Supersonic Overexpanded Impinging jet - Convergence issue
 

Hello,

I am working on impinging jets and to get confidence with Fluent, I am trying to get similar results from a template report. This report has been written by Xu and al. (2011), « Particle image velocimetry study of the impinging height effect on an overexpanded supersonic impinging free jet at Ma=1.754. Part I: global coherent structure and Part II : detailed velocity distribution ».

Here are the speficiations :

• Ma = 1.754
• Nozzle Pressure Ratio (NPR) = 4.7625
• T = 300K

https://i.postimg.cc/dkp9RqLb/Nozzle-Geometry.png

And I used this geometry :
https://i.postimg.cc/KkB1VF9S/Axisymmetric-geometry.png

I have obtained a mesh qualified as "good" by mesh metrics spectrum from ANSYS, with inflation (First layer thickness at 2.27e-6m).

I obtained quite good results with Pressure based (Axisymmetric option, SST K-omega, Energy : On, Pressure Inlet : 482 560.3 Pa from the NPR, Pressure Outlet : 101 325 Pa, Scheme : Coupled, 2nd order, steady state) but there is a physical inconsistency at the centreline :
https://i.postimg.cc/TKn3qG3Z/Centreline-Velocity.png

Given the Mach number (and so the high velocities), I have chosen to use the Density-Based Solver with the axisymmetric option, SST K-omega, Energy : On, Pressure Inlet : 482 560.3 Pa given from the NPR, Pressure Outlet : 101 325 Pa, operating condition : P = 0 Pa (absolute value), T=300K, Formulation : Implicit, Flux Type : Roe 2nd order.

BUT this time I can't have a convergent solution. I tried Pseudo Transient, no Pseudo Transient, starting with 1st Order and change in 2nd after having a stable (but no converging) solution, decrease the factors, increase and decrease the courant number (Without Pseudo Transient), using a Time Scale Factor of 0.1 , 0.01 , 0.001 and 0.0001...

And I tried these for many iterations.. :o

I can't have residuals below 1e-2 (target : 1e-6)
https://i.postimg.cc/SYyzMR4Q/Residuals.png

Yet, the velocity contours, (for this no convergent solution) show a better solution without the physical inconsistency :
https://i.postimg.cc/RNhW5jVY/Velocity-contours.png

Do you have a idea why I can't have a convergent solution ? How Can I have a more robust solution ?

Thank you ! :)

Hi Guillaume

I may have a solution for this kind of situation.

First I have to say I cannot see any of the figures in this page, and what I am doing is a rotor blade instead of a jet. So my case (research object, working condition, etc) is quite diffrent from yours.

However, the thing is that I have met exactly the same thing as you did, convergent with pressure-based solver but divergent with density-based solver. What I tried to solve this was to set the operating pressure=101325 Pa, and the gauge pressure=0 Pa. I know it is not recommended by Fluent but it was the only way I solved this problem and got a reasonable result of the flow field.

So far I don't know why and this does not seem to happen everytime. Technically when velocity is high it is more suitable to set the operating pressure as 0, but for my case it just could not converge.

Hope my solution could help you. If you know the reason for this, or you have found another solution to your issue, please tell me as well and we could discuss this in more detail.

Thanks
Nathan