I am learning simulation of centrifugal pumps and I have ran into problems with boundary conditions. When defining an inlet volume/mass flow and a pressurized outlet conditions, the simulation converges nicely and gives somewhat sensible results. However, when it comes to designing new products it would be much more useful to define pressure at the inlet and volume/mass flow at the outlet, since in reality the outlet pressure is unknown and it's what we want to calculate. Suddenly, with the boundary conditions changed around, the calculation does not converge even after triple the amount of iterations needed to finish an inlet flow + outlet pressure calculation. There are no signs of conversion and settlement in the values at all. I am using the following setup:
General: Internal, Local rotating region 300rad/s, no cavitation
Boundary Conditions: Inlet Total Pressure 2bar, Outlet Volume Flow 0,025m3/s, Real Wall with Angular Velocity on impeller surfaces not covered by rotating region
-Impeller Torque (All impeller surfaces in contact with fluid, rotation axis)
-Inlet Volume Flow (Just for convergence, should be the same as Outlet boundary condition)
-Outlet pressure (According to real tests, should stabilize at 400kPa)
-Inlet pressure (200kPa)
-Head (Pressure difference in bars * 9.81)
-Efficiency (pressure difference * volume flow / torque / angular velocity)
Mesh: 400k fluid cells, 400k partial cells. Mesh is very rough, but it doesn't explain why the calculation works with BCs the one way, and doesn't work with BCs the other way.
Any ideas what could be causing this? I am using a model of an existing pump with known specs. The pump should work with given conditions.
Thank you for help.