hexagon logo
Register Sign in
  • State Not Answered
  • Replies 9 replies
  • Subscribers 11 subscribers
  • Views 103 views
  • Users 0 members are here
Options
Related

Flexbody constraints

Josef K.
Hello,
 
I want to constrain a flexible body on multiple points, using rbe2 and fixed joints. When I use more than 1 joints I get a redundunt constraints message and the simulation stops due to lock errors. I tried removing some of the fixed joints, but then the flexible body just "pivoted" around the remaining last one.
 
Is there a more proper way to constrain flexible bodies on multible points?
 
Thanks,
 
Josef
  • 0
    There's a few thoughts:
    1. Make sure you flexible body is all one "part". No hinges, connections, CBUSH, etc within the flexbody.
    2. It's strange that the flexbody is "pivoting" about the final fixed joint. If it's really "fixed" to the fixed joint, you would see no relative moment between the flexbody and the fixed joint.
    3. If you are getting messages about redundant constraints after adding a flexible body, I think the issue lies elsewhere in the model, and not the flexible body. Flexbodies ADD degress of freedom to your model, and is a common way of eliminating redundant constraints in your model
    4. Typically, you don't use fixed joints to attach a flexbody to the rest of your model. Normally, you are replacing a rigid body that's already connected to the rest of the model with joints and bushings with a flexible body. In that case, there is no need to introduce fixed joints into the model.
    5. finally, and honestly this is the first thing to look at: make sure the model runs correctly first without the part in question being flexible. You may have redundant constraints, but it should at least have the correct kinematic behavior.
  • 0
    Hello,
     
    As far as the rigid model is concerned, it ran properly. When making the part flexible however, the fixed joints I use seem to just not work (even though they are attached to the int nodes of the flexbody).
     
    About the pivoting part, what I mean is not that the model rotates with the joint but rather it gets deformed around it. The rbe2 element where the joint is gets deformed and as such the whole body seems to pivot (as if I were pinching it by the rbe2's master node).
     
  • 0
    Update: By checking the plots and not the graphics, the model seems to work fine. May it just be a matter of graphics?
     
    What I also found weird however is the fact that while the displacements are small throughout the model, the deformations on the flexible body are huge (displacements ast +-1 mm, deformations at 1e6 mm, numerical error I assume)
  • 0
    If the part if deforming around the joint, then:
    1) double check your material properties are reasonable in the flexbody:
    a) check the mass of the flexbody that it's reasonable
    b) animate the modes of the flexbody to make sure they are reasonable.
     
    But something does not add up, if the model is performing ok, but the deformations you are seeing are that large.
     
  • 0
    Hello,
     
    I double checked the flexbody's properties and eveything is fine. I also swapped the multiple RBE2+FIXED JOINT setup with a single RBE2 comprising of all the previous areas and a single FIXED JOINT and the model seems to work properly as far as the "pivoting" problem is concerned.
     
    Checking the rigid body model, there are no redundant constraints and just two degrees of freedom (just as many as they should be). When making the part flexible, the degrees of freedom increase due to the flexibility of the part and the model is still verified successfully. The flexbody's modes are also correct as I cross referenced them with a simple SOL103 I'd run beforehand.
     
    However, the model still crashes after a while and when viewing the graphics it seemd as though the forces applied to it were making it deform extremely, which is not possible considering the fact that the forces magnitude is relatively small. I also disabled them completeley, leaving only gravity and the model still crashed. Surprisingly, disabling gravity and leaving all other force allowed for the longest possible simulation duration (40 seconds, the total time needed is 100 seconds). I also tried reducing the step size to 1e-4 and 1e-5 and the model still crashed.
     
    Does the fact that even the gravity force alone deforms the flexbody beyond reason mean that my flexbody is built poorly perhaps? Or could the large simulation time be part of the problem? Could it be a matter of the stability of the model?
     
    Thanks,
     
    Josef
     
     
  • 0
    Does it crash at approximately the same simulation time every time?
    Try stopping the dynamic simulation just before the crash and do an Eigen mode analysis on the model and see if this gives you some answers. Maybe some strange frequencies are excited in the flexible body.
  • 0
    Hello,
     
    The model crashes on different simulation times, depending on which forces I leave active and which I deactivate. What strikes me the most is the fact that just activating gravity, with no other motions/forces, makes the model crash the fastest (at 10 seconds instead of 50-60 seconds of simulation time). When I animated the model it seemed as though gravity significantly deformed the model which is not possible as the model is quite light (I double checked the units as well) and I have also run a sol101 with gravity on a nastran model with no problems whatsoever.
     
    By running an Eigen mode analysis, do you mean to exctract a nastran model from adams and run sol 103 while on its deformed state or is there a way to run an eigenmode analysis inside adams?
     
    Thanks,
     
    Josef
  • 0
    Have a look at LINEAR/EIGENSOL in the Adams Solver online help
© 2025 Hexagon AB and/or its subsidiaries. Privacy Policy | Cloud Services Agreement