I'm trying to simulate an agitator / submersible mixer with a 2-blade propeller. The speed is given and I have measured values for torque and thrust/ force. If I just simulate this case, like other classical turbomachinery(centrifugal pump), I get results that differ by more than 100% of the measured data.
The mesh looks very good, and 2000 steps were calculated (so many steps are necessary because for over 1000 steps large eddies occur around the blades until the flow aligns to the front). I've simulated many different variants of this case, but always getting poor results.
Who can tell me what are the best settings for submerged propellers (completely surrounded by fluid / free flow, without any housing)? What could be the problem?
I would say you might need more mesh downstream.
What Advection scheme are you using?
Are you running in 2 stages, ramping up over time and then a smaller 3deg/step for the rest of the analysis?
Please could you share an image of the mesh - showing the meshed blades and also a cut-plane with mesh through the blades and to the rear of the assembly.
The last model I ran like this I used motion and no Rotating Region, which seemed to work well, although really this setup should be OK.
Simulation with other submersible mixers have shown, that the size of the external area are almost irrelevant to the applied torque and the thrust.
I'm using ADV5, but also tried with ADV1 (no difference); I'm ramping up rotational speed in 50 time steps, then run 1000 steps with 40deg and further 1000 steps with 3 deg. Meshes follow...
It is possible you need more mesh on the impeller itself
We could consider SST k-o in the event that there is a bit of separation from the blades although k-e should be ok.
Did the torque curve flatten out across the final 1000 iterations?
As you are a gold subscription member, have you considered logging this as a support case where we can then investigate the model ourselves and track this properly?