Learning Forum

[Purnima Janjanam]

Hello

I am trying to generate a mesh for a 3d ring wing and the mesh quality is not good enough near the trailing edge. It has a skewness of 0.96 and Max aspect ratio of 20. I used body of influence, face sizing, edge sizing. I also tried sizing smaller near the trailing edge. What else can i do to improve it?

[Rob]

Have a careful look at the mesh in that region: what is happening to cause the poor mesh? This comes up fairly regularly, so also have a look in Search to see what's been covered in the previous threads. 

[Purnima Janjanam]

Actually, I am able to get good mesh quality with all the quality parameters. But when I run it in fluent the results are oscillating. Can you tell me why is that happening and how can I prevent it?

[Rob]

Please post some images. I suspect I know the problem, but images will confirm things. 

[Purnima Janjanam]

 

Here are some images of mesh. It’s an elliplical 3d wing. If you need any other images let me know.

 

[Rob]

You need to review the mesh at the back of the airfoil - the jump in cell size is excessive. I suspect the cell skew isn't good either. 

The other potential issue is flow related. Do you expect a single result or for the jet from between the two airfoils to oscillate?

[Purnima Janjanam]

The mesh goes like this somewhere in the middle even when the sizing settings are set to hard. Can you tell me what can i do to make it better. 

If I decrease size the number of nodes will increase and my cluster cannot handle it to run the simulation. so I'm trying to limit my number of nodes by limiting the size.

I expect the flow between to airfoils to oscillate. But I was trying to find a single result if I could get one. This is more of experimenting to understand what's happening.

[Rob]

Have a look at what sizing you're using. If you're using a mix of hard edge/surface control and size function then odd things can happen. 

If the jet is likely to oscillate you may find the residuals don't drop for a steady run. That's why we use monitor points. 

[Purnima Janjanam]

What are monitor points?

[Rob]

They're points we use to monitor flow data. Monitors in general will be covered in the early tutorials in Fluent's Help materials. 

[Purnima Janjanam]

1. These are the sizings I used. All sizing are hard. Can you suggest any changes here.

2. The oscillations for this mesh look like this. Do you think these are going to oscillate forever? I expect the flow to oscillate for this design. So can I conclude by saying these are the results for this design? 

3. The results below are for another design. These oscillate at 3e^-5 (different from the other where they oscillate at 1e^-3). My converging conditions are 1e^-6. Do you think these will converge if mesh is more fine or will they also oscillate like the previous one?

4. Can you give me a breif idea about how to plot monitor points?

[Rob]

Those will almost certainly oscillate forever. Hence monitor points to see what the effect is on the solution. You MUST review all the solution data and not just rely on residuals to tell you what's going on: that should be covered in the Fluent courses in Learning as they're fairly close to what I teach when training commercial users. 

[Purnima Janjanam]

Can you kindly share me a link of this?

also, my results for drag and lift oscillate the same way. Do you think that could conclude the results (similar to monitor points). 

[Rob]

https://innovationspace.ansys.com/learning-library/?p_type=learning-library&discipline[]=Fluids

If lift and drag are oscillating that's an indicator that the flow isn't stable. However, you need to review monitors (of which lift & drag are two), residuals (oscillating or noise at around 1e-2) and images. From what you've shown the result may be transient, but without a lot more images I can't be sure.