Learning Forum

[info]

🛈 This post originally contained file attachments which have been removed in compliance with the updated Ansys Learning Forum Terms & Conditions

Hello,

Problem: the cumulative moving average values of the Transient CFD simulation of the drag and lift forces are not assimptotic to the prior Steady State ones.

I set up a parametric study to investigate the aerodynamic resistance of a structural steel angle section in quasi-2d.

For a fixed geometry (L.100.10) I set the wind incidence angle as parameter, extracted the Cd/Cl coefficients and include the results in a chart.

See attached the archived AIM file with results included for the 125°wind incidence angle case.

Please also find some screenshots of model,, mesh and results.

Characteristics:

-          wind incidence from 0-180° in 5° steps

-          k-e standard RANS model in both Steady/Transient cases (none of the parameters changed in-between)

-          boundary layers conservatively set, resulting y+ varies between 0.53-17.13

Other than 2-3 wind incidence angles, the Steady State solution converged ’normal’, see the residual plot attached for the 125°case.

The Cd/Cl coefficients are put in a chart – see attached.

In order to ’validate’ the Steady results, I wanted to make sure about what is obvious in CFD theory, e.g. that the long term transient Cd/Cl values are assimptotic to the Steady state ones.

So by only flipping the solver from Steady to Transient (left every setting as default by ANSYS) I left AIM to run for a 10s Transient run.

Unfortunately, the steady and long time average transient values differs a lot – see the attached time history of the Cd/Cl values.

Questions:

-          can you please point out what can be the source of the problem here?

-          what turbulence or solver settings (e.g. other RANS models?) should I play around to get more reliable results?

Regards,

Csaba

[Brian Bueno]

steelxpert

I will have to inquire with our fluids experts about this.  There are several validation models you can test located here.  There are about a dozen or so fluid flow examples.

[info]

Dear Brian,

Thank you for your response, I'll check those examples if they can be relevant to my study.

One corrigendum to my original post: the transient chart shows the drag and resultant coefficients instead of drag & lift. Here I uploaded the update.

I am looking forward to the opinion of the fluid experts.

Regards,

[Brian Bueno]

steelxpert 

Here is the response:

Without going through the results, this analysis is using a turbulence model that is not appropriate for the application. KE does not do well with separation of flow as the fluid goes around an object. We recommend using SST, in either steady or transient. Try rerunning these models and then compare the Cd/Cl values.