[lcbarreto]

Hi guys,

I am simulating the flow over an aircraft with slotted flap and some complex components. I am trying to do a prims/tet mesh on Ansys meshing. All of my mesh qualities are ok except for the orthogonal quality. I can't reach the OQ criteria at the flap tip unless I increase my mesh until over 50 miliions cell, which unfortunately for me is impracticable. My min OQ is around 5e-3 in 3000 cells, does anyone knows if that can have a really big impact on the simulation?

For complex geometries like this, which mesh software do you guys use?

Thanks in advance.

[shereez234]

Originally Posted by lcbarreto

Hi guys,

I am simulating the flow over an aircraft with slotted flap and some complex components. I am trying to do a prims/tet mesh on Ansys meshing. All of my mesh qualities are ok except for the orthogonal quality. I can't reach the OQ criteria at the flap tip unless I increase my mesh until over 50 miliions cell, which unfortunately for me is impracticable. My min OQ is around 5e-3 in 3000 cells, does anyone knows if that can have a really big impact on the simulation?

For complex geometries like this, which mesh software do you guys use?

Thanks in advance.

Personally I use ICEM or Pointwise. Pointwise's TRex Feature is amazing but in ICEM also you can do the same thing by doing AutoBlock surface meshing.

By complex Aircraft I suppose you mean something like DLR F11 High Lift configuration. I have created blocked meshes for this configuration using ICEM's blocking techniques (structured).

Although I agree with you that Unstructured meshing is easier to generate on complex aircraft meshing as blocking is hard. The orthogonal quality is very important in my opinion.

When you import your mesh to ANSYS fluent (I am guessing fluent is your solver) what's the quality you get? if it is above 0.01 then you should be okay as long as you take care of the numerical setup and under relax your solution so it does not diverge.

Describe to be more or us more about your meshing methods? is it bottom up approach or top down approach? and how many elements do you have on your surface mesh? Do you use density regions? Did you laplace smooth your surface mesh before you created prisms boundary? Did you use an overall global smoother to improve global mesh quality?

[lcbarreto]

Hi Shereez,

Thanks for your answer.

Originally Posted by shereez234

When you import your mesh to ANSYS fluent (I am guessing fluent is your solver) what's the quality you get? if it is above 0.01 then you should be okay as long as you take care of the numerical setup and under relax your solution so it does not diverge.

When I import my mesh into fluent I get the same min OQ that I got from meshing (5e-3). I was able to creat a mesh with minimum OQ of 0.01 however it had 50 millions cells and when I try to run my fluent closes due to lack of memory.

Personally I use ICEM or Pointwise. Pointwise's TRex Feature is amazing but in ICEM also you can do the same thing by doing AutoBlock surface meshing.

I have no experience with either ICEM nor Pointwise. I guess I will have to start studying ICEM considering that the company that I work for has the license.

Answering your questions.

is it bottom up approach or top down approach?

I am using the advanced front option and pre inflation wich consists on a bottom up approach.

how many elements do you have on your surface mesh?

My surface mesh is has around 700,000 elements.

Do you use density regions?

I am using 3 Body of influence one around the flap, another around the whole wing and another one for the whole aircraft.

Did you use an overall global smoother to improve global mesh quality?

Fluent Has an option to smooth the cells in order to get a better skewness, I used it as High and a skewness limit of 0.9.

Did you laplace smooth your surface mesh before you created prisms boundary?

I've never heard of it, I wil look into it.

Additional informations:

The elements with low OQ are inside my inflation. It consists on 33 layers with a 1.25 Growth rate for the airplane fuselage, 1.22 for the wing and 1.18 for the flap.
I have a lot of local sizing refinements which I think are irrelevant for this discussion.

I don't think I will be able to get a better mesh quality inside ANSYS meshing, unless I increase my mesh size substantially which I won't be able to. I thought ICEM wasn't a good tool for complex 3D geometries, I will look more into it.

Thanks Again for your help

[shereez234]

It is my bad that most of the terms such as density region, laplace smoothing belongs to ICEM and not ANSYS meshing. I apologize for that. However, you seem to intelligently understand what I mean and your method of meshing is most appropriate.

If the Orthogonal quality is low inside the boundary layer then it is definitely to do with the surface mesh. One thing you can do is to convert the tris in the trailing edge to quads ( that is if the trailing edge is blunt). Remember the volume mesh quality for both boundary and beyond boundary is dependent on surface mesh on bottom up approach which is why before volume mesh generation surface mesh quality should be high and should be smoothed atleast a few times.

and ICEM is very powerful for complex geometries too if you know how to use it.
Thanks
MS

[lcbarreto]

The surface mesh seems to be ok, it does have high quality and smooth transition throughout the body.

My problem is specifically on the small gap that I have from the side of the flap to the wing (I am simulating the flap as a separated solid from the wing). This gap is really small. I think that when the inflation starts to grow in order to place all the prisms in the gap, it compromises the orthogonal quality of the elements. I have reduced the number o layers and the OQ has increased but I don't think that this is a good solution.

[shereez234]

Originally Posted by lcbarreto

The surface mesh seems to be ok, it does have high quality and smooth transition throughout the body.

My problem is specifically on the small gap that I have from the side of the flap to the wing (I am simulating the flap as a separated solid from the wing). This gap is really small. I think that when the inflation starts to grow in order to place all the prisms in the gap, it compromises the orthogonal quality of the elements. I have reduced the number o layers and the OQ has increased but I don't think that this is a good solution.

Can you post a cut plane of the mesh zoomed in to the problematic area

[lcbarreto]

Originally Posted by shereez234

Can you post a cut plane of the mesh zoomed in to the problematic area

I am sending some pictures attached. I tried to explain as much as I could. It is hard to visualize the mesh but the bottom line is that there are to types of cell with low OQ:

1) the ones in the slot are at LE of the flap. These I can get rid of by refining this area.

20 the one at the gap between the side of the flap and the wing, especifically at the side edges of the flap. For these I tried to do a edge sizing on the side edges of the flap but I had to use a really small cell size wich increased a lot my the size of my mesh.

Sorry I could not explain in a better way.

Thanks again.

[lcbarreto]

I forgot to attach the files on the previous post. I am sending them now.

thanks