Learning Forum

[Ammar Hussein]

Hi All,

I want to perform two curve fits in Workbench (One for Hyperelasticity and other for prony shear relaxation).

As far as I am concerned Ansys does not support 2 curve fits. As such, I need help on the steps I need to follow to apply the viscoelastic prony curve fitting in Mechanical APDL.

 

Thanks in Advance.

[Nanda Veralla]

Hello Ammar,

Can you be more specific what you're trying to do? Do you want to fit 2 curves for a single material?

If you want to use curve fitting using APDL, you can follow the ANSYS help here for hyperelasticity

4.6. Hyperelasticity (ansys.com)

and here for Viscoelasticity

4.7. Viscoelasticity (ansys.com)

Prony series complex modulus is used in ANSYS to represent small-deformation harmonic viscoelasticity. For more details see this Viscoelasticity in the Structural Analysis Guide

Thanks and Regards,
Nanda.

[Ammar Hussein]

Thanks for your reply Nanda,

Yes I want to have 2 curve fittings for one viscoelastic material.

The first I did was the hyperelastic Ogden curve fitting for my Uniaxial, Shear, and compression test data to define my material.

The second I want to do is the Prony shear relaxation since my model undergo relaxation, but unfortunatly, I found out that Ansys Workbench does not support 2 curve fittings.

As such I need to do it in Mechanical APDL. But I can't find any leads on how to do the viscolastic prony shear curve fitting for a beginner in APDL. 

[Bill Bulat]

I still need to get more details to satisfy your request, but I guess it might be prudent for us to first agree that it is possible to define combined hyperelastic and prony material models in MAPDL (see the two screen shots below): It seems possible to curve fit both viscoelastic and hyperelastic data to obtain the constants used to define material model constants:I see no prior mention of the TBFT command in this thread. Are you aware that this is central to curve fitting in MAPDL? Additionally, I could be wrong about this, but it may be necessary to obtain the constants for two seperate material numbers, then manually rewrite the commands that define the constants in such a way that they define both material models for the same material number.

[John Doyle]

If your relaxation data is stress vs time, this curve fit can be done directly in Engineering Data.  Please refer to online docs Workbench=>Engineering Data Users Guide=> Material Data=>Curve Fitting=> Viscoelastic Test Data...

If you have storage and loss moduli as function of frequency, then you are correct.  The curve fit needs to be done in MAPDL.  Please refer to Section 6.2 of the MAPDL Material Reference Guide...  It would also help to review the documentation on 'TBFT,,,' command in the MAPDL Commands Manual.  This procedure will result in Prony constants in the time domain (alpha and tau values), but the input is storage and loss modulii in frequency domain.  The MAPDL curve fitting results (alpha and tau terms), in this case, would need to be manually copied into WB-Engineering data. 

If you have access to the Ansys Customer Portal, please refer to KM 2036139 for a worked out example in MAPDL.

 

[Ammar Hussein]

Thank you Mr. Bulat and Mr. Doyle,

My relaxation data is Stress vs Time, and I am aware of how to curve fit it in the Engineering Data.

But, I also have already curve fitted an hyperelastic model to define my material using Ogden. As far as I am conserned, it is not possible to perform 2 curve fitting in the same material in the engineering data. Please correct me if I am wrong.

 

[John Doyle]

You have two sets of data: one for the rate independent stress-strain (Uniaxial, Shear, and compression) and one for the rate dependent relaxation (Stress relaxation vs time). 

Curve fit the rate independent ogden model first.  Then curve fit the rate dependent prony series.

These two mathematical expressions are combined under one material model to give you a large strain rate dependent material representation of your elastomer.

Your engineering data should look something like this:

[Ammar Hussein]

Thank you Mr John, this is now clear.

This might also lead to my last question regarding this query, in which I would like to know the possibility of prony curve fitting using Uniaxial Tension relaxation test data (Tensile stress vs Time) instead of shear data. Or if there is any relationship that might make it work.

Also, the possibility of using pure shear test data.

 

 

[John Doyle]

Since the alpha terms in the prony series are just unitless ratios of final/initial moduli, you can use equivalent stress vs time relaxation data directly in the prony curve fitter.  The moduli vs time relaxation data would produce the same result.