[cbooks77]

Hi all I was curious if anyone has ever experienced fluid strain rates around sharp edges for example an impeller blade that are exceedingly high. I am running a fairly high mesh resolution (y+<1) case on a hydrofoil impeller and the volume averaged shear rates (fluid strain rate) inside the MRF region is around 20/s but the max strain rates being reported are around 100,000/s. Upon inspection these exceedingly high strains are occurring on the blade edges which were not filleted in the pre processor (see attached screenshot).

Now, since the strain rate is composed of velocity gradients I am wondering if the high resolution mesh around the sharp edge is causing the strain rate to blow up much like a stress singularity in FEA. For example, consider du/dx now as dx approaches zero it stands to reason that du would blow up.

Can anyone corroborate this theory so I can feel better about my results it will be nice to hear from you especially since I have never come across this topic in any CFD blogs or articles.

Kind Regards

[FMDenaro]

Originally Posted by cbooks77

Hi all I was curious if anyone has ever experienced fluid strain rates around sharp edges for example an impeller blade that are exceedingly high. I am running a fairly high mesh resolution (y+<1) case on a hydrofoil impeller and the volume averaged shear rates (fluid strain rate) inside the MRF region is around 20/s but the max strain rates being reported are around 100,000/s. Upon inspection these exceedingly high strains are occurring on the blade edges which were not filleted in the pre processor (see attached screenshot).

Now, since the strain rate is composed of velocity gradients I am wondering if the high resolution mesh around the sharp edge is causing the strain rate to blow up much like a stress singularity in FEA. For example, consider du/dx now as dx approaches zero it stands to reason that du would blow up.

Can anyone corroborate this theory so I can feel better about my results it will be nice to hear from you especially since I have never come across this topic in any CFD blogs or articles.

Kind Regards

if you think the simple case of a 2D corner, you understand the issue as due to the fact that the derivative cannot be defined, the corner is indeed a singular point. You have different limits.

[sbaffini]

Yet, there might also be post-processing issues in your picture, as you are apparently visualizing on the nodes, which is not the actual fv solution

[cbooks77]

Originally Posted by sbaffini

Yet, there might also be post-processing issues in your picture, as you are apparently visualizing on the nodes, which is not the actual fv solution

Hi that's correct I showed the interpolated data since its a boundary face and Fluent uses a simple averaging of the adjacent cell face values I believe. Also the cell values show similar behavior around the sharp edges. Thanks for your response