Rate dependent TSL

[amassaf]

Hi, I have this new class to implement a rate dependent TSL but it's returning a zero normal and tangent traction :

void
RateDependentTractionSeparation::computeInterfaceTractionAndDerivatives()
{
  // Calculate rate of displacement jump
  RealVectorValue rateOfDisplacementJump = (_interface_displacement_jump[_qp] - _interface_displacement_jump_old[_qp]) / _dt;

  // Calculate degradation factor
  RealVectorValue degradationFactor = calculateDegradationFactor();

  // Compute total traction: elastic part + viscous part
  _interface_traction[_qp] = degradationFactor * (_K * _interface_displacement_jump[_qp] + _viscosity * rateOfDisplacementJump);

  // Update derivative of traction with respect to displacement jump
  RankTwoTensor dFactor;
  dFactor.fillFromInputVector({degradationFactor(0), degradationFactor(1), degradationFactor(2)});
  _dinterface_traction_djump[_qp] = dFactor * _K;
}

RealVectorValue
RateDependentTractionSeparation::calculateDegradationFactor()
{
  RealVectorValue critical_displacement(_normal_critical_displacement, _tangent_critical_displacement, _tangent_critical_displacement);

  RealVectorValue factor(1.0, 1.0, 1.0);  
  
  for (unsigned int i = 0; i < 3; ++i) {
    if (_interface_displacement_jump[_qp](i) >= critical_displacement(i)) {
      factor(i) = 0.00002;  
    } else {
      Real delta = _interface_displacement_jump[_qp](i) / critical_displacement(i);
      factor(i) = 1.0 - delta * (1.0 - 0.00002);  
    }
  }

  return factor;
}

if anyone could take a look,, thanks !!

[GiudGiud]

@sidharthsarmah you were working with TSL in your project could you please help?

[sidharthsarmah]

Hi @amassaf , I have just started using CZM in MOOSE, could you please send me the TSLs and the derivatives you implemented to have a look at.

[amassaf]

Hi @sidharthsarmah thanks for your help. I have changed some stuff to my code and this is what my computeInterfaceTractionAndDerivatives looks like now :

  for (int i = 0; i < 3; ++i)
  {
    if (delta(i) < critical_displacement(i)) 
    {
      _d[_qp](i) = 0;
      softening_ratio(i) = 1 - _d[_qp](i);
      traction(i) = K(i) * softening_ratio(i) * delta(i) * (1 + viscosity(i) * rateOfDisplacementJump(i)) + K(i) * delta_inactive(i);
      dTrac(i) = K(i) * softening_ratio(i) * ddelta_active_ddelta(i) * (1 + viscosity(i) * rateOfDisplacementJump(i)) + K(i) * ddelta_inactive_ddelta(i);

      ddTrac.fillFromInputVector({dTrac(0), dTrac(1), dTrac(2)});
    }
    else if (delta(i) < _failure_disp) 
    {
      _d[_qp](i) = _failure_disp * (delta(i) - critical_displacement(i)) / delta(i) /
              (_failure_disp - critical_displacement(i));

      _d[_qp](i) = (_d[_qp](i) + 1e-5 * _d_old[_qp](i) / _dt) / (1e-5 / _dt + 1);

RealVectorValue dd_ddelta(0, 0, 0);
if (_d[_qp](i) < _d_old[_qp](i))
         _d[_qp](i) = _d_old[_qp](i);
 dd_ddelta(i) = ( _failure_disp * delta(i) - _failure_disp * (delta(i) - critical_displacement(i)))  
                    / (delta(i) * delta(i) * (_failure_disp - critical_displacement(i)));

      softening_ratio(i) = 1 - _d[_qp](i);
      traction(i) = K(i) * softening_ratio(i) * delta(i) * (1 + viscosity(i) * rateOfDisplacementJump(i)) + K(i) * delta_inactive(i);

      ddTrac.fillFromInputVector({dTrac(0), dTrac(1), dTrac(2)});
    }
    else 
    {
      _d[_qp](i) = 1;
      softening_ratio(i) = 1 - _d[_qp](i);
      traction(i) = K(i) * softening_ratio(i) * delta(i) * (1 + viscosity(i) * rateOfDisplacementJump(i)) + K(i) * delta_inactive(i);
      dTrac(i) = K(i) * softening_ratio(i) * ddelta_active_ddelta(i) * (1 + viscosity(i) * rateOfDisplacementJump(i)) + K(i) * ddelta_inactive_ddelta(i);

      ddTrac.fillFromInputVector({dTrac(0), dTrac(1), dTrac(2)});
    }
  }

  _interface_traction[_qp] = traction;

  _dinterface_traction_djump[_qp] = ddTrac;

My problem is that the traction stays increasing even after delta(i) > _failure_disp, but I do not have this problem with the normal traction.

[sidharthsarmah]

Hi,
I am new to this and have just implemented the TSL in MOOSE. I need to understand the TSL which you used first and check if there is any syntax error in the code.

[amassaf]

Hi, sorry to get back late to this but I figured it out. I was not taking into account the magnitude of the tangential disp so it was all wrong.