-)

[FMDenaro]

Originally Posted by Syed Wajeeh

Hello all. I am simulating a supersonic bell nozzle in ANSYS fluent with a pressure ratio of almost 50.32 and according to the analytical relation of steady one dimensional isentropic flow, I should get an exit Mach number of 3.21. When I run the simulation on a planar axi-symmetric 2D case, I get the desired Mach number but when I move to 3D case the Mach number increases to nearly 5.5 and I am unable to get the reason behind it. Shouldn't the Mach number be less in 3D case as compared to 2D ? I am a bit confuse in this case and I will be very thankful for any valuable reason/guidance regarding this issue. Thanks in advance to all :-)

Assuming that your 3D geometry is totally equivalent to the axi-symmetric case, you should check the flow solution. Have you get a steady solution having same residuals? Could you show the Mach number along the x-direction at the centerline? Are you sure that the BCs are identical?
Last question, your 3D solution retains the azimuthal symmetry ?

[LuckyTran]

And the grids are as identical as they can be?

[Syed Wajeeh]

Originally Posted by FMDenaro

Assuming that your 3D geometry is totally equivalent to the axi-symmetric case, you should check the flow solution. Have you get a steady solution having same residuals? Could you show the Mach number along the x-direction at the centerline? Are you sure that the BCs are identical?
Last question, your 3D solution retains the azimuthal symmetry ?

Thanks FMDenaro for a positive response. I am unable to understand what you are asking in the first bullet. Secondly, I am attaching a snip showing the Mach number variation. Yes the BCs are identical. Can you please elaborate a bit more on azimuthal symmetry please ? Actually I am bit new to CFD.

[Syed Wajeeh]

Originally Posted by LuckyTran

And the grids are as identical as they can be?

Thanks LuckyTran for your reply. Can we make identical grids for a 2D and 3D geometry ?

[LuckyTran]

As much as you can. For example, the same base grid size, the same wall clustering strategy, and so on. If this is a hex dominant grid for example they should be identical up to 3 topological defects. The 3D grid will have two line defects and an extra five-way point defect compared to the 2D grid. For tets and soccer-ball meshes, you have a greater number of defects but the same principle applies.

If you mesh each with different numbers of prism layers and growth rates and different base sizes and your solution is not converged on any of them, then it is expected you have wildly different results.


Check the solution to the 3D case and make sure it is axissymmetric. The BC's are axissymmetric so the solution ought to be. If the solution does not obey circular symmetry, then obviously it's not going to be equivalent to the 2D one.

[Syed Wajeeh]

Thanks a lot Lucky. Got your point. Stay blessed

[Syed Wajeeh]

Another thing I have noticed is different values of mass low rates in both cases. In 2D, it is 36.83kg/s and in 3D, it is 0.51kg/s. Can this be the reason for different Mach number scenario ?

[Syed Wajeeh]

Also, I have applied Symmetry BCs to the two face at zero and 90 degrees. Is it the right procedure to perform this analysis? i.e., run the simulation on quarter nozzle