10 Replies Latest reply on Feb 27, 2014 8:37 PM by Jared Conway

    Stress-strain curve requirements, outlined

    Justin Strempke

      I would like some help straightening out what exactly SW requires for material information in order to do NL analysis.

       

      Based on the Help, the majority of the stress-strain curves by analysis need to be input in the fashion of True Stress, Engineering Strain (see below):

       

      SW Sim S-S curve input info.JPG

       

      That's not a huge issue, as finding Engineering stress-strain curves and transforming the stress portion is just an equation.  My issue is how to actually enter them into SW.  It seems NL elastic pulls the yield, Young's mod, etc from the S-S curve, which implies it needs the full range of stress and strain, from zero.  Plasticity models require an elastic modulus and yield strength (though I'm not certain if the yield is used in calcs or just for FOS plots), and literature/pop-up errors state the FIRST stress point should be the yield stress (very bottom bullet in the above picture).  First question:

       

      1) What should the correlating strain be?  Does it start at zero, essentially requiring an offset from a derived chart for all points, or should the first strain point be a calculated value of Sy/E?  This would be an idealized model, but may not fit most published data for materials that use a 0.2% offset to determine yield.  Is this just a discrepency that must be accepted?

       

      2) Next for some models Log-strain is required to be entered vs. True stress, this may be a stupid question but Log-strain and True-strain are the same, correct?  I'm not familiar with the Log-strain term, and the transforms I've seen are the same as for Engr strain to True strain...

       

      3) Also is there a heirarchy in SW's use of available material data?  Say a stress-strain curve is input, but (for plasticity models) a tangent modulus and hardening factor are also input.  SW takes the plot first, correct?  Reading this:

       

      SW Sim S-S curve input info 2.JPG

      does tell me the curve will be used first, based on the wording "Material properties like elastic modulus, Yield strength, etc will be taken from the stress-strain curve when it is available and not from the material properties table."  So I tend to think the first stress/strain point calculates the elastic modulus for the linear portion...  But if the material is plastic-like (literally, like PVC, etc), the low-strain portion can have a very humped look which would take multiple point to properly define.  Will SW still use the first point, effectively the next point from 0,0, as the yield?  Is it pretty much irrelevant again beside FOS?

       

      4) Lastly, it is possible to use various material models in a single simulation?  Say you have a brittle glass which can run linear elastic, and metal which uses plasticity von-mises.  Can they be combined in this fashion, or do the glass need to be at least NL elastic, and can all NL model types be mixed?

       

      Sorry for the run-around questioning, I just would like this straight in my head once-and-for-all!  I'm pretty sure I have it lined out but then I'll read something that throws me off...  Thanks!

       

      Justin

        • Re: Stress-strain curve requirements, outlined
          Shaun Densberger

          "1) What should the correlating strain be?  Does it start at zero, essentially requiring an offset from a derived chart for all points, or should the first strain point be a calculated value of Sy/E?  This would be an idealized model, but may not fit most published data for materials that use a 0.2% offset to determine yield.  Is this just a discrepency that must be accepted?"

           

          Your stress-strain curve starts at the yield point, so for a table set of data, you have have your yield strain (0.002, 0.003, or something else, depending on your mateiral) and the stress at the point. What the stress value is will depend on whether you need to use engineering or true strain.

           

           

          "2) Next for some models Log-strain is required to be entered vs. True stress, this may be a stupid question but Log-strain and True-strain are the same, correct?  I'm not familiar with the Log-strain term, and the transforms I've seen are the same as for Engr strain to True strain..."

           

          Hmm, that's interesting. Log Strain and True Strain are, as far as I know, the same quantity.

           

          "3) Also is there a heirarchy in SW's use of available material data?  Say a stress-strain curve is input, but (for plasticity models) a tangent modulus and hardening factor are also input.  SW takes the plot first, correct?  Reading this:"

           

          As I understand it, if a S-S curve is defined, SW will use that; otherwise, SW will use the tangent modulus to create a bilinear S-S curve.

          • Re: Stress-strain curve requirements, outlined
            Jared Conway

            1) what NL material do you want to use? elastic or plasticity? have you checked the help and the solidworks KB? there are a couple of really good articles on this.

            2) what NL material are you using? is large strain enabled or not? same thing here, defined in the help and the KB.

            3) generally do not recommend inputting more information than you need, why would you put a stress strain curve in and bilinear? i believe that the software will go to stress strain first and if there is a heirarchy, i do not think it is documented.

            4) you can mix and match

            • Re: Stress-strain curve requirements, outlined
              Justin Strempke

              Shaun,

               

              Thanks for the general info.  #1 was as I thought, just wasn't sure if E was calculated based on the first data point with 0,0.

               

               

               

              Jared,

               

              1) I'm using a combination, elastic for a brittle material (glass) and plasticity for plastics/metal.  From what I gather, both should start at yield, it just depends on what type of analysis as to the catagory of stress and/or strain (log, true, engineering, etc).

               

              2) Glass, PVC (plasticized and unplasticized), SS304, to name a few in this model.  No large strain at this point, though I suspect it may be needed.  I am running large deformation.

               

              3)  I understand this, but in the general interest of building the material database such that a different analysis can be ran without having to find extra properties in the future, I was hoping to plug in whatever I found when I found it so that it would be convenient.

               

              4)  Awesome.  I didn't know if it would dumb-down to the lowest model type without notice.

                • Re: Stress-strain curve requirements, outlined
                  Jared Conway

                  1 and 2, i'd focus on reviewing the help and kb for the information necessary to develop the material information. each material behavior (not necessary physical material) you use in solidworks simulation has different rules. there is a good table of the inputs needed in the kb.

                   

                  3) to do what you're describing, you'd need way more materials in the pulldown list. von mises plasticity -bi linear, von mises plasticity - curve...etc. easiest way to do that is just have individual materials for it.

                   

                  4) glad i could help.

                    • Re: Stress-strain curve requirements, outlined
                      Justin Strempke

                      Help hasn't been much help on somethings, as I comb through it as best as I can and understand before I bring up a question.  I'm not the best about using the KB, need to work on that!

                       

                      For the materials, are you saying to create a new library material for each type of model, i.e. a SS304-Plasticity-von-Mises, SS304-linear-isotropic, etc?  That I would think would crowd the materials database.

                        • Re: Stress-strain curve requirements, outlined
                          Jared Conway

                          hi justin, another thing that might help is the nonlinear class and training manual. have you taken that?

                           

                          and yes, that is what i'm suggesting. it might crowd it but it ensure that you're using the appropriate properites. as i mentioned previously, to get the function you want, the software would need to allow you to enter all the features you want but then use the pull down list to select what combination should be used. you're just working around the way the software has been programmed by putting the information in the material library.

                            • Re: Stress-strain curve requirements, outlined
                              Justin Strempke

                              I have not taken the course but would like to (high cost though), and think I saw one from our VAR in our area this April.

                               

                              Seems to me there was an SR recently about material model types changing unwantingly, may be nice if the tree display could also include which material type along with the material for quick reference.  The material label may say said SS304-Plasticity, but the model may be stuck on Linear-elastic.  Probably irrelevant as I'm sure it's fixed!