Incoherent mesh on edges - problems with meshing

[haroon.ashraf]

Hi

I am creating a detailed FE model for critical connections in a road bridge, using shell/panels to represent steel plates. Having created the geometry, I am having issues a) creating any sort of decent mesh and b) incoherent mesh on edges issue c) applying my loadings as node point loads in the local panel direction ( have to guess the angle to the global axis - you will understand when looking at the applied loads)

I've tried following a few threads with similiar issues, but can't seem to resolve these in my model - I suspect the mesh issues might be related to the complexity of the geometry. Would appreciate if anyone can have a look?

Link to model

https://we.tl/t-fkG0tEV18N

Kind regards,

Haroon

[Simau]

Hi @haroon.ashraf 

Here the model without meshing problems.

A type 3 instability is remaining du perhaps to applying load at nodes 7 and 731

[haroon.ashraf]

Hi Simau

Could you broadly explain the steps that you took to rectify the issue, as I will probably have similar issues on other models?

Thanks.

[Simau · Accepted solution]

Hi

1 - Delete existing meshing

2 - Delete remaining nodes except those with loads

3 - Check if in your Job preference you have this box unchecked

4 - Remesh with

[haroon.ashraf]

Is there a way to stop my nodes which have loads on them being deleted when deleting local mesh for panels?

i.e. I was planning to use local mesh delete to delete all nodes, then select only my nodes with loads on and invert select nodes to delete every node (except those with loads on) and then mesh from there.

Is there a way to define these nodes as being critical to the calculation model such that they are not deleted/moved etc.

[Rafal.Gaweda]

Hi @haroon.ashraf 

---

Is there a way to stop my nodes which have loads on them being deleted when deleting local mesh for panels?

Apply supports to them before deleting mesh....then delete these supports before calculations

[haroon.ashraf]

Hi

I am having similar incoherence issues on a larger scale model - the same settings (have tried playing with the parameters a bit) don't seem to work. Also have a seperate model issue, can't seem to fine where this exactly occurs.

Would appreciate if anyone could have a look? https://we.tl/t-abtEIiswOX

Thanks in advance.

[haroon.ashraf]

Also - my model is built of panels modelling steel plates (to represent steel sections on a road bridge structure) - the intention is to view extreme fiber stresses at the location of connections to understand if they require additional stiffener plates.

How do I view extreme fiber stresses in the panels? Is this what the stresses option in the detailed tab of maps on panels shows?

[Rafal.Gaweda]

Hi @haroon.ashraf 

I would star with fixing geometry first (properly connecting panels)

Some locations indicated below

[haroon.ashraf]

Hi @Rafal.Gaweda 

To correct some of the geometry issues I used the vertical mirror function on the panels, so that I could extend them and then cut them at appropriate intersections. However, I hadn't realised that I still had mesh generated, hence when doing a vertical mirror, it also copied the finite elements which wern't removed when I cut the panels using cut by plane.

Any ideas on best way to remove them? (ps. how do I add a screenshot using windows printscreen function to these posts?)

[Rafal.Gaweda]

Hi @haroon.ashraf 

I am am sorry I did not get Your explanations.

I would use "Cut by Plane" and "Detailed Correct" options.

[haroon.ashraf]

Hi @Rafal.Gaweda 

I've attempted to correct and make the geometry fit as well as possible, noting that the geometry for this structure is anything but straight forward. However, I am still getting incoherent mesh issues and separate structure errors/warnings - I have attempted to use detail correct options, but can't seem to fix these issues.. Can't quite work out where it doesn't think the structure isn't attached either. I attach a link to an updated model which includes my geometrical 'corrections'. https://we.tl/t-V3nOpRzWbm

Would appreciate if you can take another look?

[Rafal.Gaweda]

Hi @haroon.ashraf 

Your model still needs a little bit more attention.

Correct geometry of panels  1568to1579. Connect them properly to each other and to top and bottom "chords".

(Locate their edges on "chord" edges.)

Example: