8 Replies Latest reply on Dec 31, 2013 11:41 AM by Jared Conway

    Creep Constants

    Andrew Paino
    Andrew Paino Dec 30, 2013 10:50 AM

      I am trying to determine the constants described in the help file to model creep life with simulation. Does anyone have any experience doing this? So far, I haven't been able to find a valid solution using a system of four equations. I found the link showing an example problem where C2 is assumed to be 1 (http://www.motovated.co.nz/newsletter/aug13/FAQID__x394.pdf); the only problem with this is they say it is usually safe to assume C2 = 1, but don't say when it is not. Below is an image of the creep life data we're using to determine the constants. We will be using Heat No. WWS.

       

      Mar Creep.jpg

       

      Any help would be greatly appreciated.

        • Re: Creep Constants
          Shaun Densberger
          Shaun Densberger Dec 30, 2013 8:10 PM (in response to Andrew Paino)

          Any value less than 1 for C2 results in a smaller amount of creep strain, so it's a conservative assumptions to set C2 = 1. To calculate the values of C0 and C1 for Heat No. WWS, you'll need to use either the 1600 deg F or 1800 deg F data sample. This will also assume that you're analysis is done at a constant elevated temperature.

           

          For 1600 deg F you have two equations:

          1. 0.005 =[C0*(55 ksi)^C1]*(42 hr)
          2. 0.01 =[C0*(55 ksi)^C1]*(77 hr)

           

          Here you have two equations and two unknows, so a unique solution exists.

           

          For 1800 deg F you have two equations:

          1. 0.005 =[C0*(22 ksi)^C1]*(139 hr)
          2. 0.01 =[C0*(22 ksi)^C1]*(178 hr)

           

          Here you have two equations and two unknows, so a unique solution exists.

           

          Make sure to use consistant units when calculating your constants.

          • Re: Creep Constants
            Jared Conway
            Jared Conway Dec 30, 2013 8:08 PM (in response to Andrew Paino)

            C2=1 should be eliminated from the KB in the near term

            we recently did a creep project for a customer and interfaced pretty heavily with the developers and determined that was only valid for the case that it was written for.

            if you do want to keep C2=1 the method that shaun describes will work

            the other option is a guess and check

              • Re: Creep Constants
                Shaun Densberger
                Shaun Densberger Dec 30, 2013 8:26 PM (in response to Jared Conway)

                Jared, is there any additional information that came about from the developers out of this project?

                 

                If we want to leave C2 as an unknow constant, then we can still solve for C0, C1, and C2.

                 

                For 1600 def F at Heat No. WWS:

                 

                1. 0.005 =[C0*(55 ksi)^C1]*[(42 hr)^C2]
                2. 0.01 =[C0*(55 ksi)^C1]*[(77 hr)^C2]
                3. 0.02 =[C0*(55 ksi)^C1]*[(124 hr)^C2]

                 

                Here you have three equations and three unknows, so a unique solution exists.

                 

                For 1800 def F at Heat No. WWS:

                 

                1. 0.005 =[C0*(22 ksi)^C1]*[(139 hr)^C2]
                2. 0.01 =[C0*(22 ksi)^C1]*[(178 hr)^C2]
                3. 0.02 =[C0*(22 ksi)^C1]*[(214 hr)^C2]

                 

                Here you have three equations and three unknows, so a unique solution exists.

                  • Re: Creep Constants
                    Jared Conway
                    Jared Conway Dec 30, 2013 8:31 PM (in response to Shaun Densberger)

                    shaun, you are totally correct

                    in the case of this project we actually had creep curves that needed to be converted so your method could have worked but the curve would be skewed based on what points were chosen

                    we used the guess and check method to determine constants that matched the curve as much as possible

                    we also need to address that the length of the creep time far exceeded that of the curves and had to come up with a reasonable extrapolation method and also make sure our constants made a curve that matched reasonably

                     

                    i should also second your comment about making sure the units are consistent and also making sure they are in the unit system expected by the software. this is well documented in the solidworks kb.

                      • Re: Creep Constants
                        Andrew Paino
                        Andrew Paino Dec 31, 2013 8:48 AM (in response to Jared Conway)

                        Jared and Shaun,

                         

                        Thank you this is very good information. It looks like the calculation is fairly straight forward if the part is at a constant temperature, unfortunately the part I'm analyzing is not so I will need a CT value. Judging by what you both have said it doesn't sound like I have enough data, is there any way of determining all four constants with the data shown above?

                          • Re: Creep Constants
                            Jared Conway
                            Jared Conway Dec 31, 2013 11:13 AM (in response to Andrew Paino)

                            same as above

                            your constants are going ot characterize 2 curves

                            from there you should be able to build a more general curve that defines any temperature to a certain extent

                              • Re: Creep Constants
                                Andrew Paino
                                Andrew Paino Dec 31, 2013 11:38 AM (in response to Jared Conway)

                                Jared, I think I understand what you're saying to do. You mentioned in an above comment that you had to find a reasonable extrapolation method to make sure the constants matched, would you mind elaborating on how exactly you ended up doing that? The part I'm analyzing must have a creep life of 9,600 hours so I will definitely need to look into that. 

                                  • Re: Creep Constants
                                    Jared Conway
                                    Jared Conway Dec 31, 2013 11:41 AM (in response to Andrew Paino)

                                    BN is a power curve so we used that for extrapolation

                                    but we also had to use a linear interpolation in one case because we were at the very end of the curve and were fitting the curve to multiple other curves

                                     

                                    my suggestion would be to start digging for more material properties for that material, i don't think you have enough information to characterize the situation you're running into

                                     

                                    we're going to have a blog article up in the next month or so about what we learnt but it still needs some polishing before we can post it. in the meantime if you're looking for more one-on-one guidance, we can certainly help you based on your experience.