[turbo]

Originally Posted by ngoc_tran_bao

Hi, Turbo. You can see on the 1st picture, there are short vectors in dark blue which indicate low velocity flow at pressure side of blade. Meanwhile, at suction side the appearance of long,light blue vectors which indicate high velocity is observed.

Still I do not see what you see. Looking at your 1st pic, generally I see many dots near the suction surface of the pump impeller, and relatively higher velocity vectors near the pressure side. From the inlet your operating point undergoes a large negative incidence.

[ngoc_tran_bao]

Turbo, I think this picture will help, feel free to correct it if I'm wrong.

[turbo]

Originally Posted by ngoc_tran_bao

Turbo, I think this picture will help, feel free to correct it if I'm wrong.

No pics.

Have a look at leading-edge area. I see large flow separations on the suction side due to very high incidence from your part-load operation flow. No issues I see from your cfd.

[ngoc_tran_bao]

I have read some papers and I found that in their research the flow separation which occurs at suction side likes this below pic . On pressure side, the flow is attached on the blade and follows the curvature well (different from my pic) and that is the right behavior the flow should be.

[ghorrocks]

If your simulation is running at off-design conditions then all sorts of weird stuff happens.

If your simulation is not converged you can also get weird stuff happening.

And finally, if your simulation is actually transient and you just look at a single frame of a transient simulation things can look weird as well. You might need to do a time average to get a good mean flow field.