2 Replies Latest reply on Sep 4, 2015 5:38 PM by Stephen Callegari

    Mapping buckled mode shape into mesh for non linear buckling analysis

    Stephen Callegari

      Is there a means to map the buckled mode shape from an eigenvalue buckling analysis to a shell mesh of a pipe?

      This is one way of modeling a perturbation for a non linear buckling analysis.

       

      Other softwares can take the node location of the buckled mode shape and remesh to induce a

      perturbation in the column to get the bucklign started in a nonlinear analysis.

       

      If anyone has a literature reference for another method of inducing the perturbation, say by a force.

      that would be helpful as well.

       

      Thanks,

       

      Steve

        • Re: Mapping buckled mode shape into mesh for non linear buckling analysis
          Janko Stellaard

          Hi Steve,

           

          Interesting question. I don't think it is possible in SW Sim unless you apply a force that activates the buckling in the desired direction. Maybe you could try to check the eigenfreq in a frequency study. Buckling should occur when the eigenfreq is equal to zero. Would be trial and error though.

           

          I cant help you with literature.

           

          Janko Stellaard

          Cadmes BV

          The Netherlands

          • Re: Mapping buckled mode shape into mesh for non linear buckling analysis
            Stephen Callegari

            Janko,

             

            Thanks for the reply. Apparently, this technique is pretty robust.

            As I am sure you aware, the issue with nonlinear buckling is we need a "small"

            disturbance in geometry or load to induce the buckling process.

            The reason for this is, our FEA geometry is "perfect" and in reality the geometry is not.

            Also, the "perfect" FEA geometry and loading  needs a "kickstart" to the buckling process.

            Ovality in pipe, eccentricity of loads, and other manufacturing defects all contribute to an imperfect geometry or load

            which causes the structure to buckle sooner than if modeled with a linear buckling analysis with perfect geometry.

            The question then is how much disturbance load is required to design for non linear buckling.

            We dont want too large or too small of a disturbance.

             

            There are a couple of workarounds for applying the "disturbance" I am testing based on a literature search as follows:

             

            1. Apply a lateral load  (to induce buckling) with a magnitude of 0.50% of the buckling load found the eigenvalue (linear) buckling solution.

            2. Apply a lateral load such that the deflection by the lateral load is equal to ~L/300. Where L is the unsupported span.

             

            Does anyone have any comments or suggestions?

             

            Steve