[m4D]

Hello.

I'm trying to simulate flow through an axial pump. The pump itself is an archimeds screw inside a pipe with a small radial gap. Model consists of three domains:
1) inflow part - is a stationary cylinder, from which i cut out front rotor bearing stands; a BC condition of total pressure = 0 (1atm, whatever) is set on it's front surface (let's call it A);
2) rotor - is a rotating cylinder (5000-15000rpm), from which i've cut out the impeller (2-blade archimeds screw);
3) outflow part - is a stationary cylinder, from which i cut out back rotor bearing stands and flow straightener; a BC condition of mass flow is set on it's back surface (let's call it B).
These 3 domains are connected together by 2 identical interfaces on touching surfaces: frozen rotor with auto pitch change (also tried stage with val = 1). Further explanations are in the attachment.

The problem is: the difference of pressures (n.m. which, total ones or not), i.e. a pressure rise, is always around some specific number, no matter which geometry of blades there's in flow straightener. The rotational energy of flow is converted in transitional energy, so the straightener must cause pressure raise, but i don't see it. BTW, if i remove the straightener completely, pressure rise is still practically the same, around 1-3%. Also, this particular pressure rise is far-far (~50%) away from the practically measured values. So, what's wrong with this simulation?

[ghorrocks]

Sounds like an FAQ to me: http://www.cfd-online.com/Wiki/Ansys..._inaccurate.3F

[m4D]

But the results are not inaccurate, they are independent of straightener geometry. And that made me think, that the problem was setup incorrectly.