Generalized plane strain elements (eg CPE4G) with UMAT

[RGX124]

I have a UMAT for the material behaviour that works with CPE4 elements but not with CPE4G elements.
I did mention the ref node and restrict DOF 4 & 5.
CPE4G works fine for elastic material or plastic ones given with the appropriate keywords, but NOT with a UMAT.
Does anyone know if I have to specify smthg in the UMAT so that it works for CPE4G elements?

Thanks

 

[xerf]

What exactly do you mean by saying it does not work with UMAT ? Could you mention some details ?

 

[RGX124]

With CPEG4 elements, the solution is diverging. Here is what I have in the msg file:
...
AVERAGE FORCE 0.785 TIME AVG. FORCE 0.785
LARGEST RESIDUAL FORCE -1.12 AT NODE 4 DOF 2
LARGEST INCREMENT OF DISP. 1.000E-05 AT NODE 2 DOF 1
LARGEST CORRECTION TO DISP. -4.140E-06 AT NODE 1000 DOF 3
FORCE EQUILIBRIUM NOT ACHIEVED WITHIN TOLERANCE.

AVERAGE MOMENT 0.425 TIME AVG. MOMENT 0.425
ALL MOMENT RESIDUALS ARE ZERO
LARGEST INCREMENT OF ROTATION 4.248E-37 AT NODE 1000 DOF 4
LARGEST CORRECTION TO ROTATION 3.886E-52 AT NODE 1000 DOF 4
THE MOMENT EQUILIBRIUM EQUATIONS HAVE CONVERGED


***WARNING: THE SOLUTION APPEARS TO BE DIVERGING.

***ERROR: TIME INCREMENT REQUIRED IS LESS THAN MINIMUM SPECIFIED. ANALYSIS
ENDS


With CPE4, it works fine.
I don't know why it (yhe UMAT) works with CPE4 elements and not with CPEG4 elements!!!!!

 

[xerf]

If the non-convergence problem were caused by UMAT , I would check the data passed as input by Abaqus for the two element cases. It might happen to be some differences since CPEG4 allows for out-of-plane strains and the two type of elements are meant for ,let say, slightly different constitutive response.

For CPEG4 , I think, you should have NTENS=4 from which first 3 strain/stress components are the the direct stress/strain componets 11,22,33. And the 4th component should be the shear 12 component (for both stress and strains storage arrays)..i.e. (NDI=3,NSHR=1).

Anyhow , I do not know what your constitutive response described inside UMAT is , so I cannot speculate more...
but I would check the input data passed in by Abaqus for the two element types.

 

[RGX124]

In fact, that was the problem. So I sort it out, because in my UMAT the out of plane strain increment was forced to be equal to zero (DSTRAN(3)=0) which is ok for CPE4 elements.
By removing this condition, it works fine for CPEG4 element.

Anyway, thanks for answering my question