localized VERY high velocity

[pei-ying.hsieh]

Dear Autodesk experts, I am running a case that produced localized VERY high velocity. Overall, the velocity field looked OK. But, there is a very localized spot that has very high velocity (that is not reasonable). I suspect that this might be due to bad elements. What will be the best approach to resolve this type of problem? Thanks! Pei-Ying

[OmkarJ]

How about mesh adaptation? Or, if you know now the sensitive areas  of high velocity gradients, you can ceate regions at these places for a good resolution of velocity field.

[pei-ying.hsieh]

Hi,

Thanks for the suggestion.  Adaption is a possibility.  But, I would like to first resolve this issue without adpation.

How to create region?

I remembered that there is a way to find out bad quality elements, but, cannot remembered how.  Also, refining elements sometimes make things worse.

Pei-Ying

[pei-ying.hsieh]

Hi, Omkar, I found something strange. During the steady state run, I saw maximum velocity went up to something like 250 m/s. However, when I stopped the calculation, the velocity field looked OK and the maximum velocity was at a reasonable value, like 15 m/s. I am wondering why that is. Any thought on this? Pei-Ying

[OmkarJ]

Have a look at the picture, it shows where you can find Regions. You can use cylindrical, spherical or box shaped regions.

The localised high velocity is a typical sign of divergence. I would improve the mesh and give it a try, and if it still gives problems, try reducing the underrelaxation factors in Solve>>Solution Control window. Also if you want to see the quality of mesh, enabe "Stream function" in Solve>>Resullt Quantities and once you finish meshing observe the results of "Nodal aspect ratio". You may want to see the iso-surface of it so you can understand where the bad elements are located. 

Also, for a steady simulation, enabling Intelligent Soluction Control also helps in preventing divergence. 

[pei-ying.hsieh]

Hi,

I am seeing something very strange.  During steady state calculations, I saw very high maximum velcity values on the legend (something like 330 m/sec).  However, the case converged at iteration #238, and all of a sudden, the maximum velocity value droped to a very reasonable value, around 17 m/sec.

What happened?

Pei-Ying

[Royce_adsk]

Maybe you had your legend fixed at a certain value and then when the results came in it reset the legend to auto range?

[pei-ying.hsieh]

Hi, Royce,

This is a puzzle to me.  I believe that the legend was set to auto.  If the values were fixed, it will most likely be fixed to a reasonable value (ie, not very high).

On one case, when I stopped the steady state calculations prior to convergnce, I did see 3 "spots" that had very high local velocity.  However, when I re-run the case (starting from step = 0) and let it run to convergence, these local high velocity were gone (dropped from like 333 m/s to 20 m/s).

I will see if I continue to see this phenomon.

Pei-Ying