Learning Forum

[ibak91]

Hello,

I am still pretty new to ANSYS and the forum so my apologies if this is not the right place to ask. I am simulating flow over an airship and I need to locate where the flow transitions from laminar to turbulent flow. To find the point where it transitions, I am plotting velocity profiles at multiple points over the airship using CFD-Post. The velocity profile chart demonstrates that the flow is transitioning from turbulent to laminar, which I think to be inaccurate. I am using the k-ω SST transition model and the flow is flowing towards the -x direction.

[SVV]

Hi Please use Reynolds number to find the transition. Also ensure that your mesh quality is good and simulation is converged.

[ibak91]

Hi I tried using Reynolds initially but when I calculated it and displayed the contour I get this. I am showing both the formula I used and the contour. As you can see the Reynolds Number from the start of the flow is negative 10^6. which is confusing.
I believe my mesh quality is good:
Mesh Quality:
Minimum Orthogonal Quality = 1.11923e-01 cell 5358761 on zone 138 (ID: 5358762 on partition: 0) at location (-3.89613e-01 -3.00474e-01 -1.81408e-03)
Maximum Aspect Ratio = 6.79135e+02 cell 5335807 on zone 138 (ID: 5335808 on partition: 0) at location (-3.89023e-01 -3.00435e-01 -1.75374e-03)
The simulation converged at 678 iterations.

[Rob]

You're plotting a function of the x-coordinate so Re will vary over the domain by position as well as everything else.
Cell quality is OK. How well resolved is the boundary layer, streamwise on the blimp and flow in the wake region?

[ibak91]

Hi Rob I am sorry but I am not sure what you mean by resolved, do you perhaps mean refined? The boundary layer has 25 layers with a y+ < 1.
I just noticed that the boundary layer gradient does not go over the fins of the blimp, could that be a reason why my results are incorrect?

[Rob]

Possibly. By refined I mean are you picking up the flow gradients. Everyone reads about y+ and religiously applies the "rule" then forgets to refine the far field. The jump from the top layer of inflation to the poly mesh is a little large, but not excessive.
How does the above look with the velocity contour plotted?
If you're looking for when the boundary layer "trips" to turbulence you may need one of the turbulence transition models, and then monitor k and/or epsilon/omega. It's really not my field (I don't do aeroplanes) so you may need to do some reading. Nearest I got was working with some sports scans prior to the swimsuit simulation work https://www.theengineer.co.uk/speedo-uses-ansys-simulation-to-develop-racing-swimwear/

[ibak91]

I am currently using the K-omega SST model for my simulations, would the monitor residuals show when the boundary layer trips to turbulent? Also the velocity contour has a boundary layer only on the first 4 layers of the wall around the fins.

[Rob]

The residuals are a function of the total error in the solver, as such they're used to see if the solution has converged, but are otherwise useless. What you're looking for are monitors. If you plot k and w (separately) on the blimp surface does it show anything?