VUMAT and element deletion

[shripadtokekar]

Hi to all
i have some questions I am doing quasi-static simulation of lateral crushing of composite tubes in ABAQUS/Explicit. I am using VUMAT for failure criteria coding. after satisfying the failure criteria i am making the state variable equal to zero. My question is whether i am doing correct. and what values i should assign for state variable initially?? or making state variable zero will delete the element irrespective of whatever valure it contains before??
I am here putting my VUMAT code so please suggest me correction. thank you

subroutine failure (nblock, stateold, dmgfibreT, dmgfibreC, dmgMatrix, stress,statenew)

include 'vaba_param.inc'
parameter(zero=0.do, one=1.do, half=0.5do)
parameter i_xx=1
i_yy=2
i_xy=4
n_s33=4

dimension statenew(nblock,1),dmgfibreT(nblock), dmgfibreC(nblock), dmgMatrix(nblock), stress(nblock, n_s33)

c ft1,f2t-tensile strength in 1 &2 direction, f1c,f2c-compressive strength in 1 & 2 direction, f12-shearstrength


f1t=
f2t=
f1c=
f2c=
f12=

character*80 cmname

do K=1, nblock

S11=stress(K, i_xx)
S22=stress(K, i_yy)
S12=stress(K, i_xy)

c Chang-Chang Failure criteria.
c Evaluate fibre modes

if(S11.gt.zero) then

c Tensile fibre mode

rft=(S11*f1t)+(S12*f12)
if(rft.ge.one)then
statenew(K,1)=zero*stateold(K,1)

dmgfibreT(K)=one
else statenew(k,1)=stateold(k,1)
end if

c Compressive fibre mode

rfc=(abs(S11)*f1c)**2
if(rfc.ge.one)then
statenew(K,1)=zero*stateold(K,1)
dmgfibreC(K)=one
else statenew(k,1)=stateold(k,1)
end if
end if

c Evaluate Matrix failure mode

if(S22.gt.zero)then
rmt=(S22*f2t*f2c)+(S12*f12)**2+(f2c-f2t)*S22*f2C*f2t
statenew(K,1)=zero*stateold(K,1)
else statenew(k,1)=stateold(k,1)
endif
enddo

return
end

[Jorgen]

Your approach sounds fine. I would make the state variable initially equal to 1, and then make it 0 when failure occurs.

- Jorgen

[shripadtokekar]

thank you sir for ur reply i tried to implement my VUMAT subroutine with my input file. it compiles correctly but further it gives the linking error so please help me where i am wronge following are my input and VUMAT file
sory once again sir because I am asking so many question which are not that difficult but please help me

Input file
*Heading
** Job name: lateralcomp Model name: Model-1
*Preprint, echo=NO, model=NO, history=NO, contact=NO
**
** PARTS
**
*Part, name=Part-1
*Node


3040, 9.89936161, 7.9556675, 148.
*Element, type=S4R
1, 1, 7, 229, 190

2999, 3039, 3040, 175, 174
3000, 3040, 212, 6, 175
*Nset, nset=_PickedSet2, internal, generate
1, 3040, 1
*Elset, elset=_PickedSet2, internal, generate
1, 3000, 1
** Region: (Section-1:Picked)
*Elset, elset=_PickedSet2, internal, generate
1, 3000, 1
** Section: Section-1
*Shell Section, elset=_PickedSet2, composite, section integration=GAUSS
0.375, 2, Material-1, 0.
0.375, 2, Material-1, 90.
0.375, 2, Material-1, 0.
0.375, 2, Material-1, 90.
0.375, 2, Material-1, 0.
0.375, 2, Material-1, 90.
0.375, 2, Material-1, 0.
0.375, 2, Material-1, 90.
*End Part
**
*Part, name=Part-3
*Node
1, 0., 0., 0.
2, 8.37348557, 0., -79.5605698

1070, -61.0930634, 0., 8.76385689
1071, -61.1417084, 0., 4.33923721
*Element, type=R3D4
1, 10, 11, 190, 185
2, 11, 12, 191, 190

1018, 1069, 1070, 872, 871
1019, 1070, 1071, 873, 872
1020, 1071, 97, 96, 873
*Node
1072, 0., 0., 0.
*Nset, nset=Part-3-RefPt_, internal
1072,
*Elset, elset=Part-3, generate
1, 1020, 1
*End Part
**
*Part, name=TOP
*Node
1, 0., 0., 0.
2, 8.37348557, 0., -79.5605698

1070, -61.0930634, 0., 8.76385689
1071, -61.1417084, 0., 4.33923721
*Element, type=R3D4
1, 10, 11, 190, 185
2, 11, 12, 191, 190

1019, 1070, 1071, 873, 872
1020, 1071, 97, 96, 873
*Node
1072, 0., 0., 0.
*Nset, nset=TOP-RefPt_, internal
1072,
*Elset, elset=TOP, generate
1, 1020, 1
*End Part
**
**
** ASSEMBLY
**
*Assembly, name=Assembly
**
*Instance, name=Part-1-1, part=Part-1
*End Instance
**
*Instance, name=TOP-1, part=TOP
0., 12.7, 75.
*End Instance
**
*Instance, name=Part-3-1, part=Part-3
0., -12.7, 75.
*End Instance
**
*Nset, nset=_PickedSet13, internal, instance=Part-3-1
1072,
*Nset, nset=_PickedSet14, internal, instance=TOP-1
1072,
*Elset, elset=__PickedSurf9_SPOS, internal, instance=TOP-1, generate
1, 1020, 1
*Surface, type=ELEMENT, name=_PickedSurf9, internal
__PickedSurf9_SPOS, SPOS
*Elset, elset=__PickedSurf10_SPOS, internal, instance=Part-1-1, generate
1, 3000, 1
*Surface, type=ELEMENT, name=_PickedSurf10, internal
__PickedSurf10_SPOS, SPOS
*Elset, elset=__PickedSurf11_SNEG, internal, instance=Part-3-1, generate
1, 1020, 1
*Surface, type=ELEMENT, name=_PickedSurf11, internal
__PickedSurf11_SNEG, SNEG
*Elset, elset=__PickedSurf12_SPOS, internal, instance=Part-1-1, generate
1, 3000, 1
*Surface, type=ELEMENT, name=_PickedSurf12, internal
__PickedSurf12_SPOS, SPOS
*Rigid Body, ref node=TOP-1.TOP-RefPt_, elset=TOP-1.TOP
*Rigid Body, ref node=Part-3-1.Part-3-RefPt_, elset=Part-3-1.Part-3
*End Assembly
*Amplitude, name=Amp-1, definition=SMOOTH STEP
0., 0., 0.55, 1.
**
** MATERIALS
**
*Material, name=Material-1
*Density
1.8e-06,
*Depvar, delete=1
1,
*user defined UMAT
*Elastic, type=LAMINA
18330.,18330., 0.3, 7285., 4565., 4565.
*Fail Stress
255.06, 392.4, 255.06, 392.4, 147.15, 0., 0.
**
** INTERACTION PROPERTIES
**
*Surface Interaction, name=IntProp-1
*Friction
0.1,
**
** BOUNDARY CONDITIONS
**
** Name: BC-1 Type: Symmetry/Antisymmetry/Encastre
*Boundary
_PickedSet13, ENCASTRE
** Name: BC-2 Type: Displacement/Rotation
*Boundary
_PickedSet14, 1, 1
_PickedSet14, 3, 3
_PickedSet14, 4, 4
_PickedSet14, 5, 5
_PickedSet14, 6, 6
** ----------------------------------------------------------------
**
** STEP: Step-1
**
*Step, name=Step-1
*Dynamic, Explicit
, 0.55
*Bulk Viscosity
0.06, 1.2
**
** BOUNDARY CONDITIONS
**
** Name: BC-2 Type: Displacement/Rotation
*Boundary, amplitude=Amp-1
_PickedSet14, 1, 1
_PickedSet14, 2, 2, -18.
_PickedSet14, 3, 3
_PickedSet14, 4, 4
_PickedSet14, 5, 5
_PickedSet14, 6, 6
**
** INTERACTIONS
**
** Interaction: Int-1
*Contact Pair, interaction=IntProp-1, mechanical constraint=PENALTY, cpset=Int-1
_PickedSurf9, _PickedSurf10
** Interaction: Int-2
*Contact Pair, interaction=IntProp-1, mechanical constraint=PENALTY, cpset=Int-2
_PickedSurf11, _PickedSurf12
**
** OUTPUT REQUESTS
**
*Restart, write, number interval=1, time marks=NO
**
** FIELD OUTPUT: F-Output-1
**
*Output, field, variable=PRESELECT
**
** HISTORY OUTPUT: H-Output-1
**
*Output, history, variable=PRESELECT
*End Step


VUMAT subroutine


SUBROUTINE VUMAT(NBLOCK,NSTATEV,NDIR,NSHR,STRESS,CMNAME,STATE OLD,
1 STATENEW)

C
INCLUDE 'VABA_PARAM.INC'
C
C MATERIAL AND STRENGTH PARAMETERS
c
PARAMETER(ZERO=0.D0, ONE=1.D0, HALF=0.5D0)
PARAMETER(I_XX=1, I_YY=2, I_XY=4, N_SS=4)

C F1T,F2T TENSILE STRENGTH IN 1&2 DIRECTION
C F1C,F2C COMPRESSIVE STRENGTH IN 1&2 DIRECTION
c
PARAMETER(F1T=255.06, F2T=255.06,F1C=392.4,F2C=392.4, F12=147.15)
c
C
CHARACTER*80 CMNAME
DIMENSION STATENEW(NBLOCK,1),STRESS(NBLOCK,N_SS),
1 STATEOLD(NBLOCK,1)

STATEOLD(NBLOCK,1)=ONE
DO K=1,NBLOCK
S11=STRESS(K,I_XX)
S22=STRESS(K,I_YY)
S12=STRESS(K,I_XY)
C
C CHANG-CHANG FAILURE CRITERIA
C EVALUATE FIBRE MODE
C
IF(S11.GT.ZERO) THEN
C
C TENSILE FIBRE MODE
C
RFT=(S11/F1T)+(S12/F12)
IF(RFT.GT.ONE) THEN
STATENEW(K,1)=STATEOLD(K,1)*ZERO

ELSE
STATENEW(K,1)=STATEOLD(K,1)

ENDIF
ELSE
RFC=(S11/F1C)**2
IF(RFC.GT.ONE) THEN
STATENEW(K,1)=ZERO*STATEOLD(K,1)

ELSE
STATENEW(K,1)=STATEOLD(K,1)
ENDIF
ENDIF


C EVALUATE MATRIX FAILURE MODE
C TSAI-WU FAILURE CRITERIA
C
C
IF (S22.GT.ZERO) THEN
RMT=((S22)/(F2T*F2C))+(S12/F12)**2+(((F2C-F2T)*S22)/(F2C*F2T))
STATENEW(K,1)=ZERO*STATEOLD(K,1)
ELSE
STATENEW(K,1)=STATEOLD(K,1)
ENDIF
ENDDO
RETURN
END

[shripadtokekar]

Please sir reply to this question because i stopped here......if i get ans to this question i can go further...... so please help me.....:(

[Jorgen]

It appears that you VUMAT does not calculate a stress, but only attempts to predict failure. I don't see how that could work. I suggest that you also calculate the stress inside the VUMAT.

- Jorgen

[shripadtokekar]

hi sir thank you for your reply.....actually i want to predict failure and as soon as element satisfies the failure criteria i have to delete the element whether this is possible with above mentioned subroutine......and i am beginer in FEM field and hadling somewhat difficult project facing lot of problems. please helps me regarding this thank you very much

[shripadtokekar]

Hi sir, while using my subroutine it compiles fine but giving follwoing error
1)_VUMAT already defined in inputfilename.obj
2) fatal error- one or more multiply defined symbols found
these errors are coming while linking......... please help me thank you very much.

[Jorgen]

That is strange. Do you have two have two VUMAT() functions in your source file??

- Jorgen

[shripadtokekar]

thnanx sir your valuable suggesions....now the problem i talked before has been sort out but some new problem has come which i am trying to overcome.. so please help me regarding this problem.
During my quasi-static crusing analysis of composite tubes....in abaqus/explicit
I am using VUMAT for element deletion only... The subroutine compiles fine and now linking with input file but once it start analysis that is step it shows stable time increment quite low that is 2.31e-22......so that the analysis will not proceed further.......
so what should i do to overcome this difficulty. please help me regarding this
thank you very much...

regards
Shripad S. Tokekar
P.G. student
Indian Institute of Technlogy, Delhi. India

[Jorgen]

Check that you are using correct units in your simulations. Specifically check that the density is right.

-Jorgen

[shripadtokekar]

Hi Dr. Jorgen thank you for previous comments
sir, my VUMAT works fine for element deletion and i am doing lateral collapse of composite tubes actually what happens in my simulation that after small displacement load reached to its peak and which matches correctly with experimental one but after that load falls suddenly to small value or zero because of element deletion. but in experimental one it falls gradually so is this happening because of element deletion?? shall i do degradation in the stiffness insted of deleting the element?? which one is better or accurate to simulate the behavior??
One more question is that am actually doing experiments on woven composite material so to modelling woven composite in ABAQUS i am just adjusting or giving properties on macro scale and taking laminates of orientation 0 degrees is this feasible?? which failure criteria should i use for woven composite presently i am using Tsai-Wu criteria.
Sir please help me regarding this problem.
thank you very much.

[Jorgen]

It sounds like introducing degradation due to damage, before final failure, might be better for your simulation.

Tsai-Wu might be a good choice for your application. It is difficult to specify the most appropriate model without more info.

- Jorgen

[shripadtokekar]

thanx Dr. Jorgan. for ur reply........