Abaqus 6.7: Temperature viscoelastic material

[gunkerle]

Hello everyone,
I am trying to simulate viscoelastic material behaviour using Abaqus version 6.7.

As a test I am using a cyclic load process about 50-200 load cycles which is applied to block material made out of polyurethane. Plastic behaviour of the material is not involved.

Due to the speed of the loading process I am using explicit fully thermo-mechanical coupled calculations. My problem is now that there is no rising of the temperature or variation of the temperature at all in my simulation.

The calculation settings are as follows:
One surface is completely fixed (encastre) and on the opposite surface of the block a cyclic pressure load is applied. The initial temperatur is set to 10°C (283°K) via boundary condition. Values for the conductivity, specific heat, thermal expansion, "dynamic E-Moduly" in terms of a prony serie, poissons ratio are all given material parameters.

Due to the manual the option to use predefined field variables to prescribe an initial temperature condition can not be used in fully coupled explicit calculations.

My question is now if there is a heating option due to pure viscoelasting material properties implemented in abaqus.

I am asking because I found in the manuals only references that plastic behaviour is contributing to the internal heating and thus leading to a rising of the temperature but I found no description that and if how viscoelastic (dissipative) energy is contributing to the interal energy and thus lead to a variation of the temperature.

Which other options are available to simulate temperature dependent material behaviour?

ciao Gunkerle

[Jorgen]

I typically use my own user-material models when I simulate cyclic loading and the corresponding temperature rise due to internal energy loss in the material. It should work for the linear viscoelastic model too (which it sounds like you are using). One thing to check is that your material parameters are appropriate. If they are off, then you might get a case in which no energy is dissipated during the rapid cyclic loading.

- Jorgen

[gunkerle]

I typically use my own user-material models when I simulate cyclic loading and the corresponding temperature rise due to internal energy loss in the material. It should work for the linear viscoelastic model too (which it sounds like you are using). One thing to check is that your material parameters are appropriate. If they are off, then you might get a case in which no energy is dissipated during the rapid cyclic loading.

- Jorgen

Hellp Jorgen,
thus principally the pure viscous material behaviour should lead to a variation of the internal energy. I just wondered because in the manual only examples of contributions with respect to plastic energy losses are mentioned.
I am using linear visco elastic model to describe the material. For whole elements I get some amount of viscous dissipation energy which is increasing during the cyclic load process. But it seems to be that the amount is not enough to increase the temperature. What I am wondering about is that there are not even minimal variations of the temperature.


Would it be possible that you send me an example of one of your material models?

[Jorgen]

My material models are only commercially available (http://polymerfem.com/index.php?pageid=umats).

One way in which you can perhaps estimate the temperature increase is to take the dissipated energy, and the specific heat capacity of the material, and from that directly calculate the temperature increase. This calculation would be done outside the FE software and is based on adiabatic conditions.

- Jorgen