PDA

View Full Version : UMAT subroutine for Single Crystal

abhinav
2009-02-05, 09:31
Hello
I need a basic understanding from your side regarding UMAT code for Single crystals
I was following a paper on the same and it says that ABAQUS provides for each integration point, the values of total deformation gradient, F at the
current time t, the deformation gradient from the current natural conﬁguration, Fe at current time t, current value of the Cauchy stress, T and the state variables such as s(strain path length) and G(Plastic Deformation) at the current time step. It also sends an estimate of the total deformation gradient at the next time step (t + ∆t).

The UMAT supplies the values of the state variables such as s, Fe and G and Cauchy stress at time step t + ∆t and return it back to the ABAQUS. This process continues until a converged solution is obtained

Thus I would like to know how can we provide G, Fe and T at time step t + ∆t. I know how to get s from the constitutive equations.
Any help would be appreciated
Thanx

abhinav
2009-02-10, 23:41
To cut the story short, I would like to know if we have state variables at time t,
how can we use it to calculate new values of state variables at time t+dt?

Please tell me if we need to use some other subroutine to update the state variables??

Jorgen
2009-02-11, 20:25
You don't need any other subroutine. The UMAT or VUMAT subroutine need to perform that calculation.

The method to calculate the stress and state variables depends on the material model. What material model do you have in mind?

abhinav
2009-02-11, 22:22
I am trying to model the concepts given in the pdf at the following link
http://txspace.tamu.edu/handle/1969.1/4201
On pg. 77 of the above mentioned pdf.. u will find that they solved state variables using newton raphson method.
But their Newton-Raphson jacobian contains terms of state variables at time t+dt . Thus do we need to initialise state variables at time t+dt using some other subroutine like SDVINI??

Jorgen
2009-02-20, 09:54
I am afraid that I am not quite familiar with that model. I don't see, however, how using a SDVINI would help you initilize variables at t+dt. Perhaps I am not quite following your logic.

If you have a plan in mind, give it a try and let us know how it went.

-Jorgen

abhinav
2009-02-22, 00:17
Hi,
Thanx for going thru the model
I have more or less understood the logic, but have only one doubt that how do we use values of state variables at time t to calculate the values of state variables at time t+dt.
I mean how can we store the values of state variables of previous iterations to use them in next iterations??

I am presently trying to model it, if it wrks, i'll be happy to tell u all. :)

Jorgen
2009-02-23, 19:49
You don't have to store them - Abaqus will store the state variables for you if you return them back through the statev vector.

-Jorgen

Kumar Kurambakurash
2009-11-05, 08:05
NR is independent of any material model.