25 Replies Latest reply on Aug 2, 2013 5:39 PM by Justin Strempke

    Constraint for faces to move simultaneously?

    Justin Strempke

      I'm wondering if there is a way to make a sort of symmetry constraint which can be applied to a solid face, disallowing rotation only.  As an example, take a 100' long cable and put it on tall skates spaced 10' apart or so.  As the cable sags, the skates roll closer together.  I would like to be able to model the region between two skates only.  As-is, I can set the normal translation or two transverse movements, but happen to have multiple parts (say the skate is split in half) and want to define that the skate cut edge and the cable cut edge can move laterally but the elements must remain on a planar, vertical surface.

       

      I've thought of using a rigid rectangle (or a sim-only material with very high E), constraining the lower face to slide along the ground, which the cut skate and cable ends are sliding on the vertical face, but that seemed pretty tedious and possible unreliable.  The other complicating side is some cases this would have to be done in two directions, so more like a net of cable contracting toward a central point.  The models are very complex (lots of small parts repeated, with high mesh resolution) so I'm trying to avoid modeling the entire assembly.  I'm not interested in the contractual displacement, only the stresses of the cable/skate point, which would change when the cable sags.

       

      Is this doable in any fashion, or am I overlooking something simple?

        • Re: Constraint for faces to move simultaneously?
          Jared Conway

          probably worth a diagram here both simplfied version and the full version of your model. i'm really not following what the "real thing" is that you're trying to replace and what you want it to do to your model. based on what i can interpret, if i had 2 blocks sitting on the ground and had some force or "thing" that brought them together such that i wanted the faces that face each other to stay parallel, i would just restrain the "ground" in the normal direction. it would "skate" along the ground, never rotating, in any direction necessary.

            • Re: Constraint for faces to move simultaneously?
              Justin Strempke

              So here's my awesome MSPaint diagram!  Pretty much this is a partial section, with the supports and beam expading outward both ways a relatively long way.  My thoughts are the very ends are not grounded, therefore the entire structure can sort of collapse when a UDL is applied to the beam.  I want to be able to sim just a section of the pattern, outlined in red.  At the cut vertical faces however, should be a symmetric BC, or at least a no-rotate condition.  From what I can find so far, I can fix a side, apply symmetry to a side, but the opposite will be able to rotate inward if I want it to be able to translate inward.  I'm interested in the stress at the top of the beam, and not in the overall contracted length.  The top stress above the skates will be lower (I assume) if there is tension allowed from restraining both ends; they need to be able to draw inward.  Hope this helps, can't post the model.

              EXPANSE 1.jpg

              Here's my thought with the rigid blocks, but I don't know how to bond the sagging beam to them and still let them 'squish' above the skates.  I could shift the blocks to be mid-beam as the squish there would be insignificant..

              EXPANSE 2.jpg

                • Re: Constraint for faces to move simultaneously?
                  Jared Conway

                  "From what I can find so far, I can fix a side, apply symmetry to a side, but the opposite will be able to rotate inward if I want it to be able to translate inward."

                   

                  are you talking about the outside faces of the blocks in the bottom picture or something else? if you apply a fixed BC on the outside face of the right hand side and then symmetry (normal restraint) on the other side and then roller slider on the bottom (where the wheels are), when you apply a load to the top, the left block will move inwards. i'm not sure where you're getting the impression that it will rotate. but maybe i'm not following how/where you are going to apply those bcs. might need another fantastic paint drawing.

                   

                  something else to consider, rather than just analyzing one unit. just do a half unit and apply symmetry at that location.

                    • Re: Constraint for faces to move simultaneously?
                      Justin Strempke

                      Awesome paint drawing #3.  I can do a fixed on one end and symmetry on the other, and just look at the side with symmetry.  On the fixed end though, it won't allow the face to deform properly (red arrows).  I'd like the faces to move inward (green arrows, or at least one face while the other can move on the face plane, but be fixed axially), while restraining them from inward rotation (purple arrows) as they would if it were just a simply supported beam.

                       

                      EXPANSE 3.jpg

                       

                      I've tried using a Flat Face (and Fixed) on the LHS and a symmetry on the RHS, but the symmetry BC locks normal translational movement (as does the Flat Face & Fixed).

                       

                      EDIT: Additional pic, so top is what I want, with essentially symmetry on both ends (not fixed, must be able to deform on symmetry plane).  The bottom is simply supported which allows the ends to rotate inward.

                      EXPANSE 4.jpg

                      I'm probably either wanting to do something unreasonable or making an easy solution very difficult - but I just can't see it!

                        • Re: Constraint for faces to move simultaneously?
                          Jared Conway

                          not sure where my brain was previously, you're right, normal restraint will stop them from moving in the direction of the green arrow.

                           

                          for that motion, i'd put blocks on both ends and only restrain them with roller slider on the ground, then roller slider normal to the screen. make them rigids. the faces of those parts will stay parallel to each other and only move in the arrow direction. you're going to have some issues with stability, so i would cut the model at the load location and apply symmetry there and only have a block on one side.

                           

                          regarding the beam deformation, i'm not toally getting how you're expecting them to deform. the top arrow, are you expecting it to squish? ie, slide along the face of the block rather than be connected to the block bonded). if so, i could get what you're trying to show with the bottom arrow. if not, i'm not to clear on how you're thinking it is going to deform from top and bottom.

                            • Re: Constraint for faces to move simultaneously?
                              Justin Strempke

                              Yes, I'd like for the beam end to slide/deform at the rigid block face.  The loading a UDL/pressure load, and the supports are far apart so there's decent compression at the supports.  As you said though, I can use the same ideal and put the blocks at mid-span - where the compression isn't an issue and the thickness is essentially uniform, thereby allowing the supports to act on the beam as necessary.

                               

                              I think it'd be a handy BC to have, I've had to work around it a few times before but the models were only 1 or two patterns wide and I could deal with it.  I suppose it'd have to be a moving wall perpendicular to one or two planes, but the elements could slide on it.

                            • Re: Constraint for faces to move simultaneously?
                              Jared Conway

                              just want to make sure of something here, you're using the words fixed and symmetry. when you place them on face, they are basically the same thing except that with fixed it also restrains the direction out of plane and also in the upwards direction (in your pictures).

                               

                              based on the top item of your bottom picture, the method i described should work. the blocks will slide in and out but where the beam is connected at the interface it won't pivot.

                                • Re: Constraint for faces to move simultaneously?
                                  Justin Strempke

                                  Correct with fixed and symmetrical differences.  SWX just won't let you put Symmetry on two opposing faces.  The sliding blocks work well with basically a 1D line/beam, but when I tried to take it to a 2D grid I had issues.  The rigid blocks have to run the entire 'depth' (relative to the above 1D beam), and in doing so lock movement in that direction.  I suppose I'll just have to make a very large model or be okay with the assumption that the small contraction does not have a significant effect on the stress, and just Flat Face all edge faces...

                                    • Re: Constraint for faces to move simultaneously?
                                      Jared Conway

                                      not really following what you mean by symmetry on 2 opposing faces. symmetry is just restraint in the normal direction. you can apply it anywhere. same goes for what you mean by 1d vs 2d, running the entire depth or locking the movement. it sounds like you may not have had an in/out of plane restraint which would cause problems. probably easiest to post a simplified model that isn't behaving so that we can talk directly about hte setup vs theoreticals.

                                      • Re: Constraint for faces to move simultaneously?
                                        Bill McEachern

                                        I am not exactly sure what you are after and I did not read it really thoroughly. I think to do this correctly you would need to do a nl analysis to get the right sort of behavior as it would get corrupted by the small displacement assumption - but maybe not. Did you consider putting the symmetry constraint in the middle of the saging member? And, then on the other end connect the top bit to the bottom bit with a rigid connector. The bottom bit is free to slide toward the middle with the roller slider. Would that work? Not sure I really understand what you are trying ot do....but maybe

                                          • Re: Constraint for faces to move simultaneously?
                                            Justin Strempke

                                            I'm pretty green on the Rigid Connectors; I understood them to make the chosen faces completely rigid, then apply links to the other chosen, rigid face.  The benefit would be that they could be some distance away and not touching or in plane.  I may be wrong, but avoided this because I need the faces to be able to deform but remain attached to a plane, but have the plane move inward too.

                                             

                                            I'll get a sample model up tomorrow, hopefully help explain what I'd like to do.

                                              • Re: Constraint for faces to move simultaneously?
                                                Jared Conway

                                                from your original description, i can see where bill is going. i thought about something similar but figured there was probably more to what you wanted to do. but your description of how a rigid works is correct.

                                                  • Re: Constraint for faces to move simultaneously?
                                                    Bill McEachern

                                                    Well, from what was written by Justin I am not sure he understands with fidelity how the rigid connector works so I will describe it and see if adds anythig to it: Rigid bars connect the nodes of the two faces connected, thus the faces can deform but there will be no relative deformation between the faces - it behaves somewhat  like the infinitely stiff block you described above. Your other option is that you can use the remote load rigid restraint option and connect the face or faces of you entiries to a single node defined relative to a specified coordinate system (conveniently placed) and then specify any combination of the six DOF for the single node - you could allow it to slide but not rotate if htat is what you are after. I am pretty sure you can get any restrain combo you want using this and the other tools.

                                  • Re: Constraint for faces to move simultaneously?
                                    Justin Strempke

                                    So here is a sample modle of what I'm looking at.  I'd like all the support bases to slide, thinking if I constrain two complete edges of the plate that the other two are free to retract inward if necessary.  The issue is that I want to be able to use a 3x3 support section or so, as the actual model is more in the 30x30 grid size.  The supports have to be No Pen to the plate as well.  Putting a Flat Face on two sides is easy, but it's getting the other two to move as if they have the same grid on the other side that's hanging me up.

                                     

                                    test assy screenshot 1.JPG

                                    NOTE:  I put this in the wrong discussion, reposting here.  Bill McEachern had the right comment (Bill I gave you credit for above since you mentioned it):

                                     

                                    "remote restraint option in the remote loads function - easy peasy"

                                     

                                    That makes too much sense - a displacement restraint under Loads...    I did some basic testing with it using square tubing, and think this is exactly what I'm after.  I used the Displacement option, with the end face selected, and checked the non-axial rotations.  Took a few runs to find out I need to set the location pretty far away to make the 'spider' legs as perpendicular as possible for good results.

                                      • Re: Constraint for faces to move simultaneously?
                                        Jared Conway

                                        if that is your full model

                                         

                                        and you chop it so that you work with a quarter model

                                         

                                        apply symmetry at the cut planes

                                         

                                        apply no penetration between the smaller blocks and the palte

                                         

                                        put a roller slider at the bottom of each of the smaller blocks

                                         

                                        a uniform load on the top

                                         

                                        you should get what you want

                                         

                                        i must be missing something that puts you in the need for a remote load (displacement)

                                          • Re: Constraint for faces to move simultaneously?
                                            Bill McEachern

                                            Hey Jared,

                                            My understanding is that it wasn't really symmetry he was trying to model. It was him trying to get his interpretation of an approximation to what he thought was actually a decent sub model. He has a bunch of flexible stuff suppored by stuff that is free to slide and he want to mimic it on a sub scale model. Hence the need to allow sliding but to preserve rotations. Think a semi stiff netting support by a bunch of things able to slide on a stiff surface.

                                              • Re: Constraint for faces to move simultaneously?
                                                Justin Strempke

                                                That's pretty much the idea Bill.  Jared, the model posted is a fraction of the size of the actual model.  Even at a quarter of the full scale to apply proper symmetrical restraints, the model would be something like 5+ million elements or so by the time mesh controls are applied to each small support. 

                                                 

                                                I have gotten the model set up, but it keeps failing on me for some reason with no reason given.  I'll continue working on it as I'm sure this is the right way to go about it.

                                                 

                                                Thanks for the help once again you guys!

                                                  • Re: Constraint for faces to move simultaneously?
                                                    Jared Conway

                                                    based on what you've described that you're going to setup, i can see stability being an issue. so depending on what your errors are, that could be something to take a look at.

                                                     

                                                    interested to see what you come up with.

                                                     

                                                    one question i have is how you came up with the location of the remote restraint. seems like that would be the challenging part here where a normal symmetry restraint would be easier.

                                                     

                                                    also for the plate, i don't know exactly what the plate is, but if you needed to make the problem smaller, shells would be better.

                                                    • Re: Constraint for faces to move simultaneously?
                                                      Bill McEachern

                                                      use soft springs in the strudy properteis panel/dialog. That should make it work or add some soft springs manually to stablize the model and give it some stiffness. It would be easier if you put a symmetry constraing on a half model in the center of the saggy bits in the middle.

                                                        • Re: Constraint for faces to move simultaneously?
                                                          Justin Strempke

                                                          I'll need a grid at least 5x5 to look at effects that I'm interested in (unevenly spaced support), and as-is I'm trying a triangular cut (half the square), with Flat Face on the long face, and remote on the other two.  I'll throw some springs in and see if that helps, but am wondering if it has to do with the corner of the two faces at a right angle that both have a remote restraint...  Need to test that out on a small simple model.

                                                           

                                                          For the remote restraint location, I just guessed a number far away (10x the restrained face span for starters.) such that it's somewhat perpendicular.  Probably need more like 100x; I'll see how it acts (if it ever solves!)

                                                           

                                                          For the plate I'm interested in the stresses through the thickness, so thought shells would not give that sort of information as a solid would.