[kandelabr]

Hi.

I'd like to visualize spots with dissipation losses in my domain (a transient k-omega SST pimpleFoam case with dynamic mesh). I have the following options:

• stream tracer (shows where vortexes are but streamlines end on AMI patches)
• vorticity plot (not sure what that shows, though)
• Q-criterion iso-surfaces
• Entropy calculation (https://www.mdpi.com/1099-4300/16/6/2959)

I have found many examples of Q-criterion but my problem is that the values range from say -1e+17 to +1e+17. How do I choose values for isosurface plots so that they give me the most information?

Thanks!

[akidess]

AFAIK there is no real way to chose the value for the isosurface - you choose what looks good (see [1]). However, your value range seems unbounded, which makes me believe you may have a small error in your computation.

[1] Dubief, Y., Delcayre, F., 2000. On coherent-vortex identification in turbulence. Journal of Turbulence 1.
URL http://dx.doi.org/10.1088/1468-5248/1/1/011

[kandelabr]

Thanks for the reply.
Actually the values are more like -1e+7 to 1e+7; does that make any difference? What range should Q-criterion usually be in?

[LuckyTran]

The range of Q-criterion is normally some finite negative number to positive infinity. That you have a nice symmetric negative 1e7 and positive 1e7 suggests you might have some bug. But it might still be correct. Anyway, just check.



The isosurface of Q-criterion you should plot is "some positive number that you like" Q-criterion is also dimensional in its pure definition, so one type of flow you take Q=1 and another type of flow you might take Q = 1000. You could normalize the Q-criterion to help with this. But because this is so hard to decide (it's quite arbitary what this threshold is), the closely related lambda 2 criterion is often preferred over Q-criterion (where the threshold is 0).

[kandelabr]

Lambda2, you say.
Never heard of that, will try out, thanks a lot!

[akidess]

The paper I linked in my post a month ago also explains the lambda2 criterion.

[kandelabr]

...but unfortunately isn't available without paid access.

[mazhar16823]

Originally Posted by LuckyTran

The range of Q-criterion is normally some finite negative number to positive infinity. That you have a nice symmetric negative 1e7 and positive 1e7 suggests you might have some bug. But it might still be correct. Anyway, just check.



The isosurface of Q-criterion you should plot is "some positive number that you like" Q-criterion is also dimensional in its pure definition, so one type of flow you take Q=1 and another type of flow you might take Q = 1000. You could normalize the Q-criterion to help with this. But because this is so hard to decide (it's quite arbitary what this threshold is), the closely related lambda 2 criterion is often preferred over Q-criterion (where the threshold is 0).

Hi LuckyTran,


I am struggling with the same. Although I have created Q-criterion but half of the blade is taken away (maybe due to the flow separation) however I have to compare my RANSE Q-Criterion (STARCCM) with reference Q-Criterion (LES)as you can see in the attached images. Why the half of blade is taken away, is there anyway to fix this issue and represent the whole Q-criterion on the blade surface?