Supersonic TE tip in Centrifugal Comp.(Turbo)
 

Hi,

I have been working on commercial turbo compressor wheel which is 3d scanned by me (for academic purposes). Despite there is a little deviation in the process of STL to CAD transformation, I could get nearly exact Pressure ratio for first speedline in performance map which is given by producer. But in the second speedline, flow at the trailing edge exceed speed of sound and this caused, decrease of pressure in the vaneless area and it doesn't suit with map. Are there any one who face with such problem?

My BC's:
Steady state analysis

Inlet: Stationary Total Pressure 101325Pa 300K Total temperature

Outlet: Exit corrected mass flow (Referances for sea level)

Air Ideal Gas

Total Energy Viscous term on

Physical timescale 0.5/w

Pressure ratio measured between outlet of volute to inlet.

In this conditions it converges without problem even under 1e-5 rms.

Meshed in Turbogrid +y is at the trailing edge around 58-60. This value is for highest cutoff refinement level possible in Turbogrid. I can't lower further. I quadrupled mesh even I got same solutions.


Volute is not original one, I drawed it. Area/Radius ratio is 0.65 original one is 0.5 but producer says A/r of compressor volute doesn't effect turbo performance significantly.

The performance map is on attachments. I also have doubts about the outlet BC, when I use mass flow outlet it diverges in the beginning. I want to know if exit corrected bc is same thing that mentioned as corrected mass flow in the map.Thanks for your help:)

What is the MIN y+-Value, what is the MAX y+-Value? Did you do a mesh-study yet? Coarser grids tend to smooth out oscillations and stuff. So the residuals are may be low - but the solution is inaccurate

A volute can affect the compressor-stage very well. In the picture you posted here, i can see low mach-numbers in the volute (downstream, near outlet). This could be a boundary-layer detachment and this could be the reason for a low efficiency. You'll need to use the right geometry.

Are you using Mixing-Plane-Interfaces or FrozenRotor interfaces?