[diamondx]

according to a lots of posts in the forum. 0.1 can do the job... how many elements you got there. can you share your projet, i'll see if i can locate the bad elements for you. dont share the .uns file. its size is very big, share just the blocking

[Far]

check the quality in fluent. You should have orthogonal quality > 0.01. Other metrics are angle (in ICEM) which should be > 18 deg.

But you need to try, if solver does not mind the mesh quality of 0.1

[Lukas_ZH]

I hope these are the correct files. Right now, the set curve parameters are rather high. But I couldn't find guidelines for the necessary size of the elements for viscid/inviscid case. For the inviscid case I would need to create smaller elements around the wing. Could I do that with the pre-mesh paramters and inflation of the surface or is there another possibilty (easier)?

Best regards,

Lukas

[diamondx]

.tin file is missing ...

[Lukas_ZH]

I renamed the files manually in the folder. Otherwise the original name was INVISCID for all of them, in case there some troubles appear due to the name.

Best regards,

Lukas

[diamondx]

As you can see in the image above, bad elements are the results of the abscence of and ogrid inside the airfoil. I think these nodes will not be used for computation, i guess it's ok.

If your pc is not too strong to show bad elements, one thing you can do, unselect the premesh feature, like that you will only see the bad element. Another tip, use index blocking, and select premesh qualities to view the bad elements just inside the block selected in the index blocking. when number is reducing pc is more capable of showing them.

[Far]

Originally Posted by diamondx

according to a lots of posts in the forum. 0.1 can do the job... how many elements you got there. can you share your projet, i'll see if i can locate the bad elements for you. dont share the .uns file. its size is very big, share just the blocking

http://www.cfd-online.com/Forums/ans...-icem-cfd.html

[Lukas_ZH]

I decided to delete the blocks inside the airfoil. The pre-mesh statistics are within an acceptable range, but the statistics gained from the unstructured mesh don't seem that well anymore.

Are the statistics from the pre-mesh histogram more relevant than the ones from the unstructured statistics? If not, how can I try to get rid of the bad elements?

[Far]

Originally Posted by Lukas_ZH

I decided to delete the blocks inside the airfoil. The pre-mesh statistics are within an acceptable range, but the statistics gained from the unstructured mesh don't seem that well anymore.

they should be same

[Lukas_ZH]

It works by now, though I can't say what I did differently. Thought everything stayed the same except for the sizing of the edges.

Best regards,

Lukas

[Far]

Originally Posted by Lukas_ZH

It works by now, though I can't say what I did differently. Thought everything stayed the same except for the sizing of the edges.

Best regards,

Lukas

This might due to different premesh and mesh. In other words you have changed the parameters for premesh but did not convert it into unstructured mesh.

[Lukas_ZH]

Do you by chance know a source where I can find grid spacing (best practice) or resolution guidelines for an airfoil if I want to perform an inviscid simulation with Re=10^6 and Ma<0.3 ? I was only ale to find suggestions for y+ in the viscid case, but no suggestions for element sizes in the rest of the domain and how the domain should be in size related to the object around which the flow is supposed to be simulated.

Best regards

Lukas

[Lukas_ZH]

So I set a few parameters (linked edges, element number, spacing laws). Since I am new to the whole simulating enviroment I cannot estimate if that grid is useful for an inviscid calculation in Fluent.

Thanks in advance.

Best regards

Lukas

[Far]

Although I have little experience in inviscid simulation, but I can recall something from memory.

1. There is no strict requirements in the normal to wall direction i.e. Y+.

2. I feel that the mesh should be fine enough in the areas of abrupt changes either in geometry or flow. I mean you should resolve the mesh in the region of shock waves, expansion corner and also in sudden change in geometry contour and sharp corners.

3. As a rule of thump, the inviscid mesh can be of 1/10th order to that of viscous mesh. (this is from my experience while reading the literature and my friends running the Euler solution , please correct me if it is wrong)

[Lukas_ZH]

Thanks once more

Best regards

Lukas