8 Replies Latest reply on Feb 19, 2015 5:20 PM by Jared Conway

    I'm having some difficulty with a shrink-fit contact (attempting to use thermal loading to relax the interference)

    Joshua White

      Hello Simulators;

       

      This is my first time posting on the forum, although I am a frequenter.  I have been working with an issue for a little while with no progress.  I am hoping there is just something I'm simply overlooking (a quick fix).  I have confidence in you guys!

       

      I am analyzing a coupling/pipe connection for the oil and gas industry.  It's basically an API connection.  I'm doing an axisymmetrically simplified, nonlinear static study.  The threads are mated by load flank and there is some interference built into the study.  I am gradually applying contacts to each thread, but have chosen to move one at a time for now so I can catch any problems immediately as they arise.  After applying the shrink-to-fit contact to the innermost thread (left-most) of the coupling and the pin's corresponding "trough" or root, I ran the study and everything worked fine.  I then proceeded to do the same with the next thread; this is when I was prompted with my error that the incremental strain is too high (along with two other suggestions).  To remedy this, I have attempted to apply a thermal load to allow the coupling to move from an expanded state to a contracted (back to initial) state in which the interference would be built up.

       

      I excluded the other component from the analysis so I could see the effects of the thermal load.  I applied my thermal load so the temperature would begin higher than the one I specified as its "temperature of zero strain" and would then decrease to the zero strain temperature.  I also added a moving fixture to the left side of the coupling to move with the thermal contraction so as not to allow stress buildup, although I made it stationary to simplify things.  I then ran the study (again, with the other part excluded from the analysis).

       

      The study gave results that the coupling was contracting, but that it was doing so from the zero strain position, thereby ending in strain at a more contracted size.  I have been tampering with this for a while, but I cannot get it to do what I need for it to do.  Is there something I am doing wrong?

       

      Please help me out here!

       

      Edit:  Still no answer to why the thermal load can't contract the part from a strained, expanded position.  Nobody has experienced this before??

        • Re: I'm having some difficulty with a shrink-fit contact (attempting to use thermal loading to relax the interference)
          Joshua White

          Quick update:  Model can start from a strained position when using pressure load or displacement fixture.  The thermal loading, for some reason, just doesn't work (it's always from the zero-strain position to contracted or expanded).

           

          Wow, no response yet.  I suppose this isn't as simple a solution as I was hoping...

          • Re: I'm having some difficulty with a shrink-fit contact (attempting to use thermal loading to relax the interference)
            Jared Conway

            first impression is that you're picking and clicking rather than really putting together a strategy for solving this problem. for example you've chosen the large strain mode, remember this only applies to certain material models.

             

            let's take a step back and get a birds eye view of what you're actually trying to learn in your simulation. and what type of analysis you're using (nonlinear), what your boundary conditions are...etc.

             

            also, I would highly recommend that you start simple and slowly add elements to your analysis. when you run into a problem, don't jump to the complicated solution (throwing thermal onto it), really investigate what is happening.

             

            my overall guess is that shrink fit isn't the right thing for this analysis but i'm not sure exactly what you are trying to do. it looks like something i've done before where we wanted to see what kind of stresses are generated in a threaded connection but we took a completely different approach and it was very successful in the end.

              • Re: I'm having some difficulty with a shrink-fit contact (attempting to use thermal loading to relax the interference)
                Joshua White

                Thank you, Jared, for your response!

                 

                I had originally not chosen "large strain mode", but when only working with the first contact set I described, convergence didn't happen until after I had checked it off.

                 

                As for "start simple and slowly add elements", that's exactly what I was attempting to do: I worked my way through one of the contact sets (which was successful) and attempted another before running into an issue.  One of the primary fixes recommended by the error prompt was to apply a thermal load and relax it in the end to allow the interference to build.  I thought it was a logical solution.  It doesn't work the way I specify, however.

                 

                All I'm trying to do at this very moment is generate pre-load; it seems you know how to approach this in a better way.  Would you be able to provide any advice?  Also, do you know why my thermal load is not making the part react properly?

                  • Re: I'm having some difficulty with a shrink-fit contact (attempting to use thermal loading to relax the interference)
                    Jared Conway

                    the thermal load i don't think i fully understand with your description, i'd post at test model that shows what you are doing.

                     

                    preload > we put all the components in contact and then used a temp load on a split body in the part to create the preload. still not enough detail about your application to know whether this would apply or not in your case

                      • Re: I'm having some difficulty with a shrink-fit contact (attempting to use thermal loading to relax the interference)
                        Joshua White

                        Are the pictures not available to you?  I would think they are more than sufficient, along with the rather lengthy verbage I provided, to convey what I'm doing with the thermal load.  Please let me know if you can't see the pictures.

                         

                        If you can, then look at the ones titled "TempTimeCurve.PNG" and "ZeroStrainTemp.PNG" so you can see exactly what I've done.  I instruct the coupling to begin at 10 times the zero-strain temperature so it will begin in an expanded state.  Then, at 0.9 seconds pseudo-time, I have the temperature return to zero-strain temperature so the coupling will contract from said expanded state.  Again, the problem is that the coupling never starts from the expanded state, but from the zero-strain state, from which it contracts to a contracted state of smaller than the pitch diameter of interest (I included the before and after pictures of the results as well).

                         

                        In regards to the preload recommendation you've provided, do you simply put the components into contact without the interference?  Or do you build a simplified version:  basically, a hair-line profile section of each component, mate them together with no interference, then press them together?  Would the way I'm describing (if I'm even understanding you correctly) be inaccurate at all due to a lack of axisymmetric relation?  If not, I might try this out instead...

                         

                        Again, thank you for your help.  Please ask if any of this doesn't make any sense.

                          • Re: I'm having some difficulty with a shrink-fit contact (attempting to use thermal loading to relax the interference)
                            Jared Conway

                            pictures are there, but if a picture is worth 1000 words, the model is worth 1million

                             

                            and yes, the example that i state, components are not overlapping, that is only allowed for shrink fits

                            and we use 2 simplification so no hairline models need to be built

                              • Re: I'm having some difficulty with a shrink-fit contact (attempting to use thermal loading to relax the interference)
                                Joshua White

                                For this model, I am using 2-D simplification as well (axisymmetric).  I can't see any way of being able to put the components into contact without there being overlap, unless they, quite simply, were not fully "made-up".  Putting the two components together with a large amount of stand off, meaning the pin is displaced further to the right in the pictures, would allow the components to not be in overlap.  That's the only way, unless you made the pin (inner model) intentionally smaller and the box (outer model) intentionally larger so they would no longer have any interference.  But if you did that, then you would have to expand the pin and contract the box so they could build up proper interference (with no penetration contacts, of course).  Unfortunately, I think the contraction and expansion would change the geometry too much (particularly, if done thermally) and would distort the results with additional stresses and strains.  But I digress...

                                 

                                If you have any other input regarding how you would make that work, please educate me.  You probably know a better way to allow me to do what I'm trying to do, and since the thermal load is not working properly, I welcome your suggestions.  I am still really interested is why my thermal load isn't allowing the model to contract from an already-expanded, strained form.  Does anybody else experience this?  Is it only me?

                                  • Re: I'm having some difficulty with a shrink-fit contact (attempting to use thermal loading to relax the interference)
                                    Jared Conway

                                    thermal problem, have you tested on a very simple model like a block? that would eliminate your model from being the potential cause. and then also solid elements instead of 2d axisymmetric. this is what you'll need to report the issue to your var if you think there is a software issue.

                                     

                                    as for a better way to do this, i'm not really clear on what the problem is that you are trying to solve. there are a lot of descriptions about your setup and the problems that you're running into but there isn't a single shot of what your model looks like, a description of what the loading condition is or what your goals are from an analysis perspective. this might be helpful to try to come up with a strategy.

                                     

                                    and as i mentioned before, having a simplified model showing what you're trying to do so that osmeone could actually try it without having to build a model.

                                     

                                    have you spoken to your var on this? it might be helpful for you to talk to someone about it or have someone build a training plan around it. we do that for customers all the time. www.hawkridgesys.com