I am faced with the problem of simulating the strain- __and__ frequency-dependent behavior of components made from fluorocarbon rubber. Mainly, there are two fundamental simulation tasks:

- Determination of the oscillatory response of a prestressed geometry (alteration of viscoelastic behavior with respect to unloaded geometry due to large strains)
- Determination of the transient creep behavior of a prestressed geometry (alteration of creep time constants and moduli due to large, changing strains)

Because of internal constraints, the simulations will have to be carried out with Ansys Mechanical. While researching possible options in terms of material modelling, I stumbled upon the hyperviscoelastic “Bergström-Boyce” model. However, I am not sure whether it is really applicable to my problem and – if it is applicable – which material tests are required to parameterize it properly. Using the vocabulary of small-strain mechanics (engineering stress, engineering strain, Young’s modulus), I am imagining fluorocarbon rubbers to possess a Young’s modulus (e.g. for uniaxial strain) which is both frequency- and (engineering) strain-dependent, i.e. E(f, epsilon). If I would look at the two-dimensional projection E(0, epsilon) of this three-dimensional surface at zero frequency, I would obtain the quasistatic Young’s modulus as a function of engineering strain (i.e. pure hyperelasticity). If I were to consider a projection E(f, x) at an arbitrary strain x, I would obtain the (linearized) complex Young’s modulus as a function of the frequency (i.e. linearized viscoelasticity). My question is: What “fidelity” (i.e. complexity of function, e.g. whether there is an inflection point or not) would the BB material model be able to cover in these two projections? A possible answer could be – for example – that the constant-strain projection is able to represent a third-order generalized Kelvin-Voigt element. Since I – based on existing experiments – would be able to plot these functions, I would be able to judge the applicability of the BB model. Furthermore, I would like to understand which experiments would have to be conducted (in which way) to parameterize the model.

Regards,

Enrico

]]>I was able to calibrate the BB model to my data. I have 3 curves at different strain rates all coming from uniaxial tension experiments. The model seems to fit the data great, but it always fails whenever I try to run the validation for FE simulation. I try changing the solver from default to Abaqus when doing that I can't even get it to run. Additionally, I exported the model as a script and imported it into Abaqus but whenever I run a job I get deformation on the material although the stresses are zero.

Another thing that I checked is the units, I set my units to be micro newtons for force and micrometers squared for length which matches my experimental data where my stresses are in MPa. But for some reason, I get values of the order 10^-6 in my stresses. This only happens in the graph, if I edit the data it does come out to what I had input. What am I doing wrong here?

I am new to all of this, and any help would be greatly appreciated.

I've attached some pictures that might be helpful please let me know if any other details are necessary.

400]]>i am modeling my crosslinked epoxy with a parallel network by using a yeoh model and a linear elastic + Bergstrom-Boyce Network-Dependent Flow. I have a question regarding the material model.

I am trying to model my softening, which is not a lot for my epoxy and an exponential evolution is enough. Is the plastic strain calculated there by using the plastic deformation gradient F_p and calculating the strain ln(V) ?

]]>I recently used the particle swarm method to calibrate the BB model in the incompressible case, and then I will consider the compressible case. If the material is compressible, the stretch part of the rate-equation sometimes seems to be negative (I haven't rigorously proven it). If there is indeed this problem, how can I solve it?

252Thanks, by advance,

Carl

]]>In my team, we implemented the Bergstrom Boyce model in a in-house 3D FE code to simulate the mechanical behavior of TPE and adhesives. Even though we implemented an implicit formulation with no restiction on the time step for the evolution law ( network) integration, we are looking for a "natural" time scale for the model to set up our simulations

My guess would be that it would depend on the constitutive parameters, but it might also be related to the strain rate dynamics of the problem at stake.

I am struggling to find any literature on such a topic, so if anyone has a clue on the topic, it would be appreciated

Thanks, by advance,

Quentin

]]>

I am trying to model an impact problem on LS-DYNA where polyurethane is used as an adhesive between a ceramic tile and metal backing. I want to setup a BB model, but am unable to find the material properties to do so. Would you be knowing where to look for them?

To setup my model, I'll be needing the following parameters:

- density
- elastic bulk modulus
- shear bulk modulus
- viscoelastic shear modulus
- elastic segment number
- viscoelastic segment number
- inelastic strain component
- inelastic stress component
- GAMO
- Reference Kirchhoff Series

Any assistance would be much appreciated!!

]]>I am currently trying to simulate the behavior of the thermoplastic materials like HDPE and PVDF. I started reading the book of Mr. Bergstrom to understand different options of numerical modeling. As i understand the best way of modeling of these material is using the Viscoplasticity model like TNM. But as you know these models are not implicitly introduced in the abaqus and we should develop the Subroutin Umat. First I would like to know if any one here has developed the script of Umat for the viscoplastic models like TNM or BB to share with us?

secondly I would like to know how to use the nonlinear viscoelasticity in abaqus without using Subroutine? do you have any example or tutorial?

Thanks

]]>

Obviously the BB-model uses a nonlinear dampener. Therefore one needs a large number of Prony terms to describe similar material behavior, which can be harder to fit.

]]>I am working on a continuum project attempting to implement the BB model for PTFE material. I am inquiring about Network B in this model. I am not certain where the Fe deformation matrix is coming from. It is not explicitly stated in Bergstrom, Boyce (2005) paper. I am trying to match the model to experimental data using MATLAB. Could anyone provide some assistance?

Thank you,

Kate

]]>I am working with a viscoelastic polymer for which I have a single strain rate tensile test result (without unloading behavior) and a 2 step relaxation curve.

My question: Is this data enough to calibrate BB model?

I assume the single tensile test result represents hyperelastic behavior and step-strain relaxation test helps to evaluate viscoelastic behavior. Please correct me if I am wrong

Thank you!

Regards,

Sandeep Ramini