Dear FLUENT Users,

Just the other day I ran into an interesting problem concerning the use of new PBCS in FLUENT 6.3.26 for the calculation of NASA Rotor 37. I have done many previous calculations of this rotor using Fluent 6.1 and 6.2 versions, with the "classic" density-based coupled solver. The results were quite good, and I suceeded to match very closely the pressure-ratio at nominal rpm over the entire range (but I had trouble in matching efficiency...).

But when I tried to use the new PBCS, although I was very satisfied with the speed of the solver (I obtained converged solution in 500 iteration directly at nominal rpm ramping CFL from 1 at start up to about 30 within first 100 iterations), the results were rather odd: due to a low back pressure, in the blade-to-blade channel there were shocks present which looked fine and realistic, but the bow shocks were really messed up! That is, the angle between the shock and the relativ flow direction was too small (~55 deg instead of ~70 deg!!!), one bow shock from a blade almost touching the leading edge of blade next to it!!!

It's obvious that the calculated performance of the rotor was heavily altered, for example the PR being too small (1.82 instead of 1.97 at the calculated mass flow).

This is quite disturbing, because it determines me to question the applicability of this solver not only to this kind of problems, but to all shock-related simulations! Although I have to admit that in other cases that I have tested the PBCS, the results seemed OK, even the shock locations (but I never investigated the shock angle, maybe I should try a supersonic flow over a wedge, or something like that, to test that).

Has anyone else noticed such problems?

Maybe I should also mention the settings:

- discretisation: pressure 2nd order, all other QUICK;

- URFs: pressure, momentum & density 0.5, all other 0.75.

The mesh is 1.5 milion nodes, fully structured multi-block, mean y+=0.9.

Hoping for an answer,

Razvan

Dear Razvan,

I wanna ask about the Rotor 37 Geometry. From where did you get it. and can you please send it to me.

I will be thankful to you.

(sorry for asking irrelevent thing)

Regards sam

Sorry for the intervention, but I have recomputed the model switching to density-based solver, obtained similar results, the only noticeable difference being the higher resolution of the shockwaves. Now the bow wave is more visible, and after re-examining the PBCS results I concluded that what seemed to be a much too inclined bow shock, was actually the expansion fan that originates at the very leading edge of the blade. I did not noticed the bow shock the first time due to its graphic "weakness" (this PBCS seems to smooth the shocks a bit, after all it is not a true coupled solver...). By restraining the scale to the supersonic values of relative Mach number, the bow shocks became visible. And they are where supposed to be.

Once againg, I appologise for the confusion.

Nevertheless, I cannot help to notice that the SST-kw model that I used in these two calculations is far from suitable (previous calculations were made using Realizable k-e, but this time I choosed SST-kw taking into consideration the value of y+=1...). At least that is the first conclusion I can draw for now. Will further investigate...

All the best,

Razvan

Dear Sam,

I have already sent you an email about 5 days ago, but it seems that you did not receive it.

Anyway, the thing is that I cannot tell you where I got the geometry from, but I can give it to you. Just send me an email, and will continue from there.

All the best,

Razvan

Would you please also send me a copy of nasa rotor 37 geometry? thanks a lot!

[Divyaprakash]

Can you send the geomtery to me as well?

[pauldarla89]

Hello, Can you also send me the geometry of NASA Rotor 37 to this mail id paul.totty89@gmail.com.



Thank you, i appreciate it .