My name is Joseph Turner and I’ve been designing a very high speed electric pump for the last 6 months.
Here is a screenshot of pump; well a cutout
doh I can only add three images into posts because I’m new! I’ll try and add later.
Blue is fluid.
I have impeller spinning at significant speed, 30,000 RPM and have flow of water at about 3 litres a second.
I’ve only started using libPrecice a few months ago, so I probably have some things configured wrong. I’m running version version libprecice2_2.5.0 on a Debian 12 machine, virtualized. The two solvers in use are:-
- Calculix 2.20
I’m doing a one-way FSI, because I make the assumption that the impeller displacements are insignifcant. Coupling is Explicit. Fluid uses simpleFoam, Calculix uses STEP,BLGEOM, STATIC.
The simulation shows a deflection of about 0.7mm in the Z-direction and shows considerable stress in the base of the impeller, indicating that some changes are required to reduce the stresses and deflection. Stresses look quite large, but need to remember that the design is definitely not optimal and needs improvements.
Here are some screenshots.
These were created using ccx2paraview. ccx2paraview generates vtk files for paraview as cgx in calculix is rather limited.
I’m sure I’ve made mistakes somewhere along the way… but it “seems” to generate some results that look “sensible” and I’m excited by it.
Improvements to make are:-
- get nearest-projection working (need to re-mesh)
- Simulate effects (stress and displacement) on the pump housing.
Thanks for making libPrecice!