4 Replies Latest reply on Sep 22, 2009 5:10 PM by 1-AHTUV0

    Rotating Gravity

      Hi All,

       

      I am simulating a rotating drive assembly that is spinning at a constant velocity (~1500 RPM) and wish to examine the axial loading of two bolts that help hold the assembly together. The assembly has compliance and the components are massive enough that gravity will meaningfully contribute to deflection along the assembly's length.

       

      I have successfully run a nonlinear dynamic simulation of the assembly with bolt connectors, centrifugal loading, and a gravitational field that alternates sinusoidally in one plane. However, I would like to impose a trully rotating gravitational field by describing a second acceleration vector that is out of phase with the first. How might I do this? Simulation Pro only seems to allow one time-dependency curve for all 3 gravity vectors and does not allow a second, separate gravity load. Linked values came to mind, but I'm not aware of a way to reference the current solution time value in a parameter expression (for example, to define a second sinusoidal function with a phase shift).

       

      This seems like a fairly routine issue (?), and any input would be appreciated.

       

      Brian

        • Re: Rotating Gravity
          Bill McEachern

          Hi Brian,

          Nice post and good question. I have no idea for a work around for you though I have wondered whent he limitation of one gravitry load would emerge. I suppose it would be nicer to have a rotating frame option available in dynamic studies and save you the need of specifying the components and the sinusoids altogether. I think you are up against a limitation in the software which obviously shouldn't be there. You should send it in through support and see what comes back. Sorry I can't be of more help.

          Bill

            • Re: Rotating Gravity

              Bill,

               

              Glad to hear that I'm not overlooking something silly or am crazy for thinking this should be core functionality. According to my VAR, there is no work-around other than to attempt to define the second gravitational component as a time-varying remote load. Having a "gravitational" remote load that acts only on the surface of the assemly seems a little hackneyed, but I'm curious to see how it results.

               

              Does anyone else think a rotating reference frame or at least accommodation of separate time curves for load vectors should be added? I'm not a rotating machinery engineer, but it seems like there are TONS of applications where gyroscopic forces need to be considered. I guess all of those engineers use ANSYS or one of the other programs that provides a rotating reference frame.....

               

              Anyone else agree? I've put in a ER but have no sense of its viability.

               

              Brian

                • Re: Rotating Gravity
                  Bill McEachern

                  The powers that be read this stuff so they know. For them it comes down to a matter of how many requests they get for a certain piece of functionality. I am not a fan of that  approach as there are certain things you just have to have to be a general purpose code. All they really have to do is allow another gravity vector to be applied or as you said a different time curve for each component of gravity that is specified. Which should not be all that hard to pull off one would think but you never know what you are going to break when you do it and what effort it is going to take to fix it and what noise level you are going to generate when something gets broken. The current approach for anything new is you have to wait till the next release - which in this case is going to be a year away - I just checked 2010 PR1 and no help for a rotating vector or another curve and ability to add another gravity load.  If they don't want to be a "learner tool" that seeds the market for other peoples codes they need to ramp up the efforts on these sorts of short comings in my humble opinion. If this was a CAD function that had no work around, at least in the old days, they would be all over this sort of thing and it would come out in a service pack. However, I think the CAD guys are running the show now and I get the impression they just don't appreciate how analysis fits into a company's work flow - some of this stuff you just have to have or peole go looking for another code. If that is what you are going to do - look at or use another code it would be good if you let your VAR know that so they get some feedback. Other than these issues, which in all fairness are getting rarer and rarer, the speed with which you set up and can get a job running can't be beat by any other tool I am aware: ofther tools don't even come close in this department and I am looking at them all the time. If you can't get where you want to go though that just doesn't matter all that much.

                   

                  Maybe a SolidWorks person would like to comment?

                  Ciao,

                  Bill

                    • Re: Rotating Gravity

                      I happen to come up along the director of Solidworks R&D during a bike ride about a month back, and we had a very nice conversation. If only I had foreseen this issue, I would have buzzed in his ear then!

                       

                      Bill, thanks for the comments. I completely agree. 90% of the time SolidWorks makes my job a lot easier, but that other 10% can make it very hard to justify having Simulation Premium, especially when we know what other full-fledged FEA packages can do. I don't doubt SolidWorks commitment to its simulation tools, but if it wants a truly integrated platform, at some point the look-and-feel improvements have to take a back seat to butressing core functionality.

                       

                      <end digression>