This content has been marked as final. Show 15 replies
The comment I made earlier has nothing to do with meshing. I expect to remesh any time the model changes.
However, when changing a PART ONLY which consists of SURFACES on which one is meshing SHELLS, one must issue the "update all components" command prior to meshing, or changes won't be reflected in the mesh, even though the shell definitions have been updated (which takes forever, by the way.) Of course, nothing in the documentation would imply this.
There are actually 2 things going on here. If part features or assembly components are changed, a new mesh will be required. If components in an assembly are changed or added, a remesh without performing an "Update All Components" may skip the new components. That's why I said earlier that this is only an issue for assemblies. "Update all components" doesn not, in itself, force a remesh. If features within a part are changed, you'll still have to create a new mesh. This is not automatic since CW has no way of knowing if you want or need local mesh refinement on those features. Your involvement in the process is still necessary. A pioneer in the industry once told me, "FEA is not a spectator sport..."
You may want to check out the COSMOS Companion unit on "Mesh Accuracy" for more info on the imporatance of mesh control at:
The confusion (on my part) was related to the fact that bodies in a part and components in an assembly have similar characteristics.
I re-read your post on shell assemblies and wonder if you'd like to get together on this and brainstorm some ways we might improve our shell modeling. It is still relatively new functionality so we need to see how people are using it to better understand workflow. (By relatively new, I mean that with the ability to bond shells in v2006, the door to complex shell modeling was opened. Prior to that, trying to model anything beyond simple parts and assys was prohibitive.)
Can you contact me offline on this. I'm interested in your thoughts on how we can improve this in future releases.
If there's anyone else building large and/or complex shell structures, feel free to shoot me an e-mail as well.
Yeah - sorry, I read your 2nd post a little too quickly I guess. I have tried 2006 but found I have to use the Update Components command. I assumed that it was still using any given study for any configuration that happened to be active as well. I just tried and found this is wrong - it greys out a study if the configuration that the study was started for is not active - just like I wished - great. Also - I can right click on that greyed out study and make its respective configuration active, and hence the study - NICE!
In v2006, studies are tied to configurations. They weren't when the request for "Update Components" was made and the feature was implemented. I'm happy to hear you are using Configurations for assembly variations as that is a great way to keep models organized. However, not everyone is. I did a presentation at SWW on this topic and many users didn't consider configs for COSMOS model mgmt.
The usefulness of "Update Components" may be fading but since it was driven by users in the first place, we need more feedback before yanking it. Any other thoughts?
I'm a bit perplexed by the update components as well. I thought that's what using a dedicated configuration for FEA analysis was for. Isn't any one given STUDY tied to only one configuration? It seems to me it should be. IN FACT, when I define a study in one configuration, that study should disappear if I change to a different configuration (or maybe collapse and grey out would be better). I always use either a dedicated configuration or a whole different model, because inevidably I need to tweak a few features here and there for FEA analysis. Floworks studies are tied to configurations, that's how you can batch run more then one analysis.
The "Update" command was implemented based on user feedback that they didn't always want to update the analysis model when changes occurred to the CAD model. This is only an issue when dealing with assemblies. Part changes are carried thru to CW automatically.
Does anyone on this forum feel that "Update Components" is still necessary? WIth better modeling techniques and the use of configurations to organize models, maybe this is no longer an issue. Any feedback would be welcome.
I have been using cosmos works for 6 years. Can you say “update all components”? Not doing this command (right mouse menu) causes the mesher not to kick in. But it still runs the mesh anyway and results in "undefined material." So I had to go in to the part and open-exit the solidworks material editor.
THIS IS EXACTLY THE SAME THING AS SYNCHRONIZING BETWEEN 3G.AUTHOR AND SOLIDWORKS.
So what if 3g.author isn’t integrated with solidworks. From the above described experience, neither is cosmosworks. When would a user not want to “update all components”? And why would a user not want the materials to inherit? It is the same part and if I want different material properties I expect to edit them but otherwise inherit.
It should be automatic.(ie. integrated!)
"This is only an issue when dealing with assemblies. Part changes are carried thru to CW automatically. "
Not strictly true. When using a surfaces in order to shell mesh, part changes are NOT carried through to CW automatically.
Yes, a minor alteration to a previously-defined surface doesn't seem to require "update all components", only when surfaces are added or deleted, which is most of what we do. The real point was that "update all components" doesn't apply only to assemblies.
BTW, sometimes after "update all components" it will fail to mesh surfaces that were not changed, but I haven't been able to reproduce it in a sufficiently simple manner to forward to support. It takes so long to change large shell models, we can't afford to run though things twice.
For the shell modeling issue please see:
OK...let's slow this down a bit... First of all, all my comments reflect v2006. Now...
I created a part with a single rectangular surface then created a study, assigned props and meshed it.
Next, I added another surface body to the part and even though I assigned it properties, the mesher didn't see it until I issued an Update Components. I'm pretty sure that this is due to the fact that multiple bodies in a single part are handled in a similar manner to multiple parts in an assembly.
I went back into the geometry and added a hole to the sketch for the first surface. I came back into CW and issued a mesh command without an "Update Components" and the mesh recognized the geometry change.
I think the case has been made on this thread for removing the "Update Components" step and we are talking about that in product planning. If there is still something I'm missing, please let me know.
As far as the long time req'd to update shell props, can you describe the problem more specifically? Is it a workflow or a a processing related issue?
Thanks for the feedback! This really does help us!
I agree with B Carmen. If I change a feature of a part in an assembly, CW does not know the part has been changed and does not automatically remesh even with 'update all components'. I need to run the mesh manually after a part change.
I guess it depends upon how you look at it. The problem is, some of us have been using SW for quite some time now and use configurations. Then COSMOS comes along and doesn't really follow that philosophy exactly. In the end, what I want is to know that I can analyze a certain design exactly the way I specified that design in my model, try another variation on that design and analyze it, and another, and another, ... and keep all of them for reference. Not only that, But wouldn't it be prudent to be able to go back to a previous analysis and be able to see what the design actually looked like imediately? This makes sense to me, but maybe I'm missing something.
Reading back thru your previous post, I think we work with Configurations exactly as you suggested you'd like to see... Have you tried recently in v2006?
Since studies are definetely always configuration specific, the updat all components has reached obsolesence, since updating components happens by first setting up configurations and then changing configurations according to design intent changes.
Please get rid of update all components. It is a big nuisance to a daily user and a real nightmare to the rest of us who misinterpret the error message and waiste hours of solver time only to find out the mesh was incomplete and materials were not updated.