예제 #1
0
RankFourTensor
TensorMechanicsPlasticMeanCapTC::consistentTangentOperator(const RankTwoTensor & trial_stress, Real intnl_old, const RankTwoTensor & stress, Real intnl,
                                                    const RankFourTensor & E_ijkl, const std::vector<Real> & cumulative_pm) const
{
  if (!_use_custom_cto)
    return TensorMechanicsPlasticModel::consistentTangentOperator(trial_stress, intnl_old, stress, intnl, E_ijkl, cumulative_pm);

  Real df_dq;
  Real alpha;
  if (trial_stress.trace() >= tensile_strength(intnl_old))
  {
    df_dq = -dtensile_strength(intnl);
    alpha = 1.0;
  }
  else
  {
    df_dq = dcompressive_strength(intnl);
    alpha = -1.0;
  }

  RankTwoTensor elas;
  for (unsigned int i = 0; i < 3; ++i)
    for (unsigned int j = 0; j < 3; ++j)
      for (unsigned int k = 0; k < 3; ++k)
        elas(i, j) += E_ijkl(i, j, k, k);

  const Real hw = -df_dq + alpha * elas.trace();

  return E_ijkl - alpha / hw * elas.outerProduct(elas);
}
RankTwoTensor
TensorMechanicsPlasticTensile::dyieldFunction_dstress(const RankTwoTensor & stress,
                                                      Real /*intnl*/) const
{
  Real mean_stress = stress.trace() / 3.0;
  RankTwoTensor dmean_stress = stress.dtrace() / 3.0;
  Real sin3Lode = stress.sin3Lode(_lode_cutoff, 0);
  if (sin3Lode <= _sin3tt)
  {
    // the non-edge-smoothed version
    std::vector<Real> eigvals;
    std::vector<RankTwoTensor> deigvals;
    stress.dsymmetricEigenvalues(eigvals, deigvals);
    Real denom = std::sqrt(smooth(stress) + Utility::pow<2>(eigvals[2] - mean_stress));
    return dmean_stress + (0.5 * dsmooth(stress) * dmean_stress +
                           (eigvals[2] - mean_stress) * (deigvals[2] - dmean_stress)) /
                              denom;
  }
  else
  {
    // the edge-smoothed version
    Real kk = _aaa + _bbb * sin3Lode + _ccc * Utility::pow<2>(sin3Lode);
    RankTwoTensor dkk = (_bbb + 2.0 * _ccc * sin3Lode) * stress.dsin3Lode(_lode_cutoff);
    Real sibar2 = stress.secondInvariant();
    RankTwoTensor dsibar2 = stress.dsecondInvariant();
    Real denom = std::sqrt(smooth(stress) + sibar2 * Utility::pow<2>(kk));
    return dmean_stress + (0.5 * dsmooth(stress) * dmean_stress +
                           0.5 * dsibar2 * Utility::pow<2>(kk) + sibar2 * kk * dkk) /
                              denom;
  }
}
Real
TensorMechanicsPlasticDruckerPrager::yieldFunction(const RankTwoTensor & stress, Real intnl) const
{
  Real aaa;
  Real bbb;
  bothAB(intnl, aaa, bbb);
  return std::sqrt(stress.secondInvariant()) + stress.trace() * bbb - aaa;
}
Real
TensorMechanicsPlasticDruckerPrager::dyieldFunction_dintnl(const RankTwoTensor & stress,
                                                           Real intnl) const
{
  Real daaa;
  Real dbbb;
  dbothAB(intnl, daaa, dbbb);
  return stress.trace() * dbbb - daaa;
}
예제 #5
0
void
TensorMechanicsPlasticMeanCapTC::activeConstraints(const std::vector<Real> & f, const RankTwoTensor & stress, Real intnl, const RankFourTensor & Eijkl, std::vector<bool> & act, RankTwoTensor & returned_stress) const
{
  act.assign(1, false);

  if (f[0] <= _f_tol)
  {
    returned_stress = stress;
    return;
  }

  const Real tr = stress.trace();
  const Real t_str = tensile_strength(intnl);
  Real str;
  Real dirn;
  if (tr >= t_str)
  {
    str = t_str;
    dirn = 1.0;
  }
  else
  {
    str = compressive_strength(intnl);
    dirn = -1.0;
  }

  RankTwoTensor n; // flow direction
  for (unsigned i = 0; i < 3; ++i)
    for (unsigned j = 0; j < 3; ++j)
      for (unsigned k = 0; k < 3; ++k)
             n(i, j) += dirn * Eijkl(i, j, k, k);

  // returned_stress = stress - gamma*n
  // and taking the trace of this and using
  // Tr(returned_stress) = str, gives
  // gamma = (Tr(stress) - str)/Tr(n)
  Real gamma = (stress.trace() - str) / n.trace();

  for (unsigned i = 0; i < 3; ++i)
    for (unsigned j = 0; j < 3; ++j)
      returned_stress(i, j) = stress(i, j) - gamma * n(i, j);

  act[0] = true;
}
Real
TensorMechanicsPlasticOrthotropic::yieldFunction(const RankTwoTensor & stress, Real intnl) const
{
  const RankTwoTensor j2prime = _l1 * stress;
  const RankTwoTensor j3prime = _l2 * stress;
  return _b * stress.trace() +
         std::pow(std::pow(-j2prime.generalSecondInvariant(), 3.0 / 2.0) - j3prime.det(),
                  1.0 / 3.0) -
         yieldStrength(intnl);
}
예제 #7
0
RankFourTensor
TensorMechanicsPlasticTensile::dflowPotential_dstress(const RankTwoTensor & stress, const Real & intnl) const
{
  Real mean_stress = stress.trace()/3.0;
  RankTwoTensor dmean_stress = stress.dtrace()/3.0;
  Real sin3Lode = stress.sin3Lode(_lode_cutoff, 0);
  if (sin3Lode <= _sin3tt)
  {
    // the non-edge-smoothed version
    std::vector<Real> eigvals;
    std::vector<RankTwoTensor> deigvals;
    std::vector<RankFourTensor> d2eigvals;
    stress.dsymmetricEigenvalues(eigvals, deigvals);
    stress.d2symmetricEigenvalues(d2eigvals);

    Real denom = std::sqrt(_small_smoother2 + std::pow(eigvals[2] - mean_stress, 2));

    RankFourTensor dr_dstress = (eigvals[2] - mean_stress)*d2eigvals[2]/denom;
    for (unsigned i = 0 ; i < 3 ; ++i)
      for (unsigned j = 0 ; j < 3 ; ++j)
        for (unsigned k = 0 ; k < 3 ; ++k)
          for (unsigned l = 0 ; l < 3 ; ++l)
            dr_dstress(i, j, k, l) += (1 - std::pow((eigvals[2] - mean_stress)/denom, 2))*(deigvals[2](i, j) - dmean_stress(i, j))*(deigvals[2](k, l) - dmean_stress(k, l))/denom;
    return dr_dstress;
  }
  else
  {
    // the edge-smoothed version
    RankTwoTensor dsin3Lode = stress.dsin3Lode(_lode_cutoff);
    Real kk = _aaa + _bbb*sin3Lode + _ccc*std::pow(sin3Lode, 2);
    RankTwoTensor dkk = (_bbb + 2*_ccc*sin3Lode)*dsin3Lode;
    RankFourTensor d2kk = (_bbb + 2*_ccc*sin3Lode)*stress.d2sin3Lode(_lode_cutoff);
    for (unsigned i = 0 ; i < 3 ; ++i)
      for (unsigned j = 0 ; j < 3 ; ++j)
        for (unsigned k = 0 ; k < 3 ; ++k)
          for (unsigned l = 0 ; l < 3 ; ++l)
            d2kk(i, j, k, l) += 2*_ccc*dsin3Lode(i, j)*dsin3Lode(k, l);

    Real sibar2 = stress.secondInvariant();
    RankTwoTensor dsibar2 = stress.dsecondInvariant();
    RankFourTensor d2sibar2 = stress.d2secondInvariant();

    Real denom = std::sqrt(_small_smoother2 + sibar2*std::pow(kk, 2));
    RankFourTensor dr_dstress = (0.5*d2sibar2*std::pow(kk, 2) + sibar2*kk*d2kk)/denom;
    for (unsigned i = 0 ; i < 3 ; ++i)
      for (unsigned j = 0 ; j < 3 ; ++j)
        for (unsigned k = 0 ; k < 3 ; ++k)
          for (unsigned l = 0 ; l < 3 ; ++l)
          {
            dr_dstress(i, j, k, l) += (dsibar2(i, j)*dkk(k, l)*kk + dkk(i, j)*dsibar2(k, l)*kk + sibar2*dkk(i, j)*dkk(k, l))/denom;
            dr_dstress(i, j, k, l) -= (0.5*dsibar2(i, j)*std::pow(kk, 2) + sibar2*kk*dkk(i, j))*(0.5*dsibar2(k, l)*std::pow(kk, 2) + sibar2*kk*dkk(k, l))/std::pow(denom, 3);
          }
    return dr_dstress;
  }
}
예제 #8
0
RankTwoTensor
TensorMechanicsPlasticMeanCapTC::df_dsig(const RankTwoTensor & stress, Real intnl) const
{
  const Real tr = stress.trace();
  const Real t_str = tensile_strength(intnl);
  if (tr >= t_str)
    return stress.dtrace();
  const Real c_str = compressive_strength(intnl);
  if (tr <= c_str)
    return -stress.dtrace();
  return - std::cos(M_PI * (tr - c_str) / (t_str - c_str)) * stress.dtrace();
}
예제 #9
0
RankFourTensor
TensorMechanicsPlasticMeanCapTC::dflowPotential_dstress(const RankTwoTensor & stress, Real intnl) const
{
  const Real tr = stress.trace();
  const Real t_str = tensile_strength(intnl);
  if (tr >= t_str)
    return RankFourTensor();
  const Real c_str = compressive_strength(intnl);
  if (tr <= c_str)
    return RankFourTensor();
  return M_PI / (t_str - c_str) * std::sin(M_PI * (tr - c_str) / (t_str - c_str)) * stress.dtrace().outerProduct(stress.dtrace());
}
예제 #10
0
Real
TensorMechanicsPlasticMeanCapTC::hardPotential(const RankTwoTensor & stress, Real intnl) const
{
  // This is the key for this whole class!
  const Real tr = stress.trace();
  const Real t_str = tensile_strength(intnl);
  if (tr >= t_str)
    return -1.0; // this will serve to *increase* the internal parameter (so internal parameter will be a measure of volumetric strain)
  const Real c_str = compressive_strength(intnl);
  if (tr <= c_str)
    return 1.0; // this will serve to *decrease* the internal parameter (so internal parameter will be a measure of volumetric strain)
  return std::cos(M_PI * (tr - c_str) / (t_str - c_str)); // this interpolates C2 smoothly between 1 and -1
}
Real
TensorMechanicsPlasticTensile::smooth(const RankTwoTensor & stress) const
{
  Real smoother2 = _small_smoother2;
  if (_tip_scheme == "cap")
  {
    Real x = stress.trace() / 3.0 - _cap_start;
    Real p = 0;
    if (x > 0)
      p = x * (1 - std::exp(-_cap_rate * x));
    smoother2 += Utility::pow<2>(p);
  }
  return smoother2;
}
예제 #12
0
Real
TensorMechanicsPlasticMeanCapTC::dhardPotential_dintnl(const RankTwoTensor & stress, Real intnl) const
{
  const Real tr = stress.trace();
  const Real t_str = tensile_strength(intnl);
  if (tr >= t_str)
    return 0.0;
  const Real c_str = compressive_strength(intnl);
  if (tr <= c_str)
    return 0.0;
  const Real dt = dtensile_strength(intnl);
  const Real dc = dcompressive_strength(intnl);
  return - std::sin(M_PI * (tr - c_str) / (t_str - c_str)) * M_PI / std::pow(t_str - c_str, 2) * ((tr - t_str) * dc - (tr - c_str) * dt);
}
예제 #13
0
Real
TensorMechanicsPlasticMeanCapTC::dyieldFunction_dintnl(const RankTwoTensor & stress, Real intnl) const
{
  const Real tr = stress.trace();
  const Real t_str = tensile_strength(intnl);
  if (tr >= t_str)
    return -dtensile_strength(intnl);
  const Real c_str = compressive_strength(intnl);
  if (tr <= c_str)
    return dcompressive_strength(intnl);
  const Real dt = dtensile_strength(intnl);
  const Real dc = dcompressive_strength(intnl);
  return (dc - dt) / M_PI * std::sin(M_PI * (tr - c_str) / (t_str - c_str)) + 1.0 / (t_str - c_str) * std::cos(M_PI * (tr - c_str) / (t_str - c_str)) * ((tr - c_str) * dt - (tr - t_str) * dc );
}
예제 #14
0
//Obtain deviatoric stress tensor
RankTwoTensor
FiniteStrainPlasticMaterial::getSigDev(RankTwoTensor sig)
{

//  Real sig_eqv,sij;
  RankTwoTensor identity;
  RankTwoTensor sig_dev;

  for (int i=0; i<3; i++)
    identity(i,i)=1.0;

  sig_dev=sig-identity*sig.trace()/3.0;
  return sig_dev;


}
예제 #15
0
Real
TensorMechanicsPlasticMeanCapTC::yieldFunction(const RankTwoTensor & stress, Real intnl) const
{
  const Real tr = stress.trace();
  const Real t_str = tensile_strength(intnl);
  if (tr >= t_str)
    return tr - t_str;
  const Real c_str = compressive_strength(intnl);
  if (tr <= c_str)
    return -(tr - c_str);
  // the following is zero at tr = t_str, and at tr = c_str
  // it also has derivative = 1 at tr = t_str, and derivative = -1 at tr = c_str
  // it also has second derivative = 0, at these points.
  // This makes the complete yield function C2 continuous.
  return (c_str - t_str) / M_PI * std::sin(M_PI * (tr - c_str) / (t_str - c_str));
}
Real
TensorMechanicsPlasticTensile::dsmooth(const RankTwoTensor & stress) const
{
  Real dsmoother2 = 0;
  if (_tip_scheme == "cap")
  {
    Real x = stress.trace() / 3.0 - _cap_start;
    Real p = 0;
    Real dp_dx = 0;
    if (x > 0)
    {
      p = x * (1 - std::exp(-_cap_rate * x));
      dp_dx = (1 - std::exp(-_cap_rate * x)) + x * _cap_rate * std::exp(-_cap_rate * x);
    }
    dsmoother2 += 2.0 * p * dp_dx;
  }
  return dsmoother2;
}
예제 #17
0
Real
TensorMechanicsPlasticTensile::yieldFunction(const RankTwoTensor & stress, const Real & intnl) const
{
  Real mean_stress = stress.trace()/3.0;
  Real sin3Lode = stress.sin3Lode(_lode_cutoff, 0);
  if (sin3Lode <= _sin3tt)
  {
    // the non-edge-smoothed version
    std::vector<Real> eigvals;
    stress.symmetricEigenvalues(eigvals);
    return mean_stress + std::sqrt(_small_smoother2 + std::pow(eigvals[2] - mean_stress, 2)) - tensile_strength(intnl);
  }
  else
  {
    // the edge-smoothed version
    Real kk = _aaa + _bbb*sin3Lode + _ccc*std::pow(sin3Lode, 2);
    Real sibar2 = stress.secondInvariant();
    return mean_stress + std::sqrt(_small_smoother2 + sibar2*std::pow(kk, 2)) - tensile_strength(intnl);
  }
}
예제 #18
0
Real
TensorMechanicsPlasticTensile::d2smooth(const RankTwoTensor & stress) const
{
  Real d2smoother2 = 0;
  if (_tip_scheme == "cap")
  {
    Real x = stress.trace()/3.0 - _cap_start;
    Real p = 0;
    Real dp_dx = 0;
    Real d2p_dx2 = 0;
    if (x > 0)
    {
      p = x*(1 - std::exp(-_cap_rate*x));
      dp_dx = (1 - std::exp(-_cap_rate*x)) + x*_cap_rate*std::exp(-_cap_rate*x);
      d2p_dx2 = 2*_cap_rate*std::exp(-_cap_rate*x) - x*std::pow(_cap_rate, 2)*std::exp(-_cap_rate*x);
    }
    d2smoother2 += 2*std::pow(dp_dx, 2) + 2*p*d2p_dx2;
  }
  return d2smoother2;
}
void
CappedDruckerPragerCosseratStressUpdate::setStressAfterReturn(const RankTwoTensor & stress_trial,
                                                              Real p_ok,
                                                              Real q_ok,
                                                              Real /*gaE*/,
                                                              const std::vector<Real> & /*intnl*/,
                                                              const yieldAndFlow & /*smoothed_q*/,
                                                              const RankFourTensor & /*Eijkl*/,
                                                              RankTwoTensor & stress) const
{
  // symm_stress is the symmetric part of the stress tensor.
  // symm_stress = (s_ij+s_ji)/2 + de_ij tr(stress) / 3
  //             = q / q_trial * (s_ij^trial+s_ji^trial)/2 + de_ij p / 3
  //             = q / q_trial * (symm_stress_ij^trial - de_ij tr(stress^trial) / 3) + de_ij p / 3
  const Real p_trial = stress_trial.trace();
  RankTwoTensor symm_stress = RankTwoTensor(RankTwoTensor::initIdentity) / 3.0 *
                              (p_ok - (_in_q_trial == 0.0 ? 0.0 : p_trial * q_ok / _in_q_trial));
  if (_in_q_trial > 0)
    symm_stress += q_ok / _in_q_trial * 0.5 * (stress_trial + stress_trial.transpose());
  stress = symm_stress + 0.5 * (stress_trial - stress_trial.transpose());
}
RankFourTensor
TensorMechanicsPlasticTensile::dflowPotential_dstress(const RankTwoTensor & stress,
                                                      Real /*intnl*/) const
{
  Real mean_stress = stress.trace() / 3.0;
  RankTwoTensor dmean_stress = stress.dtrace() / 3.0;
  Real sin3Lode = stress.sin3Lode(_lode_cutoff, 0);
  if (sin3Lode <= _sin3tt)
  {
    // the non-edge-smoothed version
    std::vector<Real> eigvals;
    std::vector<RankTwoTensor> deigvals;
    std::vector<RankFourTensor> d2eigvals;
    stress.dsymmetricEigenvalues(eigvals, deigvals);
    stress.d2symmetricEigenvalues(d2eigvals);

    Real denom = std::sqrt(smooth(stress) + Utility::pow<2>(eigvals[2] - mean_stress));
    Real denom3 = Utility::pow<3>(denom);
    RankTwoTensor numer_part = deigvals[2] - dmean_stress;
    RankTwoTensor numer_full =
        0.5 * dsmooth(stress) * dmean_stress + (eigvals[2] - mean_stress) * numer_part;
    Real d2smooth_over_denom = d2smooth(stress) / denom;

    RankFourTensor dr_dstress = (eigvals[2] - mean_stress) * d2eigvals[2] / denom;
    for (unsigned i = 0; i < 3; ++i)
      for (unsigned j = 0; j < 3; ++j)
        for (unsigned k = 0; k < 3; ++k)
          for (unsigned l = 0; l < 3; ++l)
          {
            dr_dstress(i, j, k, l) +=
                0.5 * d2smooth_over_denom * dmean_stress(i, j) * dmean_stress(k, l);
            dr_dstress(i, j, k, l) += numer_part(i, j) * numer_part(k, l) / denom;
            dr_dstress(i, j, k, l) -= numer_full(i, j) * numer_full(k, l) / denom3;
          }
    return dr_dstress;
  }
  else
  {
    // the edge-smoothed version
    RankTwoTensor dsin3Lode = stress.dsin3Lode(_lode_cutoff);
    Real kk = _aaa + _bbb * sin3Lode + _ccc * Utility::pow<2>(sin3Lode);
    RankTwoTensor dkk = (_bbb + 2.0 * _ccc * sin3Lode) * dsin3Lode;
    RankFourTensor d2kk = (_bbb + 2.0 * _ccc * sin3Lode) * stress.d2sin3Lode(_lode_cutoff);
    for (unsigned i = 0; i < 3; ++i)
      for (unsigned j = 0; j < 3; ++j)
        for (unsigned k = 0; k < 3; ++k)
          for (unsigned l = 0; l < 3; ++l)
            d2kk(i, j, k, l) += 2.0 * _ccc * dsin3Lode(i, j) * dsin3Lode(k, l);

    Real sibar2 = stress.secondInvariant();
    RankTwoTensor dsibar2 = stress.dsecondInvariant();
    RankFourTensor d2sibar2 = stress.d2secondInvariant();

    Real denom = std::sqrt(smooth(stress) + sibar2 * Utility::pow<2>(kk));
    Real denom3 = Utility::pow<3>(denom);
    Real d2smooth_over_denom = d2smooth(stress) / denom;
    RankTwoTensor numer_full =
        0.5 * dsmooth(stress) * dmean_stress + 0.5 * dsibar2 * kk * kk + sibar2 * kk * dkk;

    RankFourTensor dr_dstress = (0.5 * d2sibar2 * Utility::pow<2>(kk) + sibar2 * kk * d2kk) / denom;
    for (unsigned i = 0; i < 3; ++i)
      for (unsigned j = 0; j < 3; ++j)
        for (unsigned k = 0; k < 3; ++k)
          for (unsigned l = 0; l < 3; ++l)
          {
            dr_dstress(i, j, k, l) +=
                0.5 * d2smooth_over_denom * dmean_stress(i, j) * dmean_stress(k, l);
            dr_dstress(i, j, k, l) +=
                (dsibar2(i, j) * dkk(k, l) * kk + dkk(i, j) * dsibar2(k, l) * kk +
                 sibar2 * dkk(i, j) * dkk(k, l)) /
                denom;
            dr_dstress(i, j, k, l) -= numer_full(i, j) * numer_full(k, l) / denom3;
          }
    return dr_dstress;
  }
}
예제 #21
0
bool
TensorMechanicsPlasticJ2::returnMap(const RankTwoTensor & trial_stress,
                                    Real intnl_old,
                                    const RankFourTensor & E_ijkl,
                                    Real ep_plastic_tolerance,
                                    RankTwoTensor & returned_stress,
                                    Real & returned_intnl,
                                    std::vector<Real> & dpm,
                                    RankTwoTensor & delta_dp,
                                    std::vector<Real> & yf,
                                    bool & trial_stress_inadmissible) const
{
  if (!(_use_custom_returnMap))
    return TensorMechanicsPlasticModel::returnMap(trial_stress,
                                                  intnl_old,
                                                  E_ijkl,
                                                  ep_plastic_tolerance,
                                                  returned_stress,
                                                  returned_intnl,
                                                  dpm,
                                                  delta_dp,
                                                  yf,
                                                  trial_stress_inadmissible);

  yf.resize(1);

  Real yf_orig = yieldFunction(trial_stress, intnl_old);

  yf[0] = yf_orig;

  if (yf_orig < _f_tol)
  {
    // the trial_stress is admissible
    trial_stress_inadmissible = false;
    return true;
  }

  trial_stress_inadmissible = true;
  Real mu = E_ijkl(0, 1, 0, 1);

  // Perform a Newton-Raphson to find dpm when
  // residual = 3*mu*dpm - trial_equivalent_stress + yieldStrength(intnl_old + dpm) = 0
  Real trial_equivalent_stress = yf_orig + yieldStrength(intnl_old);
  Real residual;
  Real jac;
  dpm[0] = 0;
  unsigned int iter = 0;
  do
  {
    residual = 3.0 * mu * dpm[0] - trial_equivalent_stress + yieldStrength(intnl_old + dpm[0]);
    jac = 3.0 * mu + dyieldStrength(intnl_old + dpm[0]);
    dpm[0] += -residual / jac;
    if (iter > _max_iters) // not converging
      return false;
    iter++;
  } while (residual * residual > _f_tol * _f_tol);

  // set the returned values
  yf[0] = 0;
  returned_intnl = intnl_old + dpm[0];
  RankTwoTensor nn = 1.5 * trial_stress.deviatoric() /
                     trial_equivalent_stress; // = dyieldFunction_dstress(trial_stress, intnl_old) =
                                              // the normal to the yield surface, at the trial
                                              // stress
  returned_stress = 2.0 / 3.0 * nn * yieldStrength(returned_intnl);
  returned_stress.addIa(1.0 / 3.0 * trial_stress.trace());
  delta_dp = nn * dpm[0];

  return true;
}
void
CappedDruckerPragerCosseratStressUpdate::consistentTangentOperator(
    const RankTwoTensor & /*stress_trial*/,
    Real /*p_trial*/,
    Real /*q_trial*/,
    const RankTwoTensor & stress,
    Real /*p*/,
    Real q,
    Real gaE,
    const yieldAndFlow & smoothed_q,
    const RankFourTensor & Eijkl,
    bool compute_full_tangent_operator,
    RankFourTensor & cto) const
{
  if (!compute_full_tangent_operator)
  {
    cto = Eijkl;
    return;
  }

  RankFourTensor EAijkl;
  for (unsigned i = 0; i < _tensor_dimensionality; ++i)
    for (unsigned j = 0; j < _tensor_dimensionality; ++j)
      for (unsigned k = 0; k < _tensor_dimensionality; ++k)
        for (unsigned l = 0; l < _tensor_dimensionality; ++l)
        {
          cto(i, j, k, l) = 0.5 * (Eijkl(i, j, k, l) + Eijkl(j, i, k, l));
          EAijkl(i, j, k, l) = 0.5 * (Eijkl(i, j, k, l) - Eijkl(j, i, k, l));
        }

  const RankTwoTensor s_over_q =
      (q == 0.0 ? RankTwoTensor()
                : (0.5 * (stress + stress.transpose()) -
                   stress.trace() * RankTwoTensor(RankTwoTensor::initIdentity) / 3.0) /
                      q);
  const RankTwoTensor E_s_over_q = Eijkl.innerProductTranspose(s_over_q); // not symmetric in kl
  const RankTwoTensor Ekl =
      RankTwoTensor(RankTwoTensor::initIdentity).initialContraction(Eijkl); // symmetric in kl

  for (unsigned i = 0; i < _tensor_dimensionality; ++i)
    for (unsigned j = 0; j < _tensor_dimensionality; ++j)
      for (unsigned k = 0; k < _tensor_dimensionality; ++k)
        for (unsigned l = 0; l < _tensor_dimensionality; ++l)
        {
          cto(i, j, k, l) -= (i == j) * (1.0 / 3.0) *
                             (Ekl(k, l) * (1.0 - _dp_dpt) + 0.5 * E_s_over_q(k, l) * (-_dp_dqt));
          cto(i, j, k, l) -=
              s_over_q(i, j) * (Ekl(k, l) * (-_dq_dpt) + 0.5 * E_s_over_q(k, l) * (1.0 - _dq_dqt));
        }

  if (smoothed_q.dg[1] != 0.0)
  {
    const RankFourTensor Tijab = _Ehost * (gaE / _Epp) * smoothed_q.dg[1] * d2qdstress2(stress);
    RankFourTensor inv = RankFourTensor(RankFourTensor::initIdentitySymmetricFour) + Tijab;
    try
    {
      inv = inv.transposeMajor().invSymm();
    }
    catch (const MooseException & e)
    {
      // Cannot form the inverse, so probably at some degenerate place in stress space.
      // Just return with the "best estimate" of the cto.
      mooseWarning("CappedDruckerPragerCosseratStressUpdate: Cannot invert 1+T in consistent "
                   "tangent operator computation at quadpoint ",
                   _qp,
                   " of element ",
                   _current_elem->id());
      return;
    }
    cto = (cto.transposeMajor() * inv).transposeMajor();
  }
  cto += EAijkl;
}
예제 #23
0
bool
TensorMechanicsPlasticMeanCapTC::returnMap(const RankTwoTensor & trial_stress, Real intnl_old, const RankFourTensor & E_ijkl,
                                    Real ep_plastic_tolerance, RankTwoTensor & returned_stress, Real & returned_intnl,
                                    std::vector<Real> & dpm, RankTwoTensor & delta_dp, std::vector<Real> & yf,
                                    bool & trial_stress_inadmissible) const
{
  if (!(_use_custom_returnMap))
    return TensorMechanicsPlasticModel::returnMap(trial_stress, intnl_old, E_ijkl, ep_plastic_tolerance, returned_stress, returned_intnl, dpm, delta_dp, yf, trial_stress_inadmissible);

  yf.resize(1);

  Real yf_orig = yieldFunction(trial_stress, intnl_old);

  yf[0] = yf_orig;

  if (yf_orig < _f_tol)
  {
    // the trial_stress is admissible
    trial_stress_inadmissible = false;
    return true;
  }

  trial_stress_inadmissible = true;

  // In the following we want to solve
  // trial_stress - stress = E_ijkl * dpm * r   ...... (1)
  // and either
  // stress.trace() = tensile_strength(intnl)  ...... (2a)
  // intnl = intnl_old + dpm                   ...... (3a)
  // or
  // stress.trace() = compressive_strength(intnl) ... (2b)
  // intnl = intnl_old - dpm                   ...... (3b)
  // The former (2a and 3a) are chosen if
  // trial_stress.trace() > tensile_strength(intnl_old)
  // while the latter (2b and 3b) are chosen if
  // trial_stress.trace() < compressive_strength(intnl_old)
  // The variables we want to solve for are stress, dpm
  // and intnl.  We do this using a Newton approach, starting
  // with stress=trial_stress and intnl=intnl_old and dpm=0
  const bool tensile_failure = (trial_stress.trace() >= tensile_strength(intnl_old));
  const Real dirn = (tensile_failure ? 1.0 : -1.0);

  RankTwoTensor n; // flow direction, which is E_ijkl * r
  for (unsigned i = 0; i < 3; ++i)
    for (unsigned j = 0; j < 3; ++j)
      for (unsigned k = 0; k < 3; ++k)
             n(i, j) += dirn * E_ijkl(i, j, k, k);
  const Real n_trace = n.trace();

  // Perform a Newton-Raphson to find dpm when
  // residual = trial_stress.trace() - tensile_strength(intnl) - dpm * n.trace()  [for tensile_failure=true]
  // or
  // residual = trial_stress.trace() - compressive_strength(intnl) - dpm * n.trace()  [for tensile_failure=false]
  Real trial_trace = trial_stress.trace();
  Real residual;
  Real jac;
  dpm[0] = 0;
  unsigned int iter = 0;
  do {
    if (tensile_failure)
    {
      residual = trial_trace - tensile_strength(intnl_old + dpm[0]) - dpm[0] * n_trace;
      jac = -dtensile_strength(intnl_old + dpm[0]) - n_trace;
    }
    else
    {
      residual = trial_trace - compressive_strength(intnl_old - dpm[0]) - dpm[0] * n_trace;
      jac = -dcompressive_strength(intnl_old - dpm[0]) - n_trace;
    }
    dpm[0] += -residual/jac;
    if (iter > _max_iters) // not converging
      return false;
    iter++;
  } while (residual*residual > _f_tol*_f_tol);

  // set the returned values
  yf[0] = 0;
  returned_intnl = intnl_old + dirn * dpm[0];
  returned_stress = trial_stress - dpm[0] * n;
  delta_dp = dpm[0] * dirn * returned_stress.dtrace();

  return true;
}