Search
Close this search box.
Notifications
Clear all

Coupled-temp-disp modelling of viscoelastic damping material

3 Posts
2 Users
0 Reactions
415 Views
Posts: 3
Topic starter
(@drexa916)
New Member
Joined: 10 years ago

Hi All,

I am new to the forum and new to ABAQUS CAE so please excuse my ignorance.

I am trying to model the self-heating effect of a viscoelastic polymer using the hyperelastic (NeoHookean) and viscoelastic (Prony coefficients with WLF) options in ABAQUS. I built a very simple 3D model of the polymer slab sandwiched between steel plates. I shear the polymer by applying a cyclic shear displacement (equivalent to 20% shear strain) on one of the steel plates while holding the other. I have a lot of difficulty making the model converge as it tends to stop at the very first step and gives me the error Too many iterations required and I get warnings saying I have excessive distortion in many of my elements. I can get the first few steps to converge when I apply a shear traction instead, but as soon as the deformations becomes large enough, I get the same warning and the analysis fails. I feel that the aspect ratio of my elements are generally good, but since the damping material is expected to go beyond 200% or maybe 300%, it worries me since the model cant even reach 20%. I have made my mesh finer but that didnt solve my problem except taking substantially longer to solve (since I need to have a 3D model for getting the heat flow analysis as well). I am not sure if the problem is in my definition of the Neo-Hookean properties or in the Prony series since I have determined these parameters outside of Abaqus as we only have pure shear tests at different frequencies and temperatures.

It would be great if someone can give me some suggestions. Any help is greatly appreciated! I can send the inp file via email if requested.

Jack

Topic Tags
2 Replies
Posts: 3998
(@jorgen)
Member
Joined: 5 years ago

As you mentioned, your simulation should work without significant convergence problems. My guess is that your material model is not defined quite right.

-Jorgen

Topic Tags
2 Replies
Posts: 3
Topic starter
(@drexa916)
New Member
Joined: 10 years ago

[QUOTE=Jorgen,11749]As you mentioned, your simulation should work without significant convergence problems. My guess is that your material model is not defined quite right.

-Jorgen

Hi Jorgen,

Thanks for your reply. I have lowered the displacement to 2mm, which is 20% shear strain and it worked. However, the temperature is way out... since the slab heated up beyond 100 degree after 10 cycles of 20% strain. Strangely, from the temperature profile, the heat seems like its flowing from the steel into the VE slab that is sandwiched in between, which makes no physical sense. I am guessing maybe I havent inputted the correct parameters for the viscoelastic material since I might not understand it fully, especially the inelastic heat fraction part. I tried again by setting the inelastic heat fraction parameter to 0 (previously it was set to 1), and this time it sees no heat up at all. I have attached an excerpt of the input file excluding the mesh since its too big for the attachment. The model is just a stack of 50mm x 50mm x 10mm slabs, with the outer slabs being steel and inner being VE material. Do you mind taking a quick look? Everything is in units of N, m and K. Finally, the Neo Hookean parameter is taken as G-inf/2 where G-inf is the static long term shear modulus in (Pa) and the Prony series parameters were calculated from an Excel sheet using frequency domain data we obtained from the manufacturer.

Thanks again.

Jack

Reply
Share: