I am looking for practical advise on Groups vs Layers in FEMAP. I have gone through the FEMAP supplied user guide, command manual & examples but still Groups vs Layers confuse the heck outta me.
For example, I am currently working with a complex installation which has several sub-assemblies. After importing the CAD model, I created a layer for each sub-assy and have created surfaces for meshing.
Since I require results in individual sub-assemblies, I intend to create a separate group for each one. I also want to mesh in groups rather than layers (don't ask me why...probably this is how I am used to doing it in other FEA packages).
So, this is what I did. I made the group I want "active" and display mode was set to "show multiple groups" and checked off the inactive groups. The layers were set to "Visible All Layers". Of course one of the layer was set to active (I don't know if its possible to keep all layers deactivated at the same time).
When I created a node, the node was placed in the active layer instead of active group. I retried the command again with "auto add" option turned ON in group, but still the node gets place in the active layer.
So is it better to model, assign BD & loads in layers and finally move them to groups later on before analysis in FEMAP?
Or any suggestions on how I can make no layer active while working with Groups so that any entities created are automatically placed in active groups.
My suggestion is clear: use LAYERS to build your FEMAP model with geometry & mesh, and use GROUPS to postprocess results, this is the clear distinction.
Think that an entity in FEMAP can belongs to ONLY ONE LAYER, but can belongs to as many groups as you want. An of course, only ONE LAYER could be active at any time.
Do not define many, many layers, try to reduce them as much as possible to not complicate things. I love to include in the same layer both geometry & mesh, for instance, midsurfaces geometry + Shell elements., or Curve lines + Beam elements, etc..
The ACTIVE layer is where goes the new entities created in FEMAP, either a midsurface or the nodes & elements when you mesh the midsurface, so you should be smart at every moment to know what LAYER is active. If you have any entity in the wrong layer, not problem, use command MODIFY > LAYER and move to the correct one.
And that's all, very easy, run this way, use LAYERS only for creating your FEMAP model and you will realize the powerful that is.
Unfortunately, I think Blas had a typo in his response to you, so to be clear... an entity in FEMAP can only be on ONE LAYER, but can be in MULTIPLE GROUPS.
It is true that only one layer can be active at a time, but that is also true with groups... only one of either can be active. For display purposes however any number of each can be drawn.
The reason we have both is that often users coming from a CAD background like to think in terms of layers where they can simply assign each entity to a layer, then turn the layers on/off for display. Many times however traditional FEA users need to look at subsets of their model in different, possibly overlapping ways... groups are perfect for this and give you many additional ways (rules, clipping, etc.) to build up the pieces you want. Multiple groups can also be "assembled" as "reference groups". As Blas pointed out, often times this is used for postprocessing.
If all you want to do is graphically select individual entities as a subset of your model which is done based on the entity IDs and there are no overlaps in your selections, then really either groups or layers can pretty much equally handle your job. When you need to do more however, groups are usually the better approach.
Groups are helpful in that there are many easy ways to create and modify them, and easy to delete and recreate if they need tidying up.
A real favourite is Group -> Operations -> Generate Solids.
This allows you to select solid(s) (including sheet solids) and put them all in a group, automatically including all their sub-entities (surfaces, curves, points), and any associated elements, nodes, loads and constraints. This is much quicker than any layer method for doing the same job.
Layers are good for major and distinct sub assemblies where you only need to remember to change the Active Layer quite rarely. If you have numerous layers and don't remember to change the active layer to the one which is relevant to what you're working on, then you can easily end up with a mess of things on wrong layers. Eg. if you start using Geometry Editing on the Meshing Toolbox with the wrong layer as Active, then your layers will end up quite disorganised. Most inconvenient is when you have (sheet) solids on multiple planned layer(s), but some of their (new) surfaces, curves, points and mesh on a different (eg.the Active) layer.
So I recommend "a few" layers for substantial distinct zones of a complex model (eg. I have an entire aircraft model with only 3 layers), Groups are useful any time for separating out local zones for either ease of modelling or viewing results. If I am doing a midsurface model, the midsurfaces would always be on different layer(s) compared to the original solids.
Groups can slow Femap's performance a bit if you have dozens of them and there are numerous rules and/or hundreds/thousands of disparate ranges of IDs used to define the lists of entities in the groups. For geometry-based models (mesh, loads and constraints all associated with solids & surfaces), groups can be tidied up after the model is complete by deleting or resetting groups and again using Group -> Operations -> Generate Solids.
Thanks a lot for the responses.
EndZ, at the moment, I think I will follow Blas' suggestion. Work primarily with Layers and create groups just for the purposes of Post-Processing.
'Cause, my original issue still remains. May be there will be better management improvements in future releases.
So just to clarify...
Groups and Layers are independent functionality.
There is always one active layer, regardless of how many layers and which layers are visible. So brand new entities go onto that layer (except eg. copied entities typically go into the original entitiy's same layer, regardless of the active layer).
As mentioned before, all entities are on one (and only ever one) layer. If they are moved to another layer, they are removed from their original layer.
There is either: (a) no active group or (b) one active group - regardless of how many groups and which groups are visible.
If Group -> Operations -> Automatic Add -> Active is switched On, then new entities will be added (based on simple ID listing) to the Active Group if there is one.
However, Groups can be defined by rules (simple examples include element by property or nodes in solid). In that case, new entities will be automatically added to multiple groups if the entities satisfy those Groups' rules.*
*If taking advantage of Group Rules, then keep in mind that you may need to periodically use Group -> Operations -> Evaluate (or Evaluate Always with a possible performance reduction for big models).
Similalry, if the model is complex with many groups and you use renumbering, then it is better to switch on Group -> Operations -> Renumber Rules for all groups.
The Femap defaults for Automatic Add and Renumber Rules might not be best, but are easily sorted.
Entities can belong to zero, one, or multiple groups - and whether by rules or simply by ID listing.
If your specific model suits using only layers for model building then that is good. However, I can guarantee there are numerous models where using Groups (as well) is modelling best-practice. This is because the Group creation / modification tools & rules substantially exceed the Layer creation / modification tools and rules.