6 Replies Latest reply on Jul 18, 2012 11:13 AM by Hannah Spurgeon

    Simulating collision between two assemblies

    Hannah Spurgeon

      Hi. I am trying to do a simulation where an object at a given velocity strikes another assembly. I have spent months trying to get this to work, and it's just not happening.

      Basically, i have a vehicle seat, with a metal frame and a thick rubber lining. I have a dummy located in this seat. The other object is a non descript heavy metal 'slab', which will collide with the back of the seat at a given velocity. I have tried doing a drop test, but i can't figure out how to accelerate the metal slab while the seat remains stationary. i have also tried a non linear analysis, and have the same problem. Any pointers would be greatly appreciated! i have a lot riding on this....

        • Re: Simulating collision between two assemblies
          Bill McEachern

          What is the object of the analysis?

            • Re: Simulating collision between two assemblies
              Hannah Spurgeon

              to see the acceleration/displacement of the occupant in the seat (neck). The rubber is supposed to increase the change in momentum time experienced by the occupant, in order to reduce the acceleration of the occupant.

                • Re: Simulating collision between two assemblies
                  Bill McEachern

                  What Tony has stated is all close to the mark in my opinion if the all singing all dancing version is absolutely required. Depending on the time scales involved an implicit code make work just fine - if the time scales of interest (or that matter) are in the order of a few milliseconds or greater you are probably fine with an implicit code. If you could make some reasonabe simplfying assumptions you can probably get something done that might be useful for whatever it is you are doing with SWX Sim but it will likely be a bit trying to get it to solve. To make this practical on SWX Sim you would need to limit the contacts to the bare minimum. Do you really need the slab to actually hit the seat? Can an acceleration be used instead? You could do a far simpler model to try and get a handle on that - say hit another block with similar mass properties to the chair and occupant or somethig similar. Can you bond the body to the chair below the neck, make the geometry really simple, etc?

                  On the all siging all dancing version - I wouldn't start there even with LS-Dyna or any other code. Start really simple and figure out what and how you can get to solve on a really simple scale that looks reasonable then complicate with geometry. NL dynamic is probably the way to go. Drop tests limitations make it hard to do anything other than a straight drop/velocity impact onto an infinite plan.

              • Re: Simulating collision between two assemblies
                Anthony Botting

                Not to rain on the parade, but it sounds like a job for something like LS-DYNA by Livermore Labs (about $70K for the software, last time I checked). I do know the drop test in SW is kind of restrictive, in that the whole model is subjected to the accelerations. The nonlinear dynamic module in SW can handle these types of simulations, but the sheer number of nonlinearities is quite a lot to handle on a PC. I am assuming you wish to model all impacts with full contact, nonlinear materials response (including rubber deformations, and metal yield), "large" material deformations, rigid body displacements, and very small time frames. The analysis computations, taken together with an assumed voluminous number of elements, it not very well-suited for a PC.

                • Re: Simulating collision between two assemblies
                  Hannah Spurgeon

                  Yes, I agree LS-DYNA would be far better for a problem such as this; however I am stuck with Solidworks for the time being.

                  I am currently working on a nonlinear time dependant model, and have removed the slab, and simply applied the force to the back of the seat. It seems to be working a lot better. However, I can’t seem to figure out how to produce a ‘time history plot’ for a constant position on the occupant. I have selected a node, but it seems that the graph is plotted for this position in relation to the original drafting origin (….it does not remain on the occupant). Is there any way that I can plot a graph for a specific component or a constant point on a component in order to see whether the simulation has actually worked as it is supposed to?