UMAT DFGRD0-1 and ELEMENT

[Ratatosk]

Hello,

I have written an UMAT subroutine which uses DFGRD1 to compute the right Cauchy strain tensor uses in my constitutive formulation

As I launch the simulations, my strain never increases.

8-node linear bricks, hybrid, constant pressure (C3D8H) elements are used. Do they provide the information of DFGRD0-1 to UMAT? Do I have to use another type of element? The hybrid formulation is necessary, since I am in the fully incompressible case.

Any suggestions?

Thank you,

Arne

[Jorgen]

Hello Arne,

I use DFGRD0 and DFGRD1 quite often in UMATS - they work fine. Why don't you rerun your simulation in displacement control and make sure that you can see the deformed shape using ABAQUS/Viewer.

Jorgen

[Ratatosk]

Hello Jorgen,

Yes, I managed to have a strain increment and fo DFGRD1 too. But I get another problem.. The simulation fails, because it seems to deverge. The load response gets really big!!

I am simulating a simple geometry (a box) with displacement boudary conditions at the bottom: the axis traction is locked (translation) on the bottom surface. At a point of the same surface, the two other translations are also locked. And at the other point of the same edge, a translation is locked to overcome rigid rotation arroud the axis of traction.

A uniform pressure load is applied at the top of the surface..

My constitutive law responds as an exponential form, therefore the jacobian is nul at rest and increases with strain increment.

ABAQUS responds with a high deformation gradient which is responded with a high stress from my UMAT, you talked about displacement control? What do you mean? Is it imposing a displacement instead of the load, to have a straightforward computation?

I have tried to impose a small displacement, but I still get errors.

I have printed all the variables in my UMAT, they seem right.

Any suggestion to control ABAQUSs deformation?

Thank you

Cheers,

Arne

[Jorgen]

My constitutive law responds as an exponential form, therefore the jacobian is nul at rest and increases with strain increment.
Hmm, are you saying that the material that you are modeling does not have any stiffness in it undeformed state? That sounds strange to me. What material are you modeling? Is the calculated Jacobian = 0 at time 0? That could certainly cause numerical problems.


ABAQUS responds with a high deformation gradient which is responded with a high stress from my UMAT, you talked about displacement control? What do you mean? Is it imposing a displacement instead of the load, to have a straightforward computation?
One thing you can try, and maybe you already have, is to apply a prescribed boundary displacement instead of a surface pressure.

If you have a problem with the Jacobian, you can always try to run you subroutine using an explicit simulation.

Best of luck,
Jorgen

[Ratatosk]

Sorry, what I said was not entirely right, the jacobian is not nul at rest..

I have tried to prescribe a very small displacement of the surface instead of a surface pressure... but it doesn't work..

The deformation chosen is 0.026 (dL/L)

*Step, name=load, nlgeom=YES
loading
*Static, direct
0.01, 10.,

Here is the last part of the corresponding msg file if that can help..

EQUILIBRIUM ITERATION 16

AVERAGE FORCE 1.980E-06 TIME AVG. FORCE 1.521E-06
LARGEST RESIDUAL FORCE -4.154E-09 AT NODE 630 DOF 2
INSTANCE: TILLON6-1
LARGEST INCREMENT OF DISP. 2.000E-04 AT NODE 1 DOF 3
INSTANCE: TILLON6-1
LARGEST CORRECTION TO DISP. -2.876E-06 AT NODE 407 DOF 2
INSTANCE: TILLON6-1
DISP. CORRECTION TOO LARGE COMPARED TO DISP. INCREMENT

COMPATIBILITY ERRORS:
TYPE NUMBER EXCEEDING TOLERANCE MAXIMUM ERROR IN ELEMENT

VOLUMETRIC 0 -1.015E-06 49
INSTANCE: TILLON6-1

***ERROR: SOLUTION FAILS TO CONVERGE IN MAXIMUM EQUILIBRIUM ITERATIONS
ALLOWED. FIXED TIME INCREMENTATION CHOSEN: ANALYSIS ENDS.



ANALYSIS SUMMARY:
TOTAL OF 10 INCREMENTS
0 CUTBACKS IN AUTOMATIC INCREMENTATION
26 ITERATIONS
26 PASSES THROUGH THE EQUATION SOLVER OF WHICH
26 INVOLVE MATRIX DECOMPOSITION, INCLUDING
0 DECOMPOSITION(S) OF THE MASS MATRIX
1 REORDERING OF EQUATIONS TO MINIMIZE WAVEFRONT
0 ADDITIONAL RESIDUAL EVALUATIONS FOR LINE SEARCHES
0 ADDITIONAL OPERATOR EVALUATIONS FOR LINE SEARCHES
1 WARNING MESSAGES DURING USER INPUT PROCESSING
0 WARNING MESSAGES DURING ANALYSIS
0 ANALYSIS WARNINGS ARE NUMERICAL PROBLEM MESSAGES
0 ANALYSIS WARNINGS ARE NEGATIVE EIGENVALUE MESSAGES
1 ERROR MESSAGES



JOB TIME SUMMARY
USER TIME (SEC) = 42.800
SYSTEM TIME (SEC) = 2.8000
TOTAL CPU TIME (SEC) = 45.600
WALLCLOCK TIME (SEC) = 84


Thanks.

Arne

[Jorgen]

Writing a UMAT can certainly be tricky. I recommend the following steps to debug your code:

:arrow: Did you get your simulation to take any increments (i.e. are there any lines in the *sta file?)
:arrow: Try to add print statements inside your UMAT subroutine that prints out the updated stress tensor values and the Jacobian matrix. Make sure these quantities are what you expected.

- Jorgen