Assigning different material properties to gauss point

[manikumd22]

Hi @GiudGiud

I am working on a project that requires varying the material properties for different gauss points.

For instance 2D plate will have Gauss points, and I want to assign different materials to each Gauss point in order to increase x

For instance,

[Materials]
[compute_stress1]
type = ComputeNeoHookeanStress
lambda = 40.0
mu = 67.0
x = 0.1, all y
[]
[compute_stress2]
type = ComputeNeoHookeanStress
lambda = 400.0
mu = 670.0
x =0.2, all y
[]
[compute_stress3]
type = ComputeNeoHookeanStress
lambda = 4000.0
mu = 6700.0
x = 0.3, all y
[]
[compute_strain]
type = ComputeLagrangianStrain
[]
[]

and so on

I am struggling with that as I don't know the location of gauss points coordinate

Or

  **if (_large_kinematics)
  {
      if ( _q_point[_qp](1) = 0.1;) {
        RankTwoTensor Cinv = (2 * _E[_qp] + RankTwoTensor::Identity()).inverse();
        _S[_qp] = 100* (_lambda[_qp] * log(_F[_qp].det()) - _mu[_qp]) * Cinv +
              _mu[_qp] * RankTwoTensor::Identity();
        _C[_qp] =100*(
            -2 * (_lambda[_qp] * log(_F[_qp].det()) - _mu[_qp]) * Cinv.times<i_, k_, l_, j_>(Cinv) +
            _lambda[_qp] * Cinv.times<i_, j_, k_, l_>(Cinv));
    }  
      if ( _q_point[_qp](1) = 0.2;) {
        RankTwoTensor Cinv = (2 * _E[_qp] + RankTwoTensor::Identity()).inverse();
        _S[_qp] = 1000* (_lambda[_qp] * log(_F[_qp].det()) - _mu[_qp]) * Cinv +
              _mu[_qp] * RankTwoTensor::Identity();
        _C[_qp] =1000*(
            -2 * (_lambda[_qp] * log(_F[_qp].det()) - _mu[_qp]) * Cinv.times<i_, k_, l_, j_>(Cinv) +
            _lambda[_qp] * Cinv.times<i_, j_, k_, l_>(Cinv));
    }  
      if ( _q_point[_qp](1) = 0.3;){
        RankTwoTensor Cinv = (2 * _E[_qp] + RankTwoTensor::Identity()).inverse();
        _S[_qp] = 10000* (_lambda[_qp] * log(_F[_qp].det()) - _mu[_qp]) * Cinv +
              _mu[_qp] * RankTwoTensor::Identity();
        _C[_qp] =10000*(
            -2 * (_lambda[_qp] * log(_F[_qp].det()) - _mu[_qp]) * Cinv.times<i_, k_, l_, j_>(Cinv) +
            _lambda[_qp] * Cinv.times<i_, j_, k_, l_>(Cinv));
    }  
  }**

[GiudGiud]

Hello

We dont have very objects to output the quadrature points.
I think the cleanest might be to create a QuadraturePointsPositions object and get the JSON output from there.
I ll get started on this right now

[GiudGiud]

#28565 has a new Positions object that will output the quadrature points to JSON

[manikumd22]

Hi @GiudGiud
I tried as per that discussion, but I am getting an error

*** ERROR ***
A 'QuadraturePointsPositions' is not a registered object.

If you are trying to find this object in a dynamically loaded library, make sure that
the library can be found either in your "Problem/library_path" parameter or in the
MOOSE_LIBRARY_PATH environment variable.�[39m

#Tensor Mechanics tutorial: the basics
#Step 1, part 1
#2D simulation of uniaxial tension with linear elasticity

[GlobalParams]
displacements = 'disp_x disp_y'
[]

[Mesh]
[msh]
type = GeneratedMeshGenerator
dim = 2
nx = 4
ny = 4
[]
[]

[Positions]
[all_mesh]
type = QuadraturePointsPositions
auto_sort = true
[]
[]

[Modules/TensorMechanics/Master]
[all]
add_variables = true
strain = small
generate_output = 'vonmises_stress stress_xx stress_yy strain_xx strain_yy'
[]
[]

[Materials]

[elasticity1]
type = ComputeIsotropicElasticityTensor
youngs_modulus = 9e4
poissons_ratio = 0.3
[]

[./stress]
type = ComputeLinearElasticStress
[../]
[]

[BCs]
[./left]
type = DirichletBC
variable = disp_x
boundary = 'top bottom'
value = 0.0
[../]
[./bottom]
type = DirichletBC
variable = disp_y
boundary = 'top'
value = 1
[../]
[./top]
type = DirichletBC
variable = disp_y
boundary = 'bottom'
value = 0
[../]
[]

[Preconditioning]
[./SMP]
type = SMP
full = true
[../]
[]

[Postprocessors]
[num_pos]
type = NumPositions
positions = all_mesh
[]
[]

[Executioner]
type = Steady

solve_type = 'NEWTON'

petsc_options = '-snes_ksp_ew'
petsc_options_iname = '-pc_type -sub_pc_type -pc_asm_overlap -ksp_gmres_restart'
petsc_options_value = 'asm lu 1 101'
[]

[Outputs]
type = JSON
exodus = true
perf_graph = true
file_base = ./inputfiles/maniklinearelastic_ir_out12
[]

[GiudGiud]

did you pull that branch and built it?
You can either do that or add the source and header to your local installation then rebuild

[GiudGiud]

for clarification, this object is not in the moose source code yet, I just made it this afternoon. So you have to either use my branch or add it to your code

[manikumd22]

Hi @GiudGiud

I add it to my local moose app,

[GiudGiud]

umm.
then I guess:

• save all your work, include this new object, with git commit
• go in the top folder of your app
• git clean -Xfd
• recompile with make -j 6
• try again

[manikumd22]

Hi @GiudGiud

I tried but getting the same error

[GiudGiud]

you can try to re-register the object by changing the registerMooseObject call in QuadraturePointsPositions.C

when you do, check that the object is being properly recompiled

[manikumd22]

Hi @GiudGiud ,

Now it worked,

thank you so much

For example, I'm getting in this form

                {
                    "x": 0.052831216351296784,
                    "y": 0.052831216351296784,
                    "z": 0.0
                },
                {
                    "x": 0.052831216351296784,
                    "y": 0.19716878364870322,
                    "z": 0.0
                },
                {
                    "x": 0.19716878364870322,
                    "y": 0.9471687836487033,
                    "z": 0.0
                },
                {
                    "x": 0.30283121635129684,
                    "y": 0.052831216351296784,
                    "z": 0.0
                },

Will this below material module work, if I define it in this way and will _q_point_qp = 0.052831216351296784 take all the 9 gauss points corresponding to this or I have to define both _q_point_qp = 0.052831216351296784 and _q_point_qp = 0.052831216351296784

if (_large_kinematics)
{
if ( _q_point_qp = 0.052831216351296784;) {
RankTwoTensor Cinv = (2 * _E[_qp] + RankTwoTensor::Identity()).inverse();
_S[_qp] = 100 (_lambda[_qp] * log(_F[_qp].det()) - _mu[_qp]) * Cinv +
_mu[_qp] * RankTwoTensor::Identity();
_C[_qp] =100(
-2 * (lambda[qp] * log(F[qp].det()) - mu[qp]) * Cinv.times<i, k, l, j>(Cinv) +
lambda[qp] * Cinv.times<i, j, k, l>(Cinv));
}
if ( _q_point_qp = 0.19716878364870322;) {
RankTwoTensor Cinv = (2 * _E[_qp] + RankTwoTensor::Identity()).inverse();
_S[_qp] = 1000* (_lambda[_qp] * log(_F[_qp].det()) - _mu[_qp]) * Cinv +
_mu[_qp] * RankTwoTensor::Identity();
_C[_qp] =1000*(
-2 * (lambda[qp] * log(F[qp].det()) - mu[qp]) * Cinv.times<i, k, l, j>(Cinv) +
lambda[qp] * Cinv.times<i, j, k, l>(Cinv));
}
if ( _q_point_qp = 0.30283121635129684;){
RankTwoTensor Cinv = (2 * _E[_qp] + RankTwoTensor::Identity()).inverse();
_S[_qp] = 10000* (_lambda[_qp] * log(_F[_qp].det()) - _mu[_qp]) * Cinv +
_mu[_qp] * RankTwoTensor::Identity();
_C[_qp] =10000*(
-2 * (lambda[qp] * log(F[qp].det()) - mu[qp]) * Cinv.times<i, k, l, j>(Cinv) +
lambda[qp] * Cinv.times<i, j, k, l>(Cinv));
}
}**

[GiudGiud]

no this wont work

if ( _q_point[_qp](https://github.com/idaholab/moose/discussions/1) = 0.052831216351296784;) {

this sets (single = instead of two ==) the value of the q point
AND
the quadrature point will never be exactly equal to 0.052831216351296784, you need a tolerance.

So instead

if (MooseUtils::absoluteFuzzyEqual(_q_point[_qp](1), 0.052831216351296784))

this used a 1e-12 tolerance iirc, enough here