For simulating composite assemblies, I frequently need to tie shells together. I understand from researching the topics on this forum, that this can be either done using tie Constraints or Contact Pairs. This works well. Alternatively, I understood that parts can be merged and a compound part be created. This would then avoid the need to apply tie constraints. I tried the latter but this seems not to give the expected result and, in fact, keeps treating the parts as separate entities. Most likely something I have done wrong and hoped to check here for advise and insight.
As an example, I tried a very simple model of two parallel shells of 150 x 100:
I was under the impression that, if now a compound part is created using “Merge Parts”, that a tied connection should have been created; guess, here I may got it wrong?:
I would have expected that there is merely compression stress in the skins and a 0 displacement at the tied centre joint but this is unfortunately not the case:
Here is the simple file in case of interest to anyone:
FYI, fully understand that this model could have been set-up better, e.g avoid shells etc. But this is just a simple example of a more complex model I tried to run. To speed up initial model set-up, I tried to avoid defining the different layers through keywords and applying all types of tie constraints manually until I get a geometry which is promising. Therefore, I just want to use some homogenised properties until I convince myself that a more detailed analysis is warranted…
Thank you for looking at it and thank you all the same if there is no simple solution.
Part → Merge only assigns elements from multiple separate parts to one common part without actually merging the mesh. You can use Model → Node → Merge Coincident Nodes with a proper tolerance instead. Then run a frequency step to verify the connection.
Just wondering, would it not make sense that this merging of nodes is done automatically when creating a merged part? Why would this need to be done separately; or better, are there any instances where node merging would not be required when merging a part?
The name is confusing, I agree. Do we have a better name? But sometimes, this functionality is needed without merging nodes, so I would like to keep it.
Or “Group Parts” maybe? … That what it effectively is, a group of individual parts. Should come with a warning that, unlike in Power Point, a group cannot be ungrouped
Personally, I would like to see this for Prepomax as well, since it is currently no longer possible to subsequently merge coincident nodes once the parts have been merged.
posibility using equal displacement constraint by equation, internally it will work similar to merge nodes but some advantages of non-averaging stress. Also, it can be removed easily later. Problem occurs for incompatible mesh due to non-coincident nodes, switch to tie constraint can be used in this case.
or maybe by introducing of sub-parts, so all available features of selection (sub), exploded view and search contact pairs (tie constraint and contact type tied) still can be accessed.
I only read the latest post and I only meant to add that you can merge nodes of the same part. For that you have to use the feature merge coincident nodes.
