Parts disconnecting during static analysi

[TractorGuy]

I have a mechanism that I am trying to analyze that utilizes 2 linkages and pinned connections. I have setup the pinned connections as contact interfaces and I have verified that there are no geometry errors, no interferences, and everything meshes properly without errors/warnings. I have also confirmed that all of the bonded/free/contact interfaces are correct.

For some reason, the middle pin in the mechanism disconnects during the analysis and I get incorrect results (one side has all the stress and one side has no stress). I have been trying to figure this out for 3 days now and so far I haven't been able to determine what the issue is.

Does anyone have any suggestions for what may cause this issue?

[SweetPeasHub]

Is there any way you can give more information?  A diagram or image or screenshot or the CAD files?

There are just too many possible issues to rule out.

One side (of what?) has no stress.  The pin? another set of link members?

To send CAD you have to zip the files first.

[TractorGuy]

Unfortunately I can not share any of it because it is proprietary, but maybe I can describe it a little better...

Imagine you have a primary lever with a force applied to the end perpendicular to the axis of rotation. The (pull) force is transferred from the primary lever to an intermediate lever by a linkage rod. The force is then carried to another lever by a second linkage rod to a remote lever where it would actuate another mechanism.

For the purposes of my FEA I am applying the force on the primary lever and holding the pivot of the remote lever fixed so I can see all of the stresses through the linkages. When I run the static analysis, the second linkage disconnects and no forces are transferred to the second linkage or the remote lever. What is really odd is that the intermediate linkage behaves as if it is fixed so it still solves. Also, the stresses are much higher than what I calculated so I don't trust any part of the solution it is giving me.

[SweetPeasHub]

sort of like this maybe? Model

Stress

Pin model interfaces

Is it possible your pins or links are under-constrained and free in axial translation?  My pin is bonded to one of the link elements and the 2nd link element is sandwiched by contacts to arrest translation, in addition to the cylindrical contact.  I made disc shaped surface regions at all the pins to accomplish this.

Are there bending stresses at the bottom of your 2nd lever, more like the remote lever in my result?  This would mean it is somehow not free to rotate.

Since for contact you must have NonLinear /use load histories, how many output steps did you ask for?  It is possible, if there is large amplification of movement due to the lever system, that the contacts move past each other too fast to be resolved, so smaller time steps, larger elements near the contacts, and/or larger penetration depth may help as well.

My model is simple enough the elements are relatively large, and I have contact penetration depth set to 15%.  My model does not have excessive mechanical advantage/disadvantage.  I set my output steps to 21 - user defined - space equally.

I also recommend setting up a simplified model to prove it out before swapping more complex link/lever geometry.

I have this model in Creo 6 and can send it if you can use it.  Hopefully I have given you some ideas to pursue.

[TractorGuy]

You have the right idea on the linkage arrangement. The main difference is the linkages have clevis that fits around the lever. The pin is bonded to the clevis and the lever has a contact between the pin as well as one on each face of the clevis so it should be constrained from axial movement.

My results show stresses similar to what you have on the primary and intermediate levers, but the 2nd linkage and remote lever are completely cold with no stress.

For now II will take your suggestions and see if I can make any progress.

[TractorGuy]

After hours of frustration I have traced the issue down to the way the way Creo generates extrudes in formed sheet metal parts...

The intermediate lever has an offset bend and the hole for the pin is created using a sheet metal extrude with the default options. When you do this, Creo generates two half cylinder surfaces instead of one continuous cylindrical surface. When the surface is split in half, it is very easy to create a contact interface to only half of the hole surface leaving it free to separate in the opposite direction.

The simple fix is to change the extrude to use the perpendicular to surface option and make sure that the contact interface includes the entire cylindrical surface. When this is done, it solves as expected.