Learning Forum

[Aras karimi]

Hi friends,

I am reading an paper where it explains how to simulation. The problem is to simulate a 3D airfoil.

The CFX solver is Reynolds flow 120,000 and the turbulence model used is   Gamma Reynolds-theta.

Now, in a part of the paper, he wrote the following text, which I don't understand what it means:

 

I think he said about temporal and spatial discretization. I do not understand it correctly.

I am simulating in Fluent. But I don't know how to apply the discretization described above in Fluent.

Please help me.

Thanks in advance

Regards

[Federico]

Hello, 

it looks like the paper is using version 12.1 of our software. Also, you mention that CFX is used, but you are using Fluent.

You can read about Discretization used in our latest release here: 24.3. Discretization (ansys.com)

[Aras karimi]

In order to be able to use the turbulence model mentioned in the paper, which is Gamma Reynolds-Theta, I chose the SST transition turbulence model in Fluent. Now, in the method part of the Fluent tree diagram, I don’t know how to put the settings :

The settings in the photo are the default settings.

Should the pressure-velocity coupling be coupled?

Also, what should the flux type and options of special discretization and transient formulation be?

Should the pseudo time method option be active?

 

According to the figure below, also in the viscous model section, I selected the model and did not change the tick of the following options. Are the following correct?

 

 

And finally, what number should be set for the intermittency value in the boundary condition of velocity inlet and pressure outlet? (1 is the default)

 

 

The problem is to simulate the flow on a three-dimensional airfoil that is in the transition regime and the Reynolds number is 120,000.

 

If you help me, I would be very grateful to you, dear Federico.

 

Regards

[Aras karimi]

Dear Federico,

I chose the SST transition turbulence model in Fluent. Now, when I put the time on the Steady, the settings of the method are as follows:

And when I put the time on the Transient, the settings of the method are as follows:

For simulation at low angles of attack, should I get a Steady solution or Transient?

I once got a steady solution at the angle of attack of 7 degrees with the SST transition model, and the residuals and lift and drag were oscillating and the residuals did not go below 1e-03. Meanwhile, at the same angle of attack of 7 degrees with the K-Omega SST model, I got a Steady solution, and the residuals are reduced to 1e-13 and the lift and drag are smooth. But the results of the K-Omega SST model are very different from the experimental results, so I have to change the turbulence model to the SST transition.

I want to plot the performance curve of the airfoil at different angles of attack. For low attack angles, should I get a transient solution considering that the residuals of the steady-state solution oscillate?

Please tell me the appropriate settings of the method section for both steady and transient modes.

Regards

[Federico]

Hello Aras, 

there is a lot in your question. To compare your results with experimental data you need to closely match the conditions for that setup (e.g. the inlet turbulence conditions) and those results will also be influenced by your mesh resolution.

We have lots of resources in the Ansys Innovation Space of simulation lift and drag for airfoils. I'll invite you to consult those resources to get started.

Compressible Flow Over An Airfoil | Ansys Innovation Courses

Flow over an Airfoil using Ansys Workbench | Ansys Courses

Flow over an Aircraft Wing - ANSYS Innovation Courses

[Aras karimi]

Federico, I raised the issue of non-validation of my CFD results with experimental results in the topic (/forum/forums/topic/validation-of-airfoil-lift-and-drag-coefficients/#post-319795).

My mesh has good resolution and quality and its Y+ is below one. I used the K-Omega SST model and obtained its performance curve up to 10° angle of attack, which is as shown below:

 

I ran all simulations from 0° to 10° angle of attack as steady state and the residuals went down to 1e-13 and the CL and CD graphs were smooth. According to the picture above, there is a huge difference in the results. Because the Reynolds flow is in the transition region, I want to use the SST transition turbulence model.

Federico I have been involved in the validation of my results for a very long time and I suffer from this issue.

Please help me

Regards

[Aras karimi]

Hi Federico,

I checked and watched the above resources you mentioned. I still have a big difference in validating my results with experimental results. I even did some transient simulations to ensure the accuracy of my results, but there is still a big difference between my results and the experimental results.

I checked the solver setup hundreds of times to make sure I set everything up correctly. I initialize all simulations as hybrid.

Is it possible to change the initialization type to make a significant change in the results?

Regards