15 Replies Latest reply on Aug 20, 2018 8:54 AM by Bill McEachern

    Stress Strain curve

    Peter Dratwa

      Hi all,

      I have few doubts about in defining SS curve. I have taken the tension test of AISI 301 steel. Which is 3/4 Hard. I got the following True stress strain curve.

      But when i take the account of the Ramberg Osgood curve fitting then curve shifted to bit above.Which is below

      Or i have to choose the another curve fitting.

      Can anyone explain how it affect simulation. Because if i choose normal SS curve without curve fitting then i get the warning  regarding ETan.

       

      Regards

        • Re: Stress Strain curve
          James Riddell

          A true-stress, true-strain curve would not dip down at the end like your blue curve shows.  Is that just the raw data?  Did you try converting the engineering stress/strain data as the help document describes?  Since you seem to have data points it would be easy to generate in a spreadsheet.

            • Re: Stress Strain curve
              Peter Dratwa

              No its true stress strain but i at the time i just convert the engineering stress strain in to true stress strain with following formula.

              σt=σ*(1 + ε)

              True Strain (εt) =ln(1 + ε)

              but i think we have to take in account of plastic strain.

                • Re: Stress Strain curve
                  James Riddell

                  Well, you have done it wrong then.  A true stress-strain curve of a ductile material will never curve down at the end.

                    • Re: Stress Strain curve
                      Peter Dratwa

                      thank you for the correction. Yes you are right. There was some error in the reading from the test. But this changes just the end part. Do you have any idea about best fit for the modification of curve.

                        • Re: Stress Strain curve
                          James Riddell

                          It's been a few years since I converted values for a non-linear problem I was investigating.  However, I dug this up (https://info.simuleon.com/blog/converting-engineering-stress-strain-to-true-stress-strain-in-abaqus) which seems to give a good overview that is similar to what SWS needs.  There is even a spreadsheet to get you to the UTS.  Of particular interest to your problem may be what is listed under "Important note 2".  Let us know if it works.

                            • Re: Stress Strain curve
                              Peter Dratwa

                              So If I consider Ramberg-Osgood equation in to account to calculate plastic strain then also I have to do subtract the elastic strain from total stain ?

                              I have done following

                              First I calculate Engineering strain from Ramberg-Osgood euation

                              Then I calculated the True Strain by =LN(1+ Eng. strain)

                              But when I read link you sent , They have mentioned that we have to subtract the elastic strain to calculate the actual plastic strain.

                              Does that mean I have to again subtract the elastic strain from Total strain ?

                              Can you please explain that.

                              • Re: Stress Strain curve
                                Peter Dratwa

                                Hi again,

                                I had a look over the link you sent. I have few doubts. In this link they talk about abaqus. First they select zero stress and zero strain ,which we cant do this in solidworks. So first point is Yield point and strain value at this point should be plastic strain or total true strain? Because then it will affect the Youngs Modulus. Which is different if we consider two different values. And same thing affect when we do unloading.

                                Regards

                                  • Re: Stress Strain curve
                                    James Riddell

                                    Well, the Young's Modulus does change since it is the slope of the curve.  Yes, that was from an Abaqus source but the theory is the same.  You still need true-stress/true-strain.  To simplify, sometimes the t-s/t-s curve would be estimated by one E up to yield then a second E past the YS point.

                                      • Re: Stress Strain curve
                                        Peter Dratwa

                                        So it means, first point that is Yield point and True strain at this point but after values is like true stress but plastic strain ? In abaqus they consider strain as zero at Yiled point.

                                          • Re: Stress Strain curve
                                            James Riddell

                                            Where are you getting 'plastic strain'?  It's true-stress and true-strain.  Yes, it is in the plastic region but the input needs to be true strain.

                                              • Re: Stress Strain curve
                                                Peter Dratwa

                                                Plastic strain i calculated by

                                                Which is less. In abaqus they consider Strain at Yield point as zero.

                                                  • Re: Stress Strain curve
                                                    James Riddell

                                                    It has been quite some time since I used Abaqus, however, the principle remains the same.  I seem to recall that their software accounts differently for the recoverable elastic strain and it's just a numbers thing to 'start over' at zero.  You still need to define the true strain.

                                                    • Re: Stress Strain curve
                                                      Bill McEachern

                                                      I do not believe this is correct. In abaqus when you define a plastic material you define the Younge's modulus and the Poisson's ratio. You also define the yield stress and you also define the yield stress as the point of zero plastic strain, not strain as mentioned above. The yield point and the zero plastic logarithmic  strain is in a table where you can add further pints to define a true plastic portion of the stress strain curve. So in abaqus, as in SWX, you make up a stress strain curve point data set with the yield point being the zero plastic strain location. You then go from there (that is to say you subtract the elastic strain from the total true strain data and enter that as the true plastic strain).

                                                      Given the shape of these things it isn't to hard to figure out why the bi-linear approximation (i.e. use of a tangent modulus) is frequently good enough. Further, in Abaqus if no other points are added beyond the yield point to the stress and plastic strain table then a perfectly plastic approximation is used. This is frequently use as a very conservative approximation to a stress strain curve. However, it can lead to stability issues if the plastic region can not redistribute the load to lower he stress below yield - which is why it is used as a conservative approximation.

                                                        • Re: Stress Strain curve
                                                          Peter Dratwa

                                                          Thank you for the answer. So if I defined the Youngs Modulus plus True Stress - True Plastic strain. Then Simulator will take the value of Youngs Modulus from Table till Yield Stress. Is it correct ? Because my approximation was that , when we defined stress strain curve then all the value simulator take from Curve not from the Table.  (I am talking about Von Mieses Plasticity Model in Solidworks). Or it uses values from both table and curve (like in abaqus). But still we cant assign zero value in Soldiworks as starting point.

                                                            • Re: Stress Strain curve
                                                              Bill McEachern

                                                              I may have been in error on the SWX curve input. I believe the first point in the stress strain curve definition table is the yield stress and the total strain and then you just keep adding points of true stress and true strain (total) - sorry its been awhile. Maybe some one can chime in on this to confirm this is correct. The material model does alter what is input.