Learning Forum

[MaFri]

Hello all,
I am pretty new to ANSYS Fluent and have learned all by myself so I probally missing a lot of theoretical background - sorry for that.
I will start to descripe my project and my problems.

My task is to to simulate a gas mixer (2 inputs / 1 outlet). The design is already fix. Later on I will do a paremtic study. I already prepared everything for it. Like wrote a ironpython script in Space Claim and so on. But my actual problem is on the solver side. To be more precise my continuity residuals only go down to 10^-3. I am observing some physical properties with are stable on a specific value (pressure drop, massbalance). Due to the fact that i already tried a lot of different solvers etc. I assume that my mesh is the problem. I am not really able to produce a good mesh (orthogonality, skwekness, aspect ration). I already watched a lot of Tutorials, read instruction etc. but I am not able to produce a good mesh. My questions are now:

Is the problem that my continiuty Resudials dont go down really related to the mesh? (I am using double precision and second order settings)?

Do you have some tips how i could improve my mesh? (I think the problems come with the inflation layers and sharp edges / I want to reach y+ ~ 1)

Best regards and thank you in advance

 

P.S.: I had to clear generated data to smaller the .wbpz file. I hope thats not a problem

[Rob]

Mesh will play a part in solver stability: you're looking at both the cell quality and the mesh quality. Ie are the cells all OK and do you have sufficient resolution to capture all of the flow features and variable gradients. 

Then there are the boundary settings and models. These need to be used appropriately or you could be setting up a solution that can't work. 

Note, it's possible that everything is correct and you've picked up a transient instability in the flow. Review all the settings and investigate the flow field. 

 

As an aside, staff are not permitted to open/download files and attachments. Please post images in with the text. 

[MaFri]

Thanks a lot for your reply. First I will ask some question and then I will try all your answers as good as possible. Pictures will follow in the next posts.

Questions:

What is the difference between cell quality and mesh quality? Mesh quality is probalby related to orthogonality, skeweness etc.?

How do I know that I have sufficien resolution? I tried that i have more then 20 cells over the diameter of a pipe.

How do I know that i picked up a transient instability? (I'm not sure if thats the case for my example because my physic parameters are constant after some iterations - only continuity stays at E-03.

Answers:

   solver settings:

         solver = k-e realizable with enhanced wall treatment, energy equation on

         Input_1 = 7m/s, T=292K

         Input_top = 34m/s, T=312K

         Outlet = 0Pa, backflow T=302K

         Double Precision, Hybrid Initzialitaion, Second order Upwind

         Everything else I didn't change in the settings

I really appreciate your help. Thank you.

Best regards

 P.S.: It says that "Your post is currently awaiting moderation. Once approved it will appear on the site" if I want to upload pictures in the text.

[MaFri]

[MaFri]

[MaFri]

[peteroznewman]

MaFri, please delete the duplicate posts.

This site automatically puts some posts into an approval queue.  You just have to wait till the post is approved. It doesn't help to resubmit the same post over and over.

[MaFri]

MaFri, please delete the duplicate posts.

This site automatically puts some posts into an approval queue.  You just have to wait till the post is approved. It doesn't help to resubmit the same post over and over.

 

I've done it. Sorry for that.

[Rob]

Mesh quality means you have good cells (ortho quality etc; with tets I'd look at skew and keep it under about 0.95) AND enough cells to resolve the flow. Read up on mesh independence studies. 

To pick up transients look at the results and use monitor points to check on the flow: they're covered in the tutorials that ship with the documentation. 

[MaFri]

Thanks for your reply. I get under 0.95 skew but my problem is - like i mentioned -  the low ortho quality in the region with sharp edges (see posted picture).

I 'm already using monitor points to check the flow. And they stay at a constant value after some iteration. But like I said - my continuity residuals don't go beneath 1-E3. So my assumption was that I have to improve my mesh quality - which now leads back to my original question for some tips how to improve my mesh quality for my specific geometry. I hope you understand my point.

Thanks a lot and I really appreciate your effort

 

Best regards

[MaFri]

Any ideas?