Strange Slice Behavior - Boundary Connections + FEPolyhedron

[kostasp]

Hello All,

I am using boundary connections for FEpolyhedral zones and boundary connections to visualize data generated in parallel. For the following case:

https://docs.google.com/file/d/0B2mFLgZ ... sp=sharing

tecplot 14.0.2 generates a strange messed up x=0.6 slice for the Ci field. The Ci field should change from 0 to 1 near z=0. For the x=0.6 slice, at the left as you view the slice from x-positive, an area of mixed contour values appear. I also noticed that the slice's mesh is a bit weird there. Could you please check if you can reproduce the behavior? Do you have any ideas of what might be causing this? I've checked the boundary connections near the area of the error and they seem fine. You can check each block separately by deactivating the other zones.

I've also noticed that tecplot generates discontinuous floods in boundary connections with cell-centered data. This a general thing (not just for the above case) and happens for every case I use boundary connections and cell-centered data. For example in the above case, for the Ci field, if you go at x=2.85 or 2.9 you will notice the discontinuity near the middle. If you visualize node data, say for example the Z coordinate the discontinuities disappear.

Do you have any suggestion on how to fix/improve the visual results?

Many thanks for you time!

Regards,
Kostas

[JerimiahLee]

Hello kostasp,

When a slice is taken, it will draw borders around each of the separate zones. You can remove these by turning off edges. Alternately, you can use the Edge pane in the Zone Style Dialog, and change the Edge Type to creases only, that way you can still see where some suspect areas are.

I am not sure I am noticing any strange discontinuities. Could you provide a screenshot of what you are seeing?

If you have current SMS please feel free to contact Tecplot Support directly regarding this.

[kostasp]

Hi Jerimiah Lee,

Thank you for your answer.

I've noted in each pics name the x=.. of the slice that this appears. If you don't get the same behavior please let me know.

The slice with the strange mesh:

https://docs.google.com/file/d/0B2mFLgZ ... sp=sharing


For the flood discontinuities near the boundary connections:

https://docs.google.com/file/d/0B2mFLgZ ... sp=sharing

https://docs.google.com/file/d/0B2mFLgZ ... sp=sharing


Here is a very simple case, a cube with four boundary connections and cell-centered variables visualization.

https://docs.google.com/file/d/0B2mFLgZ ... sp=sharing


The cube is cut in four. The boundary connections are on the x-faces. I'm visualizing the distance of each cell's center to O. The flood on slice z=0 is discontinuous near the boundary connections.


You can get the plt from here:

https://docs.google.com/file/d/0B2mFLgZ ... sp=sharing


Thanks!
Kostas

[JerimiahLee]

Kostas,

Tecplot does have some issues when crossing zone boundaries. Could you explain a little bit to me how you created these PLT file?

[kostasp]

Hello,

We produce lists of zone headers and data. Both lists are catenated into a single file. We either write ASCII files and use preplot, or write directly the plt. Both file were generated by writing directly the plt.

Kostas

[JerimiahLee]

Hello Kostas,

With a polyhedral mesh, there are known issues with taking slices. This is probably the culprit for the strange mesh you are noticing in your first pic. If you are able to create them as FE bricks, we would recommend using that style instead.

As for your second case, the discontinuity occurs because the data is stored in a cell centered location. It can be fixed by explicitly stating face neighbor connections between zones. Our Data Format guides goes into great detail about how to set up this neighboring.

[JerimiahLee]

Kostas,

When looking at the data for your second example, it appears that the boundary connections elements are not correct.

[kostasp]

Hi Jeremiah Lee,

Can you please provide me with more info? What is the problem you've notice ? Can you give me the cell's index and zone that you've found a wrong boundary connection.

Many thanks,
Kostas

[kostasp]

Hi Jeremiah Lee,

I'm probably missing something.

I've checked the cells one by one. Using the probe tool(face neighbor pane) for each connected cell I get the indices of "neighbor cell" and "neighbor zone". Moving to the neighboring cell and zone specified by the tool, I get (again using the tool) the "neighbor cell" and "neighbor zone" and they seem to match with the cell's index/zone that I began with. Isn't that the pattern that I should get after defining the boundary connections ?

The the nodes orientation in the connected faces from both sides seems to match (i turned the node index labels on). Please shed some light on this.

Thank you for you time!
Kostas

[JerimiahLee]

Kostas,

I double checked the data I had, and you are correct. All of your face connections are spot on. We have determined the issue and our developers will look into making a fix for a future release.

There are a few workarounds available, and they all involve changing the data from polyhedral or using nodal locations.

[kostasp]

Hello Jeremiah Lee,

Thank you very much for providing online support! We will interpolate the values to the nodes and we will visualize the cell-centered fields as primary cell value fields for now.

Kostas