Notifications
Clear all

Some Questions about VUMAT and VUHARD

Page 1 / 2

vstavro
Posts: 3
Topic starter
(@vstavro)
New Member
Joined: 13 years ago

I need to implement a constitutive model in Abaqus subroutine. It is used for high strain rate and large plastic deformation in which elastic strain could be omitted.

Now Im perplexed by some questions:

1: Is it ok to just define the material plasticity model in VUHARD and input E and v in material Elastic definition box? Or, do I need to write VUMAT to define the yield criterion and update the stresses explicitly?

///////////////////////////////////////////////////////////////////

2: If only VUHARD is enough, could the yield strength updated explicitly by the analytical model? Suppose we use a simplified model: FlowStress=K*(PlasticStrain^n)*(StrainRate^m), could we write the code as:

yield(k)=K*(eqps(k)**n))*(eqpsRate(k)**m)

or do we have to use Newton Iteration like:

IF(EQPLAS .GT. STATEV(1)) THEN

DO 10 KEWTON=1,NEWTON

c DF is the derivative of F with respect to yield strength

F=YieldOld-K*(eqps(k)**n))*(eqpsRate(k)**m)

DF=1-K*(eqps(k)**n))*(eqpsRate(k)**m)

DeltaYield=YieldOld-F/DF

IF(ABS(F).LT.TOLER) GOTO 20

10 CONTINUE

3: It seems that in VUHARD, the derivatives of the yield stress (or yield surface size in combined hardening models) must be defined with respect to the appropriate independent variables, such as strain, strain rate, and temperature.i.e. dyieldDeqps(nblock,1)/dyieldDeqps(nblock,2), etc. Othwise, the material looks to be very soft under shock pressure.

If use a simplied model: FlowStress=K*(PlasticStrain^n)*(StrainRate^m), could we define dyieldDeqps(nblock,1)=K*n*(PlasticStrain^(n-1))*(StrainRate^m) and dyieldDeqps(nblock,2)=K*m*(PlasticStrain^n)*(StrainRate^(m-1))?

Thank you very much for your help!

Stanley

Topic Tags
8 Replies
Jorgen
Posts: 3944
Moderator
(@jorgen)
Member
Joined: 2 years ago

1) Yes, VUHARD only contains the plasticity part.
2) My guess is that an explicit calculation is OK
3) My guess is that your expressions would work

Give it a try and let us know how it goes
-Jorgen

Topic Tags
8 Replies
vstavro
Posts: 3
Topic starter
(@vstavro)
New Member
Joined: 13 years ago

Thanks! I will summarize the simulation results and post them tomorrow.

So, it is not necessary to write VUMAT if we could only define plasticity model in VUHARD. Am I right?

Second, the material looks to be very soft (i.e. the deformation is abnormally drastic) if we only define yield(k), but without dyieldDeqps(k,2) (the derivative of yield strength with respect to strain and strain rate).

Reply
Jorgen
Posts: 3944
Moderator
(@jorgen)
Member
Joined: 2 years ago

1) Not necessary if you are only interested in a plasticity model
2) Without the derivative you will have no hardening, hence a soft response

-Jorgen

Reply
vstavro
Posts: 3
Topic starter
(@vstavro)
New Member
Joined: 13 years ago

I use VUHARD to implement 2 high strain rate plasticity models in Abaqus but find different results. One with much small deformation, the matl looks to be very stiff, the values of Yield(K) in the subroutine and Von Mise in Abaqus Field output are quite different. All these indicate that the subroutine fails to converge. The other models prediction is close to experiments, the values of Yield(K) in the subroutine and Von Mise in Abaqus Field output are identical.

It could be the reason that at the very beginning of simulation, when strain and strain rate close to zero, log(Strain rate) could be a big negative, inducing the non-convergence in the following steps. But I add lower limit to Strain rate,

StrainRate=0.d0

if (eqpsRate(k) .lt. 1E-8) then

StrainRate=1E-8

else

StrainRate=eqpsRate(k)

end if

still no use. (eqpsRate(k) is READ Only, not advised to change its value)

Could you please help?

.INP and simulation results attached.

Appreciate!

Reply
Jorgen
Posts: 3944
Moderator
(@jorgen)
Member
Joined: 2 years ago

I dont quite follow your description. Do you have a question?

-Jorgen

Reply
rune1001
Posts: 1
(@rune1001)
New Member
Joined: 8 years ago

Come on you got the wrong deritative. Maybe you should review some math, then you can

Come on you got the wrong deritatives. Maybe you should review some math, especiallly the chain rule, then you may got the correct answer.

By the way, why did you just calculate the derivative of the yield stress with respect to the equivalent plastic strain rate, but not the derivative of the yield stress with respect to the equivalent plastic strain ?

Thanks a ton in advance

Dong.Zhang

[QUOTE=stanley,4996]I use VUHARD to implement 2 high strain rate plasticity models in Abaqus but find different results. One with much small deformation, the matl looks to be very stiff, the values of Yield(K) in the subroutine and Von Mise in Abaqus Field output are quite different. All these indicate that the subroutine fails to converge. The other models prediction is close to experiments, the values of Yield(K) in the subroutine and Von Mise in Abaqus Field output are identical.

It could be the reason that at the very beginning of simulation, when strain and strain rate close to zero, log(Strain rate) could be a big negative, inducing the non-convergence in the following steps. But I add lower limit to Strain rate,

StrainRate=0.d0

if (eqpsRate(k) .lt. 1E-8) then

StrainRate=1E-8

else

StrainRate=eqpsRate(k)

end if

still no use. (eqpsRate(k) is READ Only, not advised to change its value)

Could you please help?

.INP and simulation results attached.

Appreciate![/QUOTE]

Reply
Page 1 / 2
Share: