[AMalaguera]

Hello everyone..
I am trying to simulate Terzagui's problem with elastoPlasticBiotFoam solver (Tian Tang code) using OpenFoam 1.6-ext. The following error appears:

-> FOAM FATAL ERROR:

request for volTensorField grad(U) from objectRegistry region0 failed
available objects of type volTensorField are

1
(
grad(dU)
)


From function objectRegistry::lookupObject<Type>(const word&) const
in file /home/angel/OpenFOAM/OpenFOAM-1.6-ext/src/OpenFOAM/lnInclude/objectRegistryTemplates.C at line 140.

FOAM aborting

The column is 3 m high and has an applied load of -120,000 Pa

The error is related to the boundary condition U.
I defined the U file as follows:


/*--------------------------------*- C++ -*----------------------------------*\
| ========= | |
| \\ / F ield | OpenFOAM Extend Project: Open Source CFD |
| \\ / O peration | Version: 1.6-ext |
| \\ / A nd | Web: www.extend-project.de |
| \\/ M anipulation | |
\*---------------------------------------------------------------------------*/
FoamFile
{
version 2.0;
format ascii;
class volVectorField;
location "0";
object dU;
}
// * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * //

dimensions [0 1 0 0 0 0 0];

internalField uniform (0 0 0);

boundaryField
{
right
{

type zeroTotalTraction;

}

left
{

type zeroTotalTraction;

}

front
{
type empty;

}

soilStructureInterface
{

type fixedTotalTraction;

traction uniform (0 0 -120000);
pressure uniform 0;
value uniform (0 0 0);

}

back
{
type empty;

}

soilDomainBottom
{

type fixedDisplacementZeroShear;
value uniform (0 0 0);

}


}

// ************************************************** *********************** //

I will be grateful for any suggestion..

Thanks... Angel

[HPE]

Hi,

- Is there any functionality in your setup that requests `grad(U)`? Like a function object may be? The error message clearly suggests that this request cannot be fulfilled because the only available object registered to the objectRegistry is `grad(dU)`.

[AMalaguera]

Originally Posted by HPE

Hi,

- Is there any functionality in your setup that requests `grad(U)`? Like a function object may be? The error message clearly suggests that this request cannot be fulfilled because the only available object registered to the objectRegistry is `grad(dU)`.

Hello Mr. Herpes .. Thanks for your reply

grad (U) is defined as a field. That error appears when I use the fixedDisplacementZeroShear boundary condition. That boundary condition is from the soilEpTLFoam solver of the same Tian Tang code. I would like to know if someone from the community has used the Tian Tang code

https://bitbucket.org/tiantang/minigeotechfoam.git

and they have solved the terzagi problem. My problem is that the pressure field coincides with the analytical results but the displacement field gives exactly double.

Any hold would appreciate

Thanks angel

[AMalaguera]

The problem of the Terzagui column in the elastic case was solved without fail.

I even solved it for a two-dimensional and three-dimensional geometry assigning the condition of "slip" on the side walls of the column. And the results coincide with the literature.

My problem now is to assume that the column has elastic-plastic behavior. I understand that in this case the only thing to do to Tian's code is to tell the code in the controlDict file to do plastic correction (correctPlasticity yes

When the simulation starts, the following error appears:

--> FOAM Warning :
From function eigenValues(const symmTensor&)
in file calculateCellEigens.H at line 188
return zero eigen directions and strange things detected for symmTensor: (-81911.6 1.53846e-08 1.50815e-08 -81911.6 1.51826e-08 -191127)
--> FOAM Warning :
From function eigenValues(const symmTensor&)
in file calculateCellEigens.H at line 188
return zero eigen directions and strange things detected for symmTensor: (-81911.6 1.53846e-08 1.50815e-08 -81911.6 1.51826e-08 -191127)


and reviewing the numerical results, do not see results in the plastic deformation fields.

It is a column 3 m high x 1 x 1 m in base. It has a constant load of 350,000 Pa applied to the top. In the compressed file attached the case

[AMalaguera]

If anyone can give me a sujestion, I would appreciate

Thanks

Att...Angel

[bigphil]

Hi Angel,

Can you attach the log here? I am wondering whether the "calculateCellEigens.H" error is just a symptom of an earlier error or if it really is part of the cause.

If I had to guess, I suspect that the solver is diverging which ends in this error. A solution might be to increase the amount of under-relaxation i.e. reduce under-relaxation factors.

Philip

[AMalaguera]

Originally Posted by bigphil

Hi Angel,

Can you attach the log here? I am wondering whether the "calculateCellEigens.H" error is just a symptom of an earlier error or if it really is part of the cause.

If I had to guess, I suspect that the solver is diverging which ends in this error. A solution might be to increase the amount of under-relaxation i.e. reduce under-relaxation factors.

Philip

Dr. Philip Cardiff
Thanks for your quick response

The "calculateCellEigens.H" error is just a symptom of an earlier error.

I already reduced under-relaxation factors, in both fields, pressure and displacement. I also decreased tolerance in both fields. I also thought that the solver is diverging. The error keeps popping up

Another experiment I did was to simulate the problem with a 2D and 3D geometry.

In this case, assign the boundary condition of "slip" on the side walls for displacement and "zeroGradient" for pressure. These boundary conditions guarantee that the problem remains one-dimensional, because Terzagui is indeed one-dimensional. And the same error keeps popping up

As I mentioned earlier, the simulation continues until the end of the allotted time. I compared the results of pressure and displacement with the results of the elastic problem and they are identical.

Therefore I presume that he does not make the plastic correction and solves the problem as if it were elastic.

Let me clarify something.

What I am trying to solve is the Terzagui column considering elasto-plastic deformation. This means that only deformations should appear along the column. This analysis is not in Tian's work. She solved the "stripFooting" problem to validate the elasto-plastic code. I also solved "stripFooting" and the results match Tian's.

In other words, the error appears when I perform a one-dimensional elasto-plastic simulation.

As the "stripFooting" problem is two-dimensional with a linearly varying load, so I also ran a simulation with a linearly varying load and the same error appears

I don't know if I'm erring with the boundary conditions. What seems strange to me is that in the problem of "stripFooting" the results are good and coincide with the literature

I appreciate any suggestions.

Thank you.

[bigphil]

Hi Angel,

Is there a physical solution to the 1-D elasto-plastic Terzagi column problem?
For example, as it is 1-D then it means that all plastic deformation must be related to the volume change; this would not make sense for incompressible plasticity but I guess it is OK for Mohr-Coulomb; either way, it would be good to have a reference for what you expect the solution to be.

Philip