[hebush]

I am attempting to mesh a mostly-cylindrical domain which contains some interior geometry which complicates the meshing process. I have attempted to generate a hex mesh that captures the overall geometry and interior features via a series of "nested" o-grids.

The blocking now associates to the geometry as I desire. I have attempted to achieve the resolution and spacing I wish for my initial mesh by using the Edge Spacing feature in Pre-Mesh Parameters. The mostly works, apart from one region of the domain, where the nodes adhere to the spacing where the mesh aligns with the spacing and then "jump" to the outer edge of the cylinder. This creates very poor quality and negative volumes.

Is there something immediately obvious about my associations or other blocking settings that would cause this? If it's not something that can be fixed in blocking, I would at least like to find a way to systematically move these nodes or set something in pre-mesh so that they progress radially in a way similar to those aligned with the edges.

Apologies for any sloppiness in wording or misuse of terms. I'm quite new to this and am trying to learn as I go. I've attached some images of the blocking, the edges in question, and the nodes in question, but I can provide more information or images as needed.

Overall Geometry:


Overall Blocking:


Detailed O-Grid Blocking of Relevant Region:


Side View of Mesh of Relevant Region:


Top View of Mesh of Relevant Region:

(You can see above how, for nodes not in the direction of the four o-grid edges, the nodes "jump" to the outer surface of the geometry).

[hebush]

Still looking for any help that can be offered on this. I've attempted modifying my edge spacing on the o-grid edges and simplifying my blocking by merging blocks in the simpler parts of the geometry to no avail.

[hebush]

I have a small update and (hopefully) clearer summary of my problem.

It looks as if those problem nodes are propagating in a direction that would suggest they expect a straight, not curved edge.

I looked at the "projected edge shape" and "projected mesh shape" options but am still learning about them. I'll update further if I make progress with them, but I just wanted to note two observations:

1. The "projected mesh shape" view looks as expected. All of the edges project to a curved (circular) shape that is consistent with the geometry, except for, of course, radial o-grid edges.
2. For the "projected edge shape" view, however, there are some problems. Some edges seem missing. Are they projecting in the wrong direction and therefore sitting atop one another?

Images for the unprojected edges and list items 1 and 2 are included. If anyone has any thoughts, that would be great.

[hebush]

Update: I fixed the problem mentioned in my most recent post by un-associating edges and splitting all of my circular geometry into four curves, one for each edge. For some reason, this was the only way ICEM would properly associate all edges to the circular geometry.

Unforunately, this did not fix my node problem.