Linear and non-linear viscoelastic modeling using UMAT

[Peeyush]

My eventual goal is to implement the non linear viscoelastic model of schapery into UMAT. But to gain some measure of confidence i was trying to implement the Linear viscoelastic model first using UMAT.

[Jorgen]

There are two ways you can attached equations. First, you can use the latex syntax specified on the following page (http://www.polymerfem.com/modules.php?name=Forums&file=viewtopic&t=133). Or you can create a separate pdf file which contains all equations, etc, and then post a link to that pdf-file here.

About your UMAT. Have you tried to:
- run it with a single element?
- add print statements to the code to verify that it does what you think it should?
- convert it to a VUMAT, which is easier to write since you don't need the Jacobian?

/ Jorgen

[Peeyush]

Dear Jorgen , My code runs and your suggestions really helped. I did not try vumat, looked new and complicated to me

However I still have some questions for this group.

QUESTION 1: *static and *visco in abaqus

I am confused between how these 2 commands are supposed to work.

If I take a single element and apply the *STATIC command with LOAD BC on one end and FIXED BC on another end, the LOAD is applied alomost INSTANTANEOUSLY and the element creeps which mean I DO NOT NEED a *VISCO command

If I take a single element and apply the *STATIC command with DISPLACEMENT BC on one end and FIXED BC on another end, the load is RAMPED during the step and which means I need a *VISCO command to look at viscoelastic/ non linear viscoelastic behaviour?

This is confusing me as to what options to use

QUESTION 2 : Although my UMAT works , I still have doubts about how it can invert equations etc....

The way I wrote my UMAT was that it would take stran and dstran and find stress solution. In such a case, my solution would be unique because i have to find N stress terms (N= number of terms in prony series) dependent on stran and dstran

But when I apply a load bc , there is no way N stress terms could be found out from 1 single load value (hence implied by a non unique solution) because the N stress terms are by definition a function os stran and dstran and the program prescribes load and not dispalcement.

So I am not sure how abaqus can perform such detailed calculations and still find a unique solution.

Please help

[Jorgen]

Here are a few comments:

(1a) The *STATIC command is only used for certain built-in material models in ABAQUS. For example, the creep, swelling, and viscoelasticity model. It is typically not used for UMATs.

(1b) A load BC is not always instantaneous. You can apply it using any amplitude definition you like.

(1c) You don't need *Visco to study viscoelastic effects when you use a UMAT.

(2) I am not quite sure I understand your question. Your UMAT calculates the stress from the strain state (and strain increment). If your constitutive equations are well formulated, then the solution will be unique unless you have a geometric instability (buckling) or an unstable material.

- Jorgen

[Peeyush]

In abaqus viscoelastic modeling in the example problems , section1.1.11, they follow *STATIC followed by *VISCO, I guess thats why I ended up using these 2 commands with UMAT .

In general whats the best command to use with a UMAT?

Peeyush

[Jorgen]

I typically use:

*STATIC when inertial effects can be ignored, and
*DYNAMIC when inertial effects are important and for some reason I don't want to use ABAQUS/Explicit.

- Jorgen

[hhspiny]

In abaqus viscoelastic modeling in the example problems , section1.1.11, they follow *STATIC followed by *VISCO, I guess thats why I ended up using these 2 commands with UMAT .

In general whats the best command to use with a UMAT?

Peeyush

I believe VISCOUS is only a turn-on switch for many of the built in model of abaqus, to activate the time dependent feature. if you do umat yourselve, then you don't need to be constrainted by it.

[fakoyafemi]

My eventual goal is to implement the non linear viscoelastic model of schapery into UMAT. But to gain some measure of confidence i was trying to implement the Linear viscoelastic model first using UMAT.

Hi, Based on your above past, I am tryung to develop a similar code using linear viscoelastic to implement the schapery nonlinear.

I have run into quite a lot of stops and quite confusing. I wanted to ask if you were successful at developing this code/program and if i could have a pdf copy as a guide?

Thank you

[ali rabbani]

hi sir i m a student of civil engg andi dont know how to make a subroutine ..please i want UMAT code ...