How to set constant mean pressure (hydrostatic pressure)

[sendy]

Hallo guys,

I want to simulate triaxial compression test with constant mean pressure. The initial condition is isotropic stress. Then the displacement control is applied on top of the specimen, with controlling the value of mean pressure (T11+T22+T33) to be always constant along the test.

The problem is how to set the constant mean pressure in abaqus ? I try to use subroutine DLOAD, but the information about current stress is not provided in this subroutine.

Any suggestion will be appreciated. Thank you very much for the time.

regards,
Sendy.

Note: I need to do this test in stress and displacement control boundary condition to catch the softening behavior.

 

[mrgoldthorpe]

Sendy,
I can think of an inelegant way. Excuse the Fortran 77:

Use one user defined variable so you call can subroutine SUBROUTINE UVARM(UVAR,DIRECT,T,TIME,DTIME,CMNAME,ORNAME,
NUVARM,NOEL,NPT,LAYER,KSPT,KSTEP,KINC,NDI,NSHR,COORD,JMAC,
JMATYP,MATLAYO,LACCFLA)

In your UVARM do the following:

C set up a common block array dimensioned 'no. of
C integration stations, number of elements' i.e.
COMMON /KMYBLOCK/PARRAY(4,10000)

C then get the current stress invariants at this
C integration station:
CALL GETVRM 'SINV',ARRAY,JARRAY,FLGRAY,JRCD,
JMAC,JMATYP,MATLAYO,LACCFLA)
PRESS = ARRAY (3)
C now store PRESS for the current NOEL and NPT in your own
C named common block array dimensioned 'no. of integration
C stations, number of elements':
PARRAY(NPT, NOEL)=PRESS

C it might be a good idea to say set UVAR(1)=PRESS so the
C user variable is set.

In SUBROUTINE DLOAD(F,KSTEP,KINC,TIME,NOEL,NPT,LAYER,KSPT,
1 COORDS,JLTYP,SNAME)
C access PRESS for the current NOEL and NPT from the common C block array.
C I think this will provide the value at the end of the last load increment:
PRESS=PARRAY(NPT, NOEL)
C now use it to provide your F value

Perhaps someone else can think of something more elegant?

MRG

 

[MCC1966]

Sendy


I do not think you need to use a special subroutine for that. Accuratly speaking are you talking about: total or effective mean stress ..? Are you simulating CU test or CD or UU test..? one has to know this information to give you an accurate answer.

 

[sendy]

Hallo MCC1966,

I am trying to model triaxial drained compression test under constant mean pressure. The constitutive model that I use is hypoplastic model, so basically I use UMAT in abaqus. At the moment I apply the load control (instead of displacement control) with Riks method (arc length) to capture the behavior after the peak (softening). But I found out that this method somehow take some time, so I try to use displacement control.
When I use displacement control, I need to decrease my lateral pressure while vertical displacement is prescribing. Here is the point, I don't know how to control this lateral pressure in such a way that the mean pressure is constant.
I also tried to prescribed the lateral displacement in stead of lateral pressure, but I found out that the lateral displacement is also a function of current stress.
(Note: This function would not be found if elastic isotropic model is used, so that the constant mean pressure might be controlled by prescribing the displacement)

Please kindly give me a hinge and suggestion.

sendy.

 

[MCC1966]

Sendy
P' =(sigma 1'+ sigma 2'+ sigma 3')/3
When you are applying compression load or displacment difinitely in drained test P' will increase through sigm1'
1-Realistically there no shear under drained sitution where P' becomes constant. If you know please provide me with example. is not enc possible.
2-Of course under load control one cannot track softening so you must work with displacement
3- If you work with lrage deformation stress control is not equivalent to load control
4- Is your goal to simulate shearing under constant mean stress or the softening behavior exhibited by the soil.
5- If you work with material instability softening then that is I beleive is not possible to capture in ABAQUS
I am sorry If i could not answer your question

 

[sendy]

Hallo MCC1966,

I am sorry that I wrote it wrongly, I meant the stress control test instead of load control.

In the triaxial compression that I want to simulate, the lateral stress (or lateral displacement) should be controled in such a way that the trace sigma is kept constant (S11+S22+S33=constant) while applying compression pressure (or compression displacement). I succesfully did it with stress control and I want to know how to do it under displacement control.

Then, I'd like to response your points:
1- I agree that there would be no shear when mean pressure is kept constant, which I also got from this triaxial simulation under stress control.
2- I would say that in the triaxial simulation under stress control the softening might be catched by using certain methods, i.e. modified arc length method which is already applied in Abaqus through RIKS option.
3- yes, I agree.
4- No, at the moment my goal is doing simulation of triaxial compression for broken rock under constant mean pressure to catch the effect of creep/relaxation.
5- I would say that the material stability, i.e. softening might not be catched by some available models in Abaqus (I have not used all of them). However, user defined material (UMAT) could be used to implement the constitutive model which is able to catch softening.

regards,
Sendy.

 

[sendy]

Hallo mrgoldthorpe,

Thank you for the post. It seems to be possible using your idea; it seems to be quite complicated for me.

regards,
Sendy.