SimScale CAE Forum

Optimizing mesh quality and Y+ layer formation on an FSAE car


#110

No way! i had already tests (which i think i deleted) without cartesian boxes. I had just increased the cells between layers t like 6 or 7 on one test and to 10 (the level distance between level 0 and my wing surface level) and i got the same results as you. Nothing showing up. Ideally my perfect mesh solution would have no boxes except for wake regions so that the cells could simply be reduced when they are a certain distance to the geometry. This would help a ton in terms of cell count and i could maybe if im lucky increase surface the refinement level on everything.

If this works as i envision it, along with possible absolute layering, i think this could work for my application needs!


#111

WAY BACK HERE WHEN YOU MADE THAT STATEMENT A WEEK AGO, I STARTED INVESTIGATING THE STRANGE ISSUE…

It did not make sense to me until I tried it and I thought I was stupid for not knowing why the mesh would not appear without the region refinements…

I finally discovered the Feature refinement trick and then I created that topic yesterday, where I tried to find out if there was a better way to get the mesh to ‘appear’ again… BUT it looks like so far the Feature refinement trick is working the best for me :slight_smile:


#112

The gradient between refinement level changes is controlled by the ‘Cells between levels’ HEX Parametric algorithm meshing parameter… I am using 2 instead of the default 3 for my CyberTruck mesh (saves some cell count on layering the tire treads) :wink:

Also, you need to discover the ‘Distance to surface’ region refinement… I am using it on the CyberTruck mesh now…


#113

Oh my gosh this is a game changer

I had known about this feature i just applied it to cartesian boxes. I never saw that it could be applied to the geometry. I feel so dumb, because this would have saved me a lot of frustration and mesh trials. Oh well, now a new round of meshes can be made and hopefully i can figure out how to get a good mesh.

Another side note about the pictures above. I do not like the leading edge on the element wing at all. It is extremely jagged. I am pretty sure that increasing the surface refinement level will not hlep here because the STL file size is too small. Currently I believe that it is at 0.013m but saving this again at a higher level will help. Another piece to this puzzle.

On a side note. I think i have made my decision, at least for wing surfaces, on the Y+ vlaue. After observing the meshes on the symmetry plane from the Rear wing, I have noticed that creating a boundary layer that has a max Y+ value over a certain amount, starts to warp the cells between the main and element wings. Shown below is from test 1.6 which is ok. if the boundary layer cells from both wings is too large Shown in seccond pic. Either the cells will distort too much or they will touch which i assume is no good.
image
image

However, using the distance region refinement on the geometry will allow for smaller cells in this region so maybe this distortion will not occur. The only downside of this is from what Dylan said about region refinements being a lower level then surface refinements and forcing the surface refinement down a level to whatever the region refinement was set to. I will be testing this as well.

As usual a lot to do and even more learned as the testing continues. Pumped to do this next round. I think im getting close.


#114

Personally I would make all your trailing edges square (not sloped) and make sure they are slightly wider than the level that you are refining your TEs to… Reduces jaggies…

Also, try to stick with EXACTLY 1 meter Level 0 cell size in all axes… I even generally make my BMB locations exactly on 1 meter steps…

Have a look at how I modified your top wing in that link I sent you to. It has fewer faces and all surfaces never get closer than 0.5mm from each other and all faces never get narrower than 0.5mm…

Just to be sure you understand the trick, your Feature refinement that makes the mesh appear with be to 0.5mm which is the TE thicknesses…

You also need to make the mesh finer on your leading edges to smooth them out:
image


#115

Can you give me a direct link to the mesh that this 2D image was extracted from :question: (I would like to see those weird cells in a 3D viewer):
image
.
.
.
.
Also, I have isolated another issue where meshing fails with a Floating Point Exception error when either an Inside Cartesian Box region refinement OR a Distance from surface face Region refinement is added around a geometry to increase mesh fineness around the vehicle…

Here is the project that shows the isolated issue and SimScale support is working in it now…


Porous Media Simulation Guide For FSAE Radiator
Porous Media Simulation Guide For FSAE Radiator
#116

Dale, this picture is from test 1.1. It is in the Rear wing test project i shared before but here it is again.

https://www.simscale.com/workbench/?pid=3614293863436765614&mi=spec%3Ab9873730-11a6-41b0-b076-c90c237cadc1%2Cservice%3ASIMULATION%2Cstrategy%3A14

I am working on more tests now and am making good progress, will show the results soon


#117

Great…
By the way, you can copy the browser link when you are on any tree item to send me right to the test 1.1 mesh, like this.

It is good to put direct embedded links everywhere you show a picture for the lazy helpers like me :wink:


#118

Ok Dale, i havent done one of these yet but i assume this is how it works …Just for you


#119

My preferred way to handle this is to split the airfoil surface so that the LE is a separate face at the yellow highlighted section so the you can mesh it at an EqualMaxMin level that is different than the less curved rear airfoil surfaces…
image


#120

I took a mesh clip half way out to the endplate and they are less squished but you should worry about squished cells, these are OK but later you would see bad ones in the mesh quality metrics at the end of the mesh log… (views at symmetry plane are always worse looking for layering collapses and squished cells):


.
.
.
and end of mesh log:


#121

thanks for the advice!

yea i am currently working on the squished cells by increasing the distance of the first region refinement level from the geometry.
Test 1.9 has level 9 at 0.001
Test 2.1 has level 9 at 0.005
Test 2.2 has level 9 at 0.01

I have also added decreasing levels as the distance from the geometry increases but im not sure if they are getting overwritten. Im going to finish my report of test 1.9-2.2 and report in a bit


#122

RW Test 1.9

Changes:

  • Suggested to change the bounding box to 1m
Level 0 (calculated) 1m
Level 1 0.5m
Level 2 0.25m
Level 3 0.125m
Level 4 0.0625m
Level 5 0.03125m
Level 6 0.015625m
Level 7 0.0078125m
Level 8 0.00390625m
Level 9 0.001953125m
Level 10 0.0009765625m
Level 11 0.0004828125m
Level 12 0.000244140625m
  • New boundary layer calculation to increase average Y+ values

0.001953m * X = 0.00182182m

0.00182182m /0.001953m = X

X = 0.932

Inflate boundary layer settings - layer calculation Y+ Value
Layers 3
1st 0.001078 Y+ = 50
2nd 0.0014014
3rd 0.00182182
Overall thickness 0.00430122 Y+ = 200 = 0.004311
Expansion ratio 1.3
Min thickness 0.001m
Final layer thickness (RATIO) 0.932
Relative 0.932
Absolute 0.00182182
  • Region refinements added as distance from geometry in addition to wake boxes
  • Trailing edges surface refinement level reduced
  • Wing surface refinement level reduced

Results:

Test 1.9 Pic 1: Cell squish between main wing and elements. Not many cells shown of level 9 region refinement at 0.001 - must increase distance from geometry.

Test 1.9 Pic 2: Cells are not even along wing surface - solutions could be adding feature refinement or unequal min max of refinement

Test 1.9 Pic 3: Shown again the surface refinement levels of the wing surface affect the boundary layer cells making unequal cell sizes - no bueno

Test 1.9 Pic 4: Trailing edge gurney flap boundary layer is well meshed

Test 1.9 Pic 5: Trailing edge of 1st element has jagged trailing edge

Test 1.9 Pic 6: Edge of endplate has some odd cells - solution could be unequal min max levels

Test 1.9 Pic 7: Layering chart from the meshing log. Thicknesses are good in some areas but need to be approved in others.
image
Overall results: Surface and region refinements need work

RW Test 2.0

Changes:

  • Exact same as test 1.9 but with feature refinement
    • 0.0004 level 11
    • 0.0009 level 10

Results:

Test 2.0 Pic 1: Region refinement level 9 is more prevalent even though the distance is not changed. Trailing edge boundary layer cells gone in main wing trailing edge. Squish is reduced slightly

Test 2.0 Pic 2: Gurney flap boundary layers gone, but cell refines to a lower level - most likely to surface refinements moving to lower levels

Test 2.0 Pic 3: Main wing trailing edge has many jaggies even though cell size is reduced near the edge

Test 2.0 Pic 4: Gurney flap has some bad cells

Test 2.0 Pic 5: mesh thickness log shows less coverage then test 1.9 without the feature refinement.

image
Overall results: Feature refinement allows transition to designated higher level surface refinements to show up. In test 1.9 the higher level trailing edge refinements (especially at the gurney flap) are not recognized

RW Test 2.1

Changes:

  • Distances to surface of region refinement increased
  • Feature refinement removed
  • Wing surfaces surface refinement changed to min lvl 9 max lvl 10
  • Endplate main surface refinement increased to lvl 8
  • Wake region refinement levels reduced
  • Wake region box smaller

Test 2.1 Pic 1: Increased region refinement level 9 to 0.005 gave better results then test 1.9. Increased distance of level 9 did not help with cell squish - solution, maybe this needs to be increased further to conjoin the two regions

Test 2.1 Pic 2: Same as pic 1 Showing how the surface and region refinements are related. It can be seen that in Test 1.9 Pic 1 with a region refinement distance of level 9 at 0.001m away from the geometry, this does not surpass the overall thickness of all three layers which results in 0.0043m This is why the level 9 region refinement doesn’t show up well on the mesh clip and seems to go directly to the next lower level 8, which was set at 0.005m. In this picture below the level 9 region refinement distance was changed to 0.005, and level 8 to 0.01m which is why there is better coverage outside of the boundary layer. Understanding that with distance from geometry region refinements, the first level distance value should be larger then the overall thickness of the boundary layer.

Test 2.1 Pic 3: Cell squish is still apparent even with increased level 9 region refinement distance.Also notice that the boundary layers near the trailing edge have disappeared. - solution could be to make unequal min max levels for trailing edges

Test 2.1 Pic 4: Endplate edge cells are better and wing surface cells are also better.

Overall results:

Boundary layer depletion around trailing edge - no good

Cell squish is not gone - solution is to increase the distance of level 9 region refinement

RW Test 2.2

Changes

  • Increased region refinement distance of highest level - level 9 = 0.01
  • Increased trailing edge level to min 11 max 12

Test 2.2 Pic 1: Increased level 9 distance did not remove cell squish

Test 2.2 Pic 2: Trailing edge boundary layers gone still between 1st and second element

Test 2.2 Pic 3: Trailing edge has jaggies just like test 1.9. I think the unequal cell sizes on the surface refinement is to blame. Not sure how to fix this

Overall Results: Need to find new solutions to problems


#123

Going though what i have done again. I feel that my attempts to get a really good mesh using different refinement settings might be towards its end. I think if i understood the mesh settings better, especially the snap controls and layer adding controls, i would be able to better manipulative the cells by changing these settings. This is where i need more guidance


#124

Yep, that is why I would likely do this.
.
.
.

Would be fixed by squaring off TEs and making sure TE is 0.5mm like I said here:

.
.
.

That is tough I have same issue on my CyberTruck tire tread, I tried changing all ‘Snap control’ parameters but no luck yet…


#125

Don’t worry about squish if mesh log does not complain at the end of it…
.
.
.
.

My advice now is make the CAD geometry file edits I suggested and try again…

EDIT: When you try again (just in case you are not using ‘copy mesh settings from’), just make a new sim with the new Geometry, then on ‘mesh’ dropdown, select ‘copy mesh settings from’ and select the mesh that you like best so far, say 2.2 if that is your best… That way your new edited mesh refinements will not overwrite any other mesh… I use a new sim for EVERY successful (or error-ed) mesh, I REALLY suggest you continue to do the same…

Golden rules:

  1. Best CAD file first.
  2. Then make best mesh from it…
  3. Then simulate…

Don’t be afraid to go back to CAD and fix regularly…


#126

I found an OK workaround for the ‘floating point exception’ issue… Support has not come up with anything yet :frowning:


#127

Ok dale im going to go back and change the CAD files. It was still worth it to see how everything works together. I have a lot of CAD work to fix on other parts of the car so i will be doing that at the same time. I might be able to have most of it done by tomorrow. Ill do another wing run when im done then hopefully a simulation of that will give me a good Y+ average overt the surface.

I also want to know how you got this measurement tool. Is this within simscale or your own setup?
image

After that I have to learn the Paraview post processing. So i can make the awesome Y+ histograms. SOO much to do. Ill keep chugging away


#128

Rear Wing 2 Test 1.0

  • All surface refinements reset to equal min-max levels

RW 2 Test 1.0 Pic 1: Trailing edges still looking good

RW 2 Test 1.0 Pic 2: Element and gurney flap boundary layers also looking good. Cell squish still an issue

RW 2 Test 1.0 Pic 3: Trailing edges very clean - Dale had it right with CAD geometry being the problem (not surprised he knew this)

RW 2 Test 1.0 Pic 4: Gurney flap also looking very good. Still funky cells on endplate edge though. I am not concerned about this, the wings are most important

RW 2 Test 1.0 Pic 5:
image

Overall Results: While I am very pleased with these results, and would consider stopping here in terms of overall mesh quality. I would still like to have more understanding and control over certain areas such as:

Surface refinement control

  • Even though i have given higher surface refinement designations to edge or small face areas, the finer mesh doesnt seem to be showing up. The only test where my higher level surface refinements seemed to work is when I add a feature refinement to the mesh. I believe this is because it acts as a bridge for cell reduction, just as the”Cells between levels” setting does in mesh control. In this picture RW 2 Test 1.0 Pic 4: and in many others it seems that the whole mesh is refining to level 9 only. This is also shown in the log where my test with a feature refinement is the only test that has cells up to level 11

The new problem of mesh squish

  • I am not sure how important this plays into accurate results. I would like that the cells between elements form together better and i am not sure if this has to do with region refinement levels or distance, or if it has to do with mesh control setting such as face twist, non-orthogonality, or skewness.

#129

Moving onto simulation settings. I am starting to read up on the correct numerics for my situation. I have read through some of your old threads such as Initial Values for turbulence coefficients and continued with the paper from Spalart and Rumsey. I am trying to start at the beginning with understanding the Initial Conditions settings as the FSAE tutorial said to use do the following changes:

image image

Based on my brief reading on these topics, this recommendation from the Spalart and Rumsey paper gave a good explanation.

B. Boundary-Layer Levels, and Device Reynolds Numbers

Another very concrete consideration in aerospace applications is the gap regions between different elements of a wing. These are normally designed to allow a “potential core” between the two boundary layers. It is all too easy to set the ambient νt so high that this potential core is smeared, and to be ignorant of this fact, deeply hidden in the flow field. Yet, this will make CFD inaccurate, and sensitive to ambient values.

  1. As i understand it, the initial condition values for (k) and (w) are for the ambient air in the bounding box. These values can be set so that the ambient air has laminar or turbulent characteristics, which through the paper, suggests that there is an effect on boundary layer transition, simulation convergence ability, and effects slot gap boundary layer measurement accuracy.

A. Free Decay

As mentioned earlier, for large decay rates, the coarseness of the grid far from the body may make the solver unable to support an accurate computation and the decay can be grossly underestimated.

  1. Would this be a factor when our level 0 mesh size is 1m?

Using significantly lower values for freestream ω, however, the decay rate is dramatically reduced, and CFD predicts the levels with greater precision on all three grids (lower curves).

  1. Does a low (w) value, even with a coarse mesh from #2 mean that precision is not a problem?

III. Ambient Values

We also consider the common situation in which the ambient values are picked to trigger turbulence in all boundary layers without local tripping, except perhaps at very low Reynolds numbers.

  1. Should I be purposely picking a turbulent ambient flow?

C. Recommended Values

the SST model can encounter a particular problem when the ambient value of νt/ν is large and ω is set too small. The main problem with this erratic behavior is that convergence can stall. Figure 3 suggests that values of ωc/U up to 100 would be more than safe, so that the useable range is not too narrow.

To summarize, the following ambient freestream turbulence values (in the region near the body) are recommended for general use for most typical external aerodynamic applications with two-equation turbulence models: k/U2 = 1 × 10−6 (or Tu near ≈ 0.1%), εc/U3 = 4.5 × 10−7, ωc/U = 5. These values yield an ambient eddy viscosity ratio of νt/ν ≈ 2×10−7 ×Rec. Thus, for example, for Rec = 10 million, νt/ν ≈ 2; and for Rec = 1 million, νt/ν ≈ 0.2. Again, these ambient levels are not achievable if the turbulence variables are allowed to decay naturally according to the turbulence equations. One easy way to achieve these ambient levels in practice is to impose them as both inflow and floor values.

  1. From what i have read, an FSAE car operates under low turbulence intensity of less then 1%. My (k) and (w) values are dependent on desired laminar or turbulent ambient air. turbulence model plays a factor on these values. Having multi-element wings play a part. Cell size before and i assume after the model have an effect on the consistency of results.
  • My (k) value should be be under 3.

  • My (w) value is should be close to 100 using the K-W SST model.

Am i going too deep into these values? Is this really that important for my results?