8 Replies Latest reply on Nov 1, 2018 6:54 AM by Erik Kostson


    Mallikarjuna B. S.

      Can we do Non-linear frequency analysis in SolidWorks?

        • Re: Simulation
          Benjamin Modic

          Hello Mallikarjuna,


          There is currently no ability to run a nonlinear frequency analysis, however the Simulation Premium package allows you to run linear dynamic harmonic and random vibration studies.


          File:Simulation Matrix 2018.PNG


          • Re: Simulation
            Scott Stuart

            Dang. I thought you said "stimulation". I got excited.

            • Re: Simulation
              Erik Kostson

              Not heard of such a term as nonlinear frequency analysis. Frequency analysis assumes linearity so we can establish a dynamic solution based on a linear summation of modes (with certain amplitude). In nonlinear analysis superposition is not valid.


              Other than that the dynamic equations ([K]-[M]*w^2=0) solved by these solvers assume small deformation, no change in BC and linear material.


              If we want to include these nonlinear effects then a transient nonlinear solver can be used, but that does not use natural modes to produce a solution (as say in a harmonic response analysis where modes are typically used), it integrates the nonlinear dynamic equations. The results is a general time disp./stress history data (not modes and freq.).


              Thus if you have a real contact situation (not bonded) one would need to use the nonlinear transient solver, since in solidworks only free and bonded contacts can be used in a natural frequency analysis.


              Some FEA software like Strand7, Nastran, will allow one to use some form of contacts in the natural frequency solution, but how well that works really depends on the situation I think (many time they import the contact stiffness from a nonlinear analysis in to the natural frequency solver, which assume that contacts will not change during this linear analysis, thus small displacements about a contact equilibrium found in the nonlinear static analysis – so if contact status change s this is not valid anymore).

                • Re: Simulation
                  Bill McEachern

                  FYI: I run Abaqus and what it does is that  it can take an NL analysis (say bolt tightening) and then do a frequency analysis (more explicitly  for the OP it does an eigenvalue extraction of the stiffness matrix). However, what it assumes is that there is no change in any contact status from the previous analysis - if a contact is closed then it remains closed (ie bonded, tied or whatever you choose to call it) when the extraction is done. The value in this is that the span  of the contact patch is a better approximation for small amplitudes of vibration which what happens when vibrations start to emerge. It also allows any other pre-loads which may contribute to stress stiffening of softening effects to be included (SWX Sim can do some of this but not the contact status carry over)

                  • Re: Simulation
                    Mallikarjuna B. S.



                    Thank you for the info.

                    As per your statement, I understand that there is NO FEA solver, has facility to deal non-linear frequency analysis.


                    What is the best approximation method (whether bonding or rigid connection etc) for imposing contacts while doing a modal analysis in SolidWorks. just for e.g., we have a block of size 1mx1mx1m and is clamped onto the floor with few screws and assume that there is no gap between the block and the floor. In this situation, what kind of boundary condition you impose, whether you arrest at the screw region or the whole surface where the block and floor touches?, assuming that you only consider the block for analysis, treating that floor is infinitely rigid.