[310toumad]

Hello,

I'm running analysis on a centrifugal blower and trying to use CFD To match the performance curve based on lab data. My boundary conditions are volume flow inlet, and static pressure outlet (atmospheric). I adjust the volume flow based on the performance curve data and let SW calculate the pressure rise over the blower.

All of my results are lower than the lab data by quite a bit, the first two data points at higher pressures are only off by about 5-7% however as I get closer to a free flow condition the disparity becomes greater. At free flow SW is telling me there is a negative pressure rise, so what I take from that is it means the solver is essentially saying this blower is incapable of pushing through that much air on its own.

The RPM's of this blower are quite high in the range of 7600-10000. I am wondering if the turbulence model built into SWFS is not accurately capturing the physics in this instance. I believe I've done all I can to implement the best practice for this application in this software. I have it set as a transient simultion with sliding mesh, I'm confident my rotating region is set up correctly, I have done a grid convergence study, and tried every combination of boundary conditions. Always the result is an underprediction.

Many white papers I've read use the k-omega SST model and seem to get results that closely match the performance curve for blowers, so is k-e not the appropriate choice in this instance? Unfortunately I cannot try it in SWFS as the software does not grant the user that option.