[afzaal]

Dear all,

I am trying to build mesh in an axial turbine for unsteady calculations using AutiGrid5 8.6-1, please let me know the best size and the quality criteria for this particular application.

Best Regards to all

[DarylMusashi]

Dear Afzaal,

after the 3D mesh is generated a grid quality report is presented, which you should study carefully. I am going through the columns one by one now and naming a rough estimation of the desired ranges:

Negative cells: 0
Number of Grid Levels: >=3
Minimum Skewness: >15
Maximum Aspect Ratio: <15000
Maximum Expansion Ratio: <2
Spanwise Angular Deviation: <30

To achieve a low Maximum Aspect Ratio many cells might be necessary, if you wish to create a fine mesh in the boundary layer.

The question about the "best" size of the mesh cannot be answered...as many cells as needed and as few as possible. You should think about 300.000 to 800.000 cells per blade row to get an appropriate mesh. The "row wizard" will assist you concerning this question, I hope.

[afzaal]

Dear Sir,

I am a new user of NUMECA, this is the first mesh i have generated using AutoGrid. I have attached the mesh quality report for your consideration. Please view and comment.

Best Regards,
Afzaal

[afzaal]

Dear Sir DarylMusashi,

My mesh do not meet the criteria defined by you in terms of minimum skewness & maximum expansion ratio. Please give me some guidelines to improve these two parameters.

Best Regards,
Afzaal

[DarylMusashi]

I think row 1 is okay, you should try to improve row 2. In the grid report overview (which you uploaded) you can simply click on row 2 to see the distribution of the cells concerning the quality criteria. How many cells are problematic there?

Your "Minimum Skewness Angle" (=minimum orthogonality) might be problematic in row 2. You should try to improve your mesh in row 2 concerning this. At first you need to find out where these bad quality cells are.

1) Check B2B layers
You should check your B2B meshes at different spanwise positions at first. Please see the attached picture where to find the dialog. I suggest to check the mesh at hub (0%), shroud (100%) and mid-span (50%). Just enter 0, 50 or 100, press enter and click "generate B2B mesh" in the upper right corner to update your B2B view.
After you generated the B2B mesh at a certain span position you can check your mesh visually, maybe you see some critical regions with skewed cells immediately. Secondly please go to "Grid" -> "Grid Quality" to check the quality of the selected B2B mesh. Please click "Show Chart" and "More Info" (to see the minimum and maximum values). You can enter the range, you are interested in (e.g. 0-25 concerning Orthogonality) and click "Display all cells". Now all low quality cells concerning Orthogonality are displayed.

2) Check flow paths control
Which values did you use in the Flow Path Control Dialog? Please see the attached picture where to find it. The values you see in this picture are a good first attempt. As your spanwise expansion ratio is below 2 57 flow paths are enough. The cell width at hub and shroud = 1E-05 is in most cases small enough to resolve the boundary layer. You should check your y+ values at the blades after your first calculations.

3) Check your 3D mesh
After you generated your 3D mesh please make it in the lower right part visible by clicking "View" -> "View 3D Solid mesh". Please click at the lower right window (a red frame is around it, then). Now click "Grid" -> "Grid Quality" again. You can inspect block by block or check the whole mesh by entering block "0" in the upper part of the grid quality window now. In the 3D grid quality dialog you can check your mesh for angular deviation, too. This value should be as low as possible, as it represents the angular variation of two neighboring cells.

[afzaal]

Dear Sir DarylMusashi,

Thank you very much for your detailed guide lines, i really appreciate your dedication. I will follow the guidelines provided by you and let you know the results. I will need a lot more assistance from you for this particular case study.

Thanks & Best Wishes,
Afzaal

[Hamidzoka]

Dear all;
Just one general point to add.
Setting criteria for mesh quality ,for example skewness, like what daryl states is of course useful. But these are often limited to univerisity basic research cases with relatively simple domains. When real industrial cases are delt , it is rarely possible to keep mesh qualities high enough . As a matter of fact, In these cases the idealized criteria will turn into some minimum requirements. So talking about qualities always depends on complexity of domains.

Regards