6 Replies Latest reply on May 23, 2012 10:59 AM by Blake Rudy

    Including Residual Stress In Analysis

    Andy Krispie

      Afternoon Folks,

       

      I am doing to FEA analysis on rolled steel W-Shapes (also known as I-beams) in axial compression.  I am wondering if there is a simple way to include residual stresses in the cross section.  I thought about trying to emulate the stresses using some sort of temperature gradients, but my feeling is there must be an easier way.

       

      I've includeda picture to illustrate the stresses that are created by hot rolling, for those who may not be familiar with hot-rolled structural shapes.

       

      Thanks very much.

       

      -Andrew

        • Re: Including Residual Stress In Analysis
          Andy Krispie

          Anyone have any thoughts?  Really hopeful that I don't need to resort to something like Ansys to include residual stress due to manufacturing...

          • Re: Including Residual Stress In Analysis
            Bill McEachern

            I suppose you could try adding some loads to impart the residual stress distribution though that may not be all that simple. The way it is typically done is to conduct a manufacturing simulation to obtain the distribution in an NL analysis and use that as the start in a subsequent analysis step. Another option maybe  to use the pressure vessel study: do an analysis that produces the residual stress and another with the service loads and use the pressure vessel study to combine the results with one of the available methods.

              • Re: Including Residual Stress In Analysis
                Andy Krispie

                Great thanks for the suggestion Bill.  I'll try playing with the pressure vessel study.

                 

                Do you know if there is a way to separate the flange and web plates of a typical steel section during the first pressure vessel study (so it's easier to impart the desired stress in the plates without affecting connected plates) and then reassemble them for the second load study so they can once again interact together?  Seems to me that it might be pretty finicky if the plates are connected - in my model I've just extruded a section from the toolbox within SW so the steel is solid and continous.

              • Re: Including Residual Stress In Analysis
                Andy Krispie

                Ok so I managed to use several independent static studies to create stresses in the flanges and the web, and then I used a pressure vessel study to combine them.  Now I would like to take the combined stress and use it in a nonlinear analysis with my axial loading.  These internal stresses will affect how elastic and inelastic buckling transform so it is important that my nonlinear study accounts for these stresses.  There is no option that I can find to include the residual stresses I created.

                 

                If I do like you suggested Bill and try and do a nonlinear study independently of the residual stresses and then combine them using a pressure vessel study, I am not confident that I will get a realistic buckling response, especially when I start incorporating some more complex geometry.

                 

                -Andrew

                  • Re: Including Residual Stress In Analysis
                    Blake Rudy

                    Is there something preventing you from doing what Bill suggested above..

                     

                    "The way it is typically done is to conduct a manufacturing simulation to obtain the distribution in an NL analysis and use that as the start in a subsequent analysis step." 

                     

                    Designate loads as desired.  Run through them consecutively while saving each run's results and choosing "restart" for each successive run.