This approach may have an interesting usage when applied multibody systems.
If it is of general interest, it would require some changes internally in PREPOMAX compression only definition.
Actually, all the bases of the gapuni elements associated to compression only supports are grouped for later fixation.
That should be changed.
Instead of fixing each node base, they should be part of a rigid body definition with its corresponding Ref node so the user would be free to apply BC and loads to that REF Node.
Fixing the REF and ROT could be the default option to mimic the actual behavior.
I have done some tests, and it works nicely. See Vid made using the MECWAY interface cause I couldn’t manage to isolate the Gapuni bases to create the RB in Prepomax.
What does it mean in terms of improvement?. Loads transferred between different components of a multibody system by means of sleeves would only transfer load in the compressed contact area as they should.
Now, I don’t think there is a way to get that without introducing contacts (Pain) and detailed modeling of the sleeve.
Can you share the input file for reference ? So those are 2 GAPUNI elements (with spring-like symbols) attached to a common rigid body reference node ?
Regarding the compression-only constraint, it would be also helpful to have an option to invert its direction: Tension-only constraint
Perhaps there could be a compression-only spring so that the user could apply it like Point Spring to selected nodes (although, currently Point Spring is grounded so I guess two-node springs would need to be added first).
No. The two springs are real springs just to act as sleeves. They could be Truss elements or whatever.
Each REF node (two in this case) is at the center of each hole. Thinking on your YouTube video of the bearing preasure on LUG as reference (simulated with only compression). In this approach, the lug doesn’t move. I’m moving the only compression surface against the lug.
yup -i played with springs, trusses and beams -its an approach that appears pretty robust and with the API in Mecway it is actually quicker than adding a handfull of other types of connections
Ah, ok. Nowadays, most preprocessors kind of abandoned such classic spring elements (not to mention GAP elements which used to be applied in 1D contact problems) and replaced them with various connectors (elements or predefined connection types). For example, Abaqus doesn’t support GAP elements in GUI and recommends connector elements in place of regular springs. However, as you said, the old-fashioned springs can still be useful even instead of classic spiders (couplings or rigid bodies in CalculiX).
But back to the original request - the problem in such cases is that there has to be balance between user-friendliness and versatility/customizability. For example, if a preprocessor defines compression-only constraints automatically, this is very convenient and sufficient in most cases. But sometimes more control is needed as discussed here. Perhaps advanced settings could be introduced to control the direction of the GAPUNI elements, their base nodes and so on. Scripting could also suffice, though (after all, preprocessor can’t handle all possible custom cases).
There are some really powerful open-source MBD solvers such as MBDyn, but they lack graphical interfaces and connection with FEA. There’s a project to combine MBDyn with CalculiX using FreeCAD’s interface, though.
Of course, it would be really nice to have some features that could aid with mechanism analyses using FEA, such as connector elements in Abaqus.
this is similar to pin or bolt simplified connections proposed at another threads, required to add spring at out of plane directions also. It can be complex implementations for three-dimensional problems when bolt/pin nonlinear response need to be considered, linear one is less complicated.
there”s two approach to modeling such as connection model, using two nodes spring without tension stiffness and gap element. First approach allows consider nonlinear response by force deformation relationship, the last is limited to linear in compression stiffness.
as i mentioned before and at another threads, it’s SIMPLIFICATION of bolt connections. Example is actually 2D approach since it ignore eccentricity of multiple parts connected. Did models show out of plane bending? probably not.
and sure, i can understand the problems of pliers like models mechanism.
Ideally, such a tool for simplified pinned joint models could offer multiple approaches. Then the user would select which MPCs (rigid body constraints or kinematic/distributing couplings) and 1D elements (springs/beams/GAP) are used internally (although, it’s usually named more generally such as rigid/flexible).
testing shown pinned joint models using kinematic coupling have problem in large deformation analysis, so it’s probably not general solution. Or these example test need to define penalty contact between surface parts also.
new feature in equation of CalculiX seems really helpful, user defining single master node and large slave nodes instead of two nodes only. Maybe this approach can be used in PrePoMax also in generating two nodes spring or gap element.
below my test related to modeling such connections, it’s a nonlinear spring connected to shell element. I’m not remember about the years since old enough, i did not continue due to problems with layered shell element and no resolve even posted the problems. Another complexity is in shear and clamping transfer mechanism, but i can continue if working in the right direction and purpose of implementations.
maybe i’m not good in explanation, but someone can build full solid models including pins of these part interactions and proof that out of plane deformation is zero. To understand clearly can be simple by plotting contact or bearing stress passing through all part, no out of plane bending means stress are uniform and the value along the line is exactly the same.
