9 Replies Latest reply on Nov 11, 2013 11:03 PM by Jared Conway

    Simulating a ramp?

    Mildred Mattice

      For example, I would like to simulate the rectangular block traveling at 1 m/s (in the direction towards the triangular block, parallel with the bottom plate) and hitting the stationary triangular block.

      Capture.PNG

       

      What is expected to happen:

      The rectangular block should hit the triangular block, and be deflected somewhat.

       

      What I want to measure:

      The stress on the wall of the triangular block. I want to make this block hollow, and would like to run simulations to see just how thin I can make the walls/supports before such an event results in a failure.

       

      Questions:

      1) How do I set the mass of the rectangular block? For example, I want the block to be 1kg with an impact footprint of 1cm x 1cm. Do I do m = pV, and just calculate for the depth of the block, or is there an easy way to just set the mass for this part?

       

       

      2) How would I set this simulation up? I tried non-linear dynamic, but the rectangular block does not "deflect"; it stays in its plane (despite me not putting any displacement boundary conditions)... it just keeps entering the triangular block, without any deflection.

      Settings:

      -No penetration between the inclined surface of the triangular block, and the surface closest to that block on the rectangular block, as well as the bottom of the rectangular block.

      -0mm translation normal to the bottom plate of the triangular plate (to fix it in the y-direction)

      -0mm translation in the direction of the bottom plate of the triangular plate (to fix it in the x-direction)

      -roller slider on one of the sides of the triangular plate (to fix it in the z-direction, but allow for deformations)

      -Velocity of 1 m/s on the rectangular block, heading towards the triangular block

      -Meshing on lowest, draft quality for quick calculations as a test

      3.PNG

      But the results:

      4.PNG

      And:

      5.PNG

       

       

      Any ideas?

        • Re: Simulating a ramp?
          Mildred Mattice

          I tried adding a roller/slider property on one of the sides of the rectangular block (such that it can't move in the z-direction), and the simulation completed with no errors, but the block does not move:

          6.PNG

          • Re: Simulating a ramp?
            Mildred Mattice

            Here is an absolute great example of what I want to do:

            http://www.youtube.com/watch?v=RT_xT3b58Zc

             

            I just have no idea how to do it...

              • Re: Simulating a ramp?
                Anthony Botting

                HI Mildred: You can do this but the NL solver is very sensitive to finite displacement with no deformation, esp. at the start of the analysis (the nonlinear dynamic solver is not really designed to solve for that case). However you should be able to get around it if you put the block right up against the triangular component. Provide an initial velocity to the rectangular block. This way, the two components start-out already in-contact.

                Since the rectangular block has a sharp corner touching the flat face, the solver might have some difficulty with the contacts. You can try beveling the initial contact edge of the rectangular block such that a larger surface starts-off in-contact with the triangular block face.

                 

                For quick turn-around trial runs you can also take advantage of symmetry by cutting the components along a centerline vertical plane in the longitudinal direction and applying a sliding bc on the cut faces.

                  • Re: Simulating a ramp?
                    Mildred Mattice

                    Thank you for your response.

                     

                    This is a simplified, example model, to test the feasibility of the simulation, before spending time on modeling the full, complex model in SolidWorks. I suppose even in this simplified model, I could save time by using symmetry .

                     

                    I will try what you suggested: fillet/round the retangular block corner face, and try running it.

                  • Re: Simulating a ramp?
                    Jared Conway

                    Why not fix the bottom of the triangle just to simplify the restraints?

                     

                    Did you use initial velocity? I haven't had much luck with that. I'd use another method. Often the parts don't move with initial velocities.

                     

                    Regarding mass, the software uses density and volume of the part.

                     

                    You may want to post the model for a bit more help.

                      • Re: Simulating a ramp?
                        Mildred Mattice

                        Thank you for your response as well.

                         

                        I do not want to fix the bottom of the triangle, as I want the trianglular base to be able to deform. It is to my knowledge that if I use fixed geometry, the model will not be able to expand or contract in the said axes of that plane.

                         

                        As for the intiial velocity, I too haven't had much luck with that... it doesn't seem to move!

                         

                        With the mass, I suppose I have to use m = pV = p*(l*w*d), and just calculate the depth...

                         

                        And finally, as for the model, this was an over simplification, and not the final, intended, simulation.

                          • Re: Simulating a ramp?
                            Jared Conway

                            I do not want to fix the bottom of the triangle, as I want the trianglular base to be able to deform. It is to my knowledge that if I use fixed geometry, the model will not be able to expand or contract in the said axes of that plane.

                             

                            >what do you wnat to expand/contract? what do you think your current setup allows? also my recommendation here for fixed is to simplify the problem to get started. you can add more complexity later

                             

                            As for the intiial velocity, I too haven't had much luck with that... it doesn't seem to move!

                             

                            >it should but i have seen this problem before dependent on the way you've setup other things. posting the model is the best way to  troubleshoot this.

                             

                            With the mass, I suppose I have to use m = pV = p*(l*w*d), and just calculate the depth...

                             

                            > i'm not sure what you're looking to do here. mass is taken into account by the software automatically. unless you're trying to get a specific mass by back calculating the density.

                             

                            And finally, as for the model, this was an over simplification, and not the final, intended, simulation.

                             

                            > this is the best way to start any new analyses. especially if you're new to the study type. i'd even recommend trying to make this a static analysis to start with an idealized impact. that should help you understand what trends to expect from the dynamic.

                              • Re: Simulating a ramp?
                                Mildred Mattice

                                Jared Conway wrote:

                                 

                                >what do you wnat to expand/contract? what do you think your current setup allows? also my recommendation here for fixed is to simplify the problem to get started. you can add more complexity later

                                Because it is fixed in the x-y direction, it won't be able to move in said directions. However, it can move in the z-direction. This is what I expect the "roller/slider" and the "0mm translation" options to do.

                                 

                                 

                                Jared Conway wrote:


                                > i'm not sure what you're looking to do here. mass is taken into account by the software automatically. unless you're trying to get a specific mass by back calculating the density.

                                The final model (not this over simplification) must be able to withstand a 5kg round object hitting it at 10 m/s. So I was wondering if I can just create a spherical ball (of any size) and give it a mass of 5kg, instead of choosing a material, and making a specific-sized spherical ball.

                                 

                                 

                                Jared Conway wrote:

                                 

                                > this is the best way to start any new analyses. especially if you're new to the study type. i'd even recommend trying to make this a static analysis to start with an idealized impact. that should help you understand what trends to expect from the dynamic.

                                Yup. I've never done dynamic analysis in SolidWorks, so trying this out. Problem with static is that I do not know what type of force will be exerted by the round object.

                                 

                                For example, how would you translate the kinetic energy of that 5kg object (traveling at 10 m/s) into a static force? I guess you can make the kinetic energy = work required to stop the car such that:

                                0.5*m*v^2 = Favg*d

                                Favg = (0.5*m*v^2)/d

                                 

                                So if I set the distance required to stop the ball at 10cm, that would give me:

                                Favg = (0.5*5kg*(10m/s)^2)/0.1m = 2500N

                                 

                                That, however, is the average force, and is also of it stopping at 10cm. I have no idea if this assumption of stopping at 10cm is valid or not, because the design will have a curved wall.