Ansys Learning Forum Maintenance

NOTICE: We will be performing backend maintenance of our Learning Forum from April 5 to April 12, 2021. The result will be a new infrastructure but with little impact to user experience and design. Currently the forum is accessible in read-only mode as we complete our final migration. Thank you for your patience. For urgent issues please visit HERE.

Viscoelastic curve fitting from DMA data. — Ansys Learning Forum

Viscoelastic curve fitting from DMA data.

gouthambrajgouthambraj Member Posts: 1

We've been trying to model a hyperviscoelastic material by inputting hyperelastic (Yeoh 3rd order) and viscoelastic(Prony 4 parameter) parameters. We are facing high abnormalities with the results obtained on simulating shear relaxation test on static structural while applying those properties to the material(ramped loading for 1 second and allowed to relax). The maximum value obtained when I inputted both visco and hyperelastic parameters are too small0.20681 Pascal) compared to the maximum value obtained on inputting only the hyperelastic parameters(754.54 Pascal)

Result on both hyperelastic and viscoelastic parameter input

 Result obtained on inputting only hyperelastic parameters.

                            The maximum value obtained on using only hyperelastic parameters is comparatively closer to the desired output value. And it makes us believe that there could be error with the prony parameters we inputted. 

                            We obtained prony parameters by curve fitting our storage modulus vs frequency data of DMA ( obtained from a high impact journal) with the help of APDL prony curve fitting tool. We uploaded dynamic modulus(both storage and loss) along with corresponding frequency in the format they're demanding in the below mentioning video (modulus data in Pascals and frequency in cycles/sec) even though the frequency data given in journal is in rad/sec. 

I fixed time values and tried fitting but the values of A0 was most of the time smaller than A1 or A2. I believe that's not how it should be. But the prony values (a1,a2) were always less than 1.

So I applied those values to the workbench, and it displayed a shear modulus vs time data in the range of 1e-34 pascal which seem to be an absurd value for a pure ABS rubber.

                                    I have attached our DMA data (.csv) and the workbench archive we were working on. Is there any problem with the trend of our data?(The data was extracted from a figure given in a journal) . We even tried curve fitting with a normalized DMA data, and it showed the same trend.   

       (figure from which we extracted dynamic data)


                          Or is there any other subroutine available in ANSYS for curve fitting frequency domain data?


  • 1shan1shan Ansys Employee Posts: 262

    Hello @gouthambraj,

    In the last image it looks like you might have inserted the wrong elastic moduli. Could you check the data that you have inserted for the neo-hookean model especially the units? For small times your hyper elastic and visco elastic models should give you nearly the same result (which I dont see).



Sign In or Register to comment.