I would advise against using Nonlinear adaptive mesh region here. This is really for models that become highly deformed during analysis, and so the highly deformed regions will have bad mesh so it remeshes these. The nonlinear adaptive meshing can take experience to get it to work well, and it has a large set of limitations. Your description also shows bad elements at the start of analtysis, not during analysis.
This model really looks best done as a shell model. Whether to treat as a thin body with solid mesh or shell model is a tradeoff. You either spend your time trying to get a good mesh or you spend your time on geometry midsurfacing. Either way takes some work.
- Shell model. Use midsrface tool in SpaceClaim or whatever CAD tool you use. Mesh should be easy on shell bodies, and the soluton is accurate for a wide range of element aspect ratios, and runs much faster than a solid model, taking less memory. Then set up contacts in Mechanical with the option "shell thickness effect" and be mindful of the top/bottom side choosen for contact/target sides.
- To mesh as solid without too many elements and get good quality, and especially to get multiple elements across, you will need to ensure it does a sweep mesh. Many of these bodies look sweepable, and it will do sweepable by default it you do not set a different mesh method and you don't set element sizes to large or different on faces it would use for source and target. But you can manually set a mesh method on these bodies and set to sweepable. Also, multizone method can be thought of as a more intelligent sweep which can handle some non-matching face imprints between source and target sides. If bodies are not sweepable you can split off the features that prevent this in SpaceClaim as separate bodies. You can right click on "Mesh > Show > Sweepable Bodies" in Mechanical to let you know. Element aspect ratio is a concern in thin solid models. As a rough number, once they get to around 8 times the length on another side, that's about the limit you want for accuracy, and the optimal aspect ratio is 1. This means you need to have a lot of elements when solid meshing thin bodies, and using multiple layers across a thickness results in a very large mesh. Shell models do not have this limitation. If you use a sweep method with "Src/Trg Selection" set to automatic or manual thin, you can set the "Element Option" to "Solid shell." This can handle a gradient across the thickness similar to multiple elements across the thickness, but obviously with much less elements. The "Solid shell" option only allows one element across. However, its range of aspect ratios for accuracy is much less than shell elements, with a max acceptable thickness over span ratio of around 0.5. Thin model methods are accurate when elements are "thin" which rough means they must be thinner than their span distance.
My recommendation for thin models is that it is worth the time spent in geometry midsurfacing to model as shell bodies.
You are navigating away from the AIS Discovery experience