Finite Deformations and Plastic Materials

[MarkusB]

Hi!

Although this might be a stupid question:
  Why does parFEAP not work upon using finite deformations and plastic materials?

Input:

Code: [Select]

! .... omitted a few lines  ...

GLOBal
  FINIte

MATErial 1
  SOLID
    ELAStic isotropic 20000. 0.3
    PLAStic MISEs 200
    PLAStic HARDening 10 0

! ... more here ...


Output (number of processes = 2):
"
  *ERROR* Residual norm is NaN or Inf

  *ERROR* Residual norm is NaN or Inf
"

What would I need to change in the code to get it working?

Thanks.


Markus

[FEAP_Admin]

Markus,
  Does your input file work in serial?

[MarkusB]

Hi Admin.

Oh, I did not do that.
It does not work with serial FEAP. I get the same message.
This means there's a problem with my 'solver' configuration, right?

I want to do 10 proportional loading steps with a time-delta of 0.1 seconds and print the von Mises stress as well as the reaction forces.
The mesh is a simple cube (tetrahedronal mesh) and serves as a testcase.

If it is not too much to ask (my hope is you immediately spot the mistake I made): What am I doing wrong?

My config (parFEAP):

Code: [Select]

! ... mesh here
END

BATCh
  PROP,,1
END
  2, 2
  0 0  1 1

BATCh
  PETSc ON ! omitted when using serial FEAP
  DT,,0.1
  LOOP,time,10
    TIME
    LOOP, newton, 100
      FORM
      TANGent
      SOLVe
    NEXT
    STREss NODE
    Reac, all
  NEXT
END

STOP

Many thanks, you already saved me a lot of trouble.   

Cheers

[FEAP_Admin]

It is is hard to say what is going on but I suggest that you run the problem interactively.  Then you will be able to see how the nonlinear solver is working.  The likely problem is that your load steps are just too large.

[MarkusB]

Thanks!

I just had to adjust the parameters and dt. 

Solved...

btw:
  This is why I thought it (finite deformations + plastic material) would not work: http://tuprints.ulb.tu-darmstadt.de/1336/1/main.pdf  (See page 39; 7.1)

[FEAP_Admin]

Interesting post (albeit an old version of the code).
I read the test case and the reason the plasticity did not work had nothing to do with plasticity.   In the version he tested there was no parallel contact.  Parallel contact will be part of the next release (version 8.4).