libMesh
Functions
systems_of_equations_ex2.C File Reference

Go to the source code of this file.

Functions

void assemble_stokes (EquationSystems &es, const std::string &system_name)
 
void set_lid_driven_bcs (TransientLinearImplicitSystem &system)
 
void set_stagnation_bcs (TransientLinearImplicitSystem &system)
 
void set_poiseuille_bcs (TransientLinearImplicitSystem &system)
 
int main (int argc, char **argv)
 
void assemble_stokes (EquationSystems &es, const std::string &libmesh_dbg_var(system_name))
 

Function Documentation

◆ assemble_stokes() [1/2]

void assemble_stokes ( EquationSystems es,
const std::string &  system_name 
)

Referenced by main().

◆ assemble_stokes() [2/2]

void assemble_stokes ( EquationSystems es,
const std::string &  libmesh_dbg_varsystem_name 
)

Definition at line 345 of file systems_of_equations_ex2.C.

References libMesh::FEGenericBase< OutputType >::build(), libMesh::FEType::default_quadrature_order(), dim, libMesh::DofMap::dof_indices(), libMesh::Parameters::get(), libMesh::EquationSystems::get_mesh(), libMesh::EquationSystems::get_system(), libMesh::DofMap::heterogenously_constrain_element_matrix_and_vector(), libMesh::libmesh_ignore(), mesh, libMesh::MeshBase::mesh_dimension(), libMesh::TransientSystem< Base >::old_solution(), libMesh::EquationSystems::parameters, libMesh::Real, libMesh::DenseSubVector< T >::reposition(), libMesh::DenseSubMatrix< T >::reposition(), libMesh::DenseVector< T >::resize(), and libMesh::DenseMatrix< T >::resize().

347 {
348  // It is a good idea to make sure we are assembling
349  // the proper system.
350  libmesh_assert_equal_to (system_name, "Navier-Stokes");
351 
352 #if LIBMESH_DIM > 1
353  // Get a constant reference to the mesh object.
354  const MeshBase & mesh = es.get_mesh();
355 
356  // The dimension that we are running
357  const unsigned int dim = mesh.mesh_dimension();
358 
359  // Get a reference to the Stokes system object.
360  TransientLinearImplicitSystem & navier_stokes_system =
361  es.get_system<TransientLinearImplicitSystem> ("Navier-Stokes");
362 
363  // Numeric ids corresponding to each variable in the system
364  const unsigned int u_var = navier_stokes_system.variable_number ("vel_x");
365  const unsigned int v_var = navier_stokes_system.variable_number ("vel_y");
366  const unsigned int p_var = navier_stokes_system.variable_number ("p");
367 
368  // Get the Finite Element type for "u". Note this will be
369  // the same as the type for "v".
370  FEType fe_vel_type = navier_stokes_system.variable_type(u_var);
371 
372  // Get the Finite Element type for "p".
373  FEType fe_pres_type = navier_stokes_system.variable_type(p_var);
374 
375  // Build a Finite Element object of the specified type for
376  // the velocity variables.
377  std::unique_ptr<FEBase> fe_vel (FEBase::build(dim, fe_vel_type));
378 
379  // Build a Finite Element object of the specified type for
380  // the pressure variables.
381  std::unique_ptr<FEBase> fe_pres (FEBase::build(dim, fe_pres_type));
382 
383  // A Gauss quadrature rule for numerical integration.
384  // Let the FEType object decide what order rule is appropriate.
385  QGauss qrule (dim, fe_vel_type.default_quadrature_order());
386 
387  // Tell the finite element objects to use our quadrature rule.
388  fe_vel->attach_quadrature_rule (&qrule);
389  fe_pres->attach_quadrature_rule (&qrule);
390 
391  // Here we define some references to cell-specific data that
392  // will be used to assemble the linear system.
393  //
394  // The element Jacobian * quadrature weight at each integration point.
395  const std::vector<Real> & JxW = fe_vel->get_JxW();
396 
397  // The element shape functions evaluated at the quadrature points.
398  const std::vector<std::vector<Real>> & phi = fe_vel->get_phi();
399 
400  // The element shape function gradients for the velocity
401  // variables evaluated at the quadrature points.
402  const std::vector<std::vector<RealGradient>> & dphi = fe_vel->get_dphi();
403 
404  // The element shape functions for the pressure variable
405  // evaluated at the quadrature points.
406  const std::vector<std::vector<Real>> & psi = fe_pres->get_phi();
407 
408  // The value of the linear shape function gradients at the quadrature points
409  // const std::vector<std::vector<RealGradient>> & dpsi = fe_pres->get_dphi();
410 
411  // A reference to the DofMap object for this system. The DofMap
412  // object handles the index translation from node and element numbers
413  // to degree of freedom numbers. We will talk more about the DofMap
414  // in future examples.
415  const DofMap & dof_map = navier_stokes_system.get_dof_map();
416 
417  // Define data structures to contain the element matrix
418  // and right-hand-side vector contribution. Following
419  // basic finite element terminology we will denote these
420  // "Ke" and "Fe".
423 
425  Kuu(Ke), Kuv(Ke), Kup(Ke),
426  Kvu(Ke), Kvv(Ke), Kvp(Ke),
427  Kpu(Ke), Kpv(Ke), Kpp(Ke);
428 
430  Fu(Fe),
431  Fv(Fe),
432  Fp(Fe);
433 
434  // References to momentum equation right hand sides
435  std::reference_wrapper<DenseSubVector<Number>> F[2] = {Fu, Fv};
436 
437  // References to velocity-velocity coupling blocks
438  std::reference_wrapper<DenseSubMatrix<Number>> K[2][2] = {{Kuu, Kuv}, {Kvu, Kvv}};
439 
440  // References to velocity-pressure coupling blocks
441  std::reference_wrapper<DenseSubMatrix<Number>> B[2] = {Kup, Kvp};
442  std::reference_wrapper<DenseSubMatrix<Number>> BT[2] = {Kpu, Kpv};
443 
444  // This vector will hold the degree of freedom indices for
445  // the element. These define where in the global system
446  // the element degrees of freedom get mapped.
447  std::vector<dof_id_type> dof_indices;
448  std::vector<dof_id_type> dof_indices_u;
449  std::vector<dof_id_type> dof_indices_v;
450  std::vector<dof_id_type> dof_indices_p;
451 
452  // Find out what the timestep size parameter is from the system, and
453  // the value of theta for the theta method. We use implicit Euler (theta=1)
454  // for this simulation even though it is only first-order accurate in time.
455  // The reason for this decision is that the second-order Crank-Nicolson
456  // method is notoriously oscillatory for problems with discontinuous
457  // initial data such as the lid-driven cavity. Therefore,
458  // we sacrifice accuracy in time for stability, but since the solution
459  // reaches steady state relatively quickly we can afford to take small
460  // timesteps. If you monitor the initial nonlinear residual for this
461  // simulation, you should see that it is monotonically decreasing in time.
462  const Real dt = es.parameters.get<Real>("dt");
463  const Real theta = 1.;
464 
465  // The kinematic viscosity, multiplies the "viscous" terms.
466  const Real nu = es.parameters.get<Real>("nu");
467 
468  // The system knows whether or not we need to do a pressure pin.
469  // This is only required for problems with all-Dirichlet boundary
470  // conditions on the velocity.
471  const bool pin_pressure = es.parameters.get<bool>("pin_pressure");
472 
473  // The global system matrix
474  SparseMatrix<Number> & matrix = navier_stokes_system.get_system_matrix();
475 
476  // Now we will loop over all the elements in the mesh that
477  // live on the local processor. We will compute the element
478  // matrix and right-hand-side contribution. Since the mesh
479  // will be refined we want to only consider the ACTIVE elements,
480  // hence we use a variant of the active_elem_iterator.
481  for (const auto & elem : mesh.active_local_element_ptr_range())
482  {
483  // Get the degree of freedom indices for the
484  // current element. These define where in the global
485  // matrix and right-hand-side this element will
486  // contribute to.
487  dof_map.dof_indices (elem, dof_indices);
488  dof_map.dof_indices (elem, dof_indices_u, u_var);
489  dof_map.dof_indices (elem, dof_indices_v, v_var);
490  dof_map.dof_indices (elem, dof_indices_p, p_var);
491 
492  const unsigned int n_dofs = dof_indices.size();
493  const unsigned int n_u_dofs = dof_indices_u.size();
494  const unsigned int n_v_dofs = dof_indices_v.size();
495  const unsigned int n_p_dofs = dof_indices_p.size();
496 
497  // Compute the element-specific data for the current
498  // element. This involves computing the location of the
499  // quadrature points (q_point) and the shape functions
500  // (phi, dphi) for the current element.
501  fe_vel->reinit (elem);
502  fe_pres->reinit (elem);
503 
504  // Zero the element matrix and right-hand side before
505  // summing them. We use the resize member here because
506  // the number of degrees of freedom might have changed from
507  // the last element. Note that this will be the case if the
508  // element type is different (i.e. the last element was a
509  // triangle, now we are on a quadrilateral).
510  Ke.resize (n_dofs, n_dofs);
511  Fe.resize (n_dofs);
512 
513  // Reposition the submatrices... The idea is this:
514  //
515  // - - - -
516  // | Kuu Kuv Kup | | Fu |
517  // Ke = | Kvu Kvv Kvp |; Fe = | Fv |
518  // | Kpu Kpv Kpp | | Fp |
519  // - - - -
520  //
521  // The DenseSubMatrix.reposition () member takes the
522  // (row_offset, column_offset, row_size, column_size).
523  //
524  // Similarly, the DenseSubVector.reposition () member
525  // takes the (row_offset, row_size)
526  Kuu.reposition (u_var*n_u_dofs, u_var*n_u_dofs, n_u_dofs, n_u_dofs);
527  Kuv.reposition (u_var*n_u_dofs, v_var*n_u_dofs, n_u_dofs, n_v_dofs);
528  Kup.reposition (u_var*n_u_dofs, p_var*n_u_dofs, n_u_dofs, n_p_dofs);
529 
530  Kvu.reposition (v_var*n_v_dofs, u_var*n_v_dofs, n_v_dofs, n_u_dofs);
531  Kvv.reposition (v_var*n_v_dofs, v_var*n_v_dofs, n_v_dofs, n_v_dofs);
532  Kvp.reposition (v_var*n_v_dofs, p_var*n_v_dofs, n_v_dofs, n_p_dofs);
533 
534  Kpu.reposition (p_var*n_u_dofs, u_var*n_u_dofs, n_p_dofs, n_u_dofs);
535  Kpv.reposition (p_var*n_u_dofs, v_var*n_u_dofs, n_p_dofs, n_v_dofs);
536  Kpp.reposition (p_var*n_u_dofs, p_var*n_u_dofs, n_p_dofs, n_p_dofs);
537 
538  Fu.reposition (u_var*n_u_dofs, n_u_dofs);
539  Fv.reposition (v_var*n_u_dofs, n_v_dofs);
540  Fp.reposition (p_var*n_u_dofs, n_p_dofs);
541 
542  // Now we will build the element matrix and right-hand-side.
543  // Constructing the RHS requires the solution and its
544  // gradient from the previous timestep. This must be
545  // calculated at each quadrature point by summing the
546  // solution degree-of-freedom values by the appropriate
547  // weight functions.
548  for (unsigned int qp=0; qp<qrule.n_points(); qp++)
549  {
550  // Values to hold the solution & its gradient at the previous timestep.
551  NumberVectorValue U_old;
553  Number p_old = 0.;
554 
555  // {grad_u, grad_v}, initialized to zero
556  std::array<Gradient, 2> grad_uv{};
557 
558  // {grad_u_old, grad_v_old}, initialized to zero
559  std::array<Gradient, 2> grad_uv_old{};
560 
561  // Compute the velocity & its gradient from the previous timestep
562  // and the old Newton iterate.
563  for (unsigned int l=0; l<n_u_dofs; l++)
564  {
565  // From the old timestep:
566  U_old(0) += phi[l][qp]*navier_stokes_system.old_solution (dof_indices_u[l]);
567  U_old(1) += phi[l][qp]*navier_stokes_system.old_solution (dof_indices_v[l]);
568  grad_uv_old[0].add_scaled (dphi[l][qp],navier_stokes_system.old_solution (dof_indices_u[l]));
569  grad_uv_old[1].add_scaled (dphi[l][qp],navier_stokes_system.old_solution (dof_indices_v[l]));
570 
571  // From the previous Newton iterate:
572  U(0) += phi[l][qp]*navier_stokes_system.current_solution (dof_indices_u[l]);
573  U(1) += phi[l][qp]*navier_stokes_system.current_solution (dof_indices_v[l]);
574  grad_uv[0].add_scaled (dphi[l][qp],navier_stokes_system.current_solution (dof_indices_u[l]));
575  grad_uv[1].add_scaled (dphi[l][qp],navier_stokes_system.current_solution (dof_indices_v[l]));
576  }
577 
578  // Compute the old pressure value at this quadrature point.
579  for (unsigned int l=0; l<n_p_dofs; l++)
580  p_old += psi[l][qp]*navier_stokes_system.old_solution (dof_indices_p[l]);
581 
582  // First, an i-loop over the velocity degrees of freedom.
583  // We know that n_u_dofs == n_v_dofs so we can compute contributions
584  // for both at the same time.
585  for (unsigned int i=0; i<n_u_dofs; i++)
586  {
587  for (unsigned int k=0; k<2; ++k)
588  F[k](i) += JxW[qp] *
589  (U_old(k) * phi[i][qp] - // mass-matrix term
590  (1.-theta) * dt * (U_old * grad_uv_old[k]) * phi[i][qp] + // convection term
591  (1.-theta) * dt * p_old * dphi[i][qp](k) - // pressure term on rhs
592  (1.-theta) * dt * nu * (grad_uv_old[k] * dphi[i][qp]) + // diffusion term on rhs
593  theta * dt * (U * grad_uv[k]) * phi[i][qp]); // Newton term
594 
595  // Matrix contributions for the uu and vv couplings.
596  for (unsigned int j=0; j<n_u_dofs; j++)
597  for (unsigned int k=0; k<2; ++k)
598  for (unsigned int l=0; l<2; ++l)
599  {
600  // "Diagonal" contribution
601  if (k==l)
602  K[k][k](i,j) += JxW[qp]*(phi[i][qp]*phi[j][qp] + // mass matrix term
603  theta*dt*nu*(dphi[i][qp]*dphi[j][qp]) + // diffusion term
604  theta*dt*(U*dphi[j][qp])*phi[i][qp]); // convection term
605 
606  // Newton term
607  K[k][l](i,j) += JxW[qp] * theta * dt * grad_uv[k](l) * phi[i][qp] * phi[j][qp];
608  }
609 
610  // Matrix contributions for the up and vp couplings.
611  for (unsigned int j=0; j<n_p_dofs; j++)
612  for (unsigned int k=0; k<2; ++k)
613  B[k](i,j) += JxW[qp] * -theta * dt * psi[j][qp] * dphi[i][qp](k);
614  }
615 
616  // Now an i-loop over the pressure degrees of freedom. This code computes
617  // the matrix entries due to the continuity equation. Note: To maintain a
618  // symmetric matrix, we may (or may not) multiply the continuity equation by
619  // negative one. Here we do not.
620  for (unsigned int i=0; i<n_p_dofs; i++)
621  for (unsigned int j=0; j<n_u_dofs; j++)
622  for (unsigned int k=0; k<2; ++k)
623  BT[k](i,j) += JxW[qp] * psi[i][qp] * dphi[j][qp](k);
624  } // end of the quadrature point qp-loop
625 
626 
627  // At this point the interior element integration has been
628  // completed. We now need to pin the pressure to zero at global
629  // node number "pressure_node". This effectively removes the
630  // non-trivial null space of constant pressure solutions. The
631  // pressure pin is not necessary in problems that have "outflow"
632  // BCs, like Poiseuille flow with natural BCs. In fact it is
633  // actually wrong to do so, since the pressure is not
634  // under-specified in that situation.
635  if (pin_pressure)
636  {
637  const Real penalty = 1.e10;
638  const unsigned int pressure_node = 0;
639  const Real p_value = 0.0;
640  for (auto c : elem->node_index_range())
641  if (elem->node_id(c) == pressure_node)
642  {
643  Kpp(c,c) += penalty;
644  Fp(c) += penalty*p_value;
645  }
646  }
647 
648  // Since we're using heterogeneous DirichletBoundary objects for
649  // the boundary conditions, we need to call a specific function
650  // to constrain the element stiffness matrix.
651  dof_map.heterogenously_constrain_element_matrix_and_vector (Ke, Fe, dof_indices);
652 
653  // The element matrix and right-hand-side are now built
654  // for this element. Add them to the global matrix and
655  // right-hand-side vector. The SparseMatrix::add_matrix()
656  // and NumericVector::add_vector() members do this for us.
657  matrix.add_matrix (Ke, dof_indices);
658  navier_stokes_system.rhs->add_vector (Fe, dof_indices);
659  } // end of element loop
660 #else
661  libmesh_ignore(es);
662 #endif
663 }
class FEType hides (possibly multiple) FEFamily and approximation orders, thereby enabling specialize...
Definition: fe_type.h:182
void dof_indices(const Elem *const elem, std::vector< dof_id_type > &di) const
Fills the vector di with the global degree of freedom indices for the element.
Definition: dof_map.C:1992
unsigned int dim
void resize(const unsigned int n)
Resize the vector.
Definition: dense_vector.h:374
MeshBase & mesh
Manages storage and variables for transient systems.
Order default_quadrature_order() const
Definition: fe_type.h:357
This class defines a vector in LIBMESH_DIM dimensional Real or Complex space.
const T_sys & get_system(std::string_view name) const
This is the MeshBase class.
Definition: mesh_base.h:74
Defines a dense subvector for use in finite element computations.
This class handles the numbering of degrees of freedom on a mesh.
Definition: dof_map.h:169
Definition: assembly.h:38
void libmesh_ignore(const Args &...)
const T & get(std::string_view) const
Definition: parameters.h:426
Defines a dense submatrix for use in Finite Element-type computations.
Number old_solution(const dof_id_type global_dof_number) const
void heterogenously_constrain_element_matrix_and_vector(DenseMatrix< Number > &matrix, DenseVector< Number > &rhs, std::vector< dof_id_type > &elem_dofs, bool asymmetric_constraint_rows=true, int qoi_index=-1) const
Definition: dof_map.h:1237
DIE A HORRIBLE DEATH HERE typedef LIBMESH_DEFAULT_SCALAR_TYPE Real
const MeshBase & get_mesh() const
void resize(const unsigned int new_m, const unsigned int new_n)
Resizes the matrix to the specified size and calls zero().
Definition: dense_matrix.h:895
This class implements specific orders of Gauss quadrature.
unsigned int mesh_dimension() const
Definition: mesh_base.C:324
Parameters parameters
Data structure holding arbitrary parameters.

◆ main()

int main ( int  argc,
char **  argv 
)

Definition at line 94 of file systems_of_equations_ex2.C.

References libMesh::EquationSystems::add_system(), assemble_stokes(), libMesh::MeshTools::Generation::build_square(), libMesh::command_line_next(), libMesh::default_solver_package(), libMesh::FIRST, libMesh::Parameters::get(), libMesh::EquationSystems::get_system(), libMesh::TriangleWrapper::init(), libMesh::EquationSystems::init(), libMesh::INVALID_SOLVER_PACKAGE, mesh, libMesh::MeshTools::n_elem(), libMesh::TransientSystem< Base >::old_local_solution, libMesh::out, libMesh::EquationSystems::parameters, libMesh::PerfLog::pop(), libMesh::EquationSystems::print_info(), libMesh::MeshBase::print_info(), libMesh::PerfLog::push(), libMesh::QUAD9, libMesh::Real, libMesh::SECOND, libMesh::Parameters::set(), set_lid_driven_bcs(), and libMesh::TRILINOS_SOLVERS.

95 {
96  // Initialize libMesh.
97  LibMeshInit init (argc, argv);
98 
99  // This example requires a linear solver package.
100  libmesh_example_requires(libMesh::default_solver_package() != INVALID_SOLVER_PACKAGE,
101  "--enable-petsc, --enable-trilinos, or --enable-eigen");
102 
103  // Skip this 2D example if libMesh was compiled as 1D-only.
104  libmesh_example_requires(2 <= LIBMESH_DIM, "2D support");
105 
106  // We use Dirichlet boundary conditions here
107 #ifndef LIBMESH_ENABLE_DIRICHLET
108  libmesh_example_requires(false, "--enable-dirichlet");
109 #endif
110 
111  // This example NaNs with the Trilinos solvers
112  libmesh_example_requires(libMesh::default_solver_package() != TRILINOS_SOLVERS, "--enable-petsc or --enable-laspack");
113 
114  // Create a mesh, with dimension to be overridden later, distributed
115  // across the default MPI communicator.
116  Mesh mesh(init.comm());
117 
118  // Get the mesh size from the command line.
119  const int n_elem = libMesh::command_line_next("-n_elem", 20);
120 
121  // Use the MeshTools::Generation mesh generator to create a uniform
122  // 2D grid on the square [-1,1]^2. We instruct the mesh generator
123  // to build a mesh of 8x8 Quad9 elements in 2D. Building these
124  // higher-order elements allows us to use higher-order polynomial
125  // approximations for the velocity.
127  n_elem, n_elem,
128  0., 1.,
129  0., 1.,
130  QUAD9);
131 
132  // Print information about the mesh to the screen.
133  mesh.print_info();
134 
135  // Create an equation systems object.
136  EquationSystems equation_systems (mesh);
137 
138  // Declare the system and its variables.
139  // Creates a transient system named "Navier-Stokes"
141  equation_systems.add_system<TransientLinearImplicitSystem> ("Navier-Stokes");
142 
143  // Add the variables "vel_x" & "vel_y" to "Navier-Stokes". They
144  // will be approximated using second-order approximation.
145  system.add_variable ("vel_x", SECOND);
146  system.add_variable ("vel_y", SECOND);
147 
148  // Add the variable "p" to "Navier-Stokes". This will
149  // be approximated with a first-order basis,
150  // providing an LBB-stable pressure-velocity pair.
151  system.add_variable ("p", FIRST);
152 
153  // Give the system a pointer to the matrix assembly
154  // function.
155  system.attach_assemble_function (assemble_stokes);
156 
157  // Note: only pick one set of BCs!
158  set_lid_driven_bcs(system);
159  // set_stagnation_bcs(system);
160  // set_poiseuille_bcs(system);
161 
162  // Initialize the data structures for the equation system.
163  equation_systems.init ();
164 
165  // Prints information about the system to the screen.
166  equation_systems.print_info();
167 
168  // Create a performance-logging object for this example
169  PerfLog perf_log("Systems Example 2");
170 
171  // Get a reference to the Stokes system to use later.
172  TransientLinearImplicitSystem & navier_stokes_system =
173  equation_systems.get_system<TransientLinearImplicitSystem>("Navier-Stokes");
174 
175  // Now we begin the timestep loop to compute the time-accurate
176  // solution of the equations.
177  const Real dt = 0.1;
178  navier_stokes_system.time = 0.0;
179  const unsigned int n_timesteps = 15;
180 
181  // The number of steps and the stopping criterion are also required
182  // for the nonlinear iterations.
183  const unsigned int n_nonlinear_steps = 15;
184  const Real nonlinear_tolerance = 1.e-5;
185 
186  // We also set a standard linear solver flag in the EquationSystems object
187  // which controls the maximum number of linear solver iterations allowed.
188  const int max_iter = libMesh::command_line_next("-max_iter", 25);
189 
190  equation_systems.parameters.set<unsigned int>("linear solver maximum iterations") = max_iter;
191 
192  // Tell the system of equations what the timestep is by using
193  // the set_parameter function. The matrix assembly routine can
194  // then reference this parameter.
195  equation_systems.parameters.set<Real> ("dt") = dt;
196 
197  // The kinematic viscosity, nu = mu/rho, units of length**2/time.
198  equation_systems.parameters.set<Real> ("nu") = .007;
199 
200  // The first thing to do is to get a copy of the solution at
201  // the current nonlinear iteration. This value will be used to
202  // determine if we can exit the nonlinear loop.
203  std::unique_ptr<NumericVector<Number>>
204  last_nonlinear_soln (navier_stokes_system.solution->clone());
205 
206  // Since we are not doing adaptivity, write all solutions to a single Exodus file.
207  ExodusII_IO exo_io(mesh);
208 
209  for (unsigned int t_step=1; t_step<=n_timesteps; ++t_step)
210  {
211  // Increment the time counter, set the time step size as
212  // a parameter in the EquationSystem.
213  navier_stokes_system.time += dt;
214 
215  // A pretty update message
216  libMesh::out << "\n\n*** Solving time step "
217  << t_step
218  << ", time = "
219  << navier_stokes_system.time
220  << " ***"
221  << std::endl;
222 
223  // Now we need to update the solution vector from the
224  // previous time step. This is done directly through
225  // the reference to the Stokes system.
226  *navier_stokes_system.old_local_solution = *navier_stokes_system.current_local_solution;
227 
228  // At the beginning of each solve, reset the linear solver tolerance
229  // to a "reasonable" starting value.
230  const Real initial_linear_solver_tol = 1.e-6;
231  equation_systems.parameters.set<Real> ("linear solver tolerance") = initial_linear_solver_tol;
232 
233  // We'll set this flag when convergence is (hopefully) achieved.
234  bool converged = false;
235 
236  // Now we begin the nonlinear loop
237  for (unsigned int l=0; l<n_nonlinear_steps; ++l)
238  {
239  // Update the nonlinear solution.
240  last_nonlinear_soln->zero();
241  last_nonlinear_soln->add(*navier_stokes_system.solution);
242 
243  // Assemble & solve the linear system.
244  perf_log.push("linear solve");
245  equation_systems.get_system("Navier-Stokes").solve();
246  perf_log.pop("linear solve");
247 
248  // Compute the difference between this solution and the last
249  // nonlinear iterate.
250  last_nonlinear_soln->add (-1., *navier_stokes_system.solution);
251 
252  // Close the vector before computing its norm
253  last_nonlinear_soln->close();
254 
255  // Compute the l2 norm of the difference
256  const Real norm_delta = last_nonlinear_soln->l2_norm();
257 
258  // How many iterations were required to solve the linear system?
259  const unsigned int n_linear_iterations = navier_stokes_system.n_linear_iterations();
260 
261  // What was the final residual of the linear system?
262  const Real final_linear_residual = navier_stokes_system.final_linear_residual();
263 
264  // If the solver did no work (sometimes -ksp_converged_reason
265  // says "Linear solve converged due to CONVERGED_RTOL
266  // iterations 0") but the nonlinear residual norm is above
267  // the tolerance, we need to pick an even lower linear
268  // solver tolerance and try again. Note that the tolerance
269  // is relative to the norm of the RHS, which for this
270  // particular problem does not go to zero, since we are
271  // solving for the full solution rather than the update.
272  //
273  // Similarly, if the solver did no work and this is the 0th
274  // nonlinear step, it means that the delta between solutions
275  // is being inaccurately measured as "0" since the solution
276  // did not change. Decrease the tolerance and try again.
277  if (n_linear_iterations == 0 &&
278  (navier_stokes_system.final_linear_residual() >= nonlinear_tolerance || l==0))
279  {
280  Real old_linear_solver_tolerance = equation_systems.parameters.get<Real> ("linear solver tolerance");
281  equation_systems.parameters.set<Real> ("linear solver tolerance") = 1.e-3 * old_linear_solver_tolerance;
282  continue;
283  }
284 
285  // Print out convergence information for the linear and
286  // nonlinear iterations.
287  libMesh::out << "Linear solver converged at step: "
288  << n_linear_iterations
289  << ", final residual: "
290  << final_linear_residual
291  << " Nonlinear convergence: ||u - u_old|| = "
292  << norm_delta
293  << std::endl;
294 
295  // Terminate the solution iteration if the difference between
296  // this nonlinear iterate and the last is sufficiently small, AND
297  // if the most recent linear system was solved to a sufficient tolerance.
298  if ((norm_delta < nonlinear_tolerance) &&
299  (navier_stokes_system.final_linear_residual() < nonlinear_tolerance))
300  {
301  libMesh::out << " Nonlinear solver converged at step "
302  << l
303  << std::endl;
304  converged = true;
305  break;
306  }
307 
308  // Otherwise, decrease the linear system tolerance. For the inexact Newton
309  // method, the linear solver tolerance needs to decrease as we get closer to
310  // the solution to ensure quadratic convergence. The new linear solver tolerance
311  // is chosen (heuristically) as the square of the previous linear system residual norm.
312  //Real flr2 = final_linear_residual*final_linear_residual;
313  Real new_linear_solver_tolerance = std::min(Utility::pow<2>(final_linear_residual), initial_linear_solver_tol);
314  equation_systems.parameters.set<Real> ("linear solver tolerance") = new_linear_solver_tolerance;
315  } // end nonlinear loop
316 
317  // Don't keep going if we failed to converge.
318  libmesh_error_msg_if(!converged, "Error: Newton iterations failed to converge!");
319 
320 #ifdef LIBMESH_HAVE_EXODUS_API
321  // Write out every nth timestep to file.
322  const unsigned int write_interval = 1;
323 
324  if ((t_step+1)%write_interval == 0)
325  {
326  exo_io.write_timestep("out.e",
327  equation_systems,
328  t_step+1, // we're off by one since we wrote the IC and the Exodus numbering is 1-based.
329  navier_stokes_system.time);
330  }
331 #endif // #ifdef LIBMESH_HAVE_EXODUS_API
332  } // end timestep loop.
333 
334  // All done.
335  return 0;
336 }
T command_line_next(std::string name, T default_value)
Use GetPot&#39;s search()/next() functions to get following arguments from the command line...
Definition: libmesh.C:1011
This is the EquationSystems class.
void set_lid_driven_bcs(TransientLinearImplicitSystem &system)
dof_id_type n_elem(const MeshBase::const_element_iterator &begin, const MeshBase::const_element_iterator &end)
Count up the number of elements of a specific type (as defined by an iterator range).
Definition: mesh_tools.C:850
The ExodusII_IO class implements reading meshes in the ExodusII file format from Sandia National Labs...
Definition: exodusII_io.h:52
MeshBase & mesh
Manages storage and variables for transient systems.
void build_square(UnstructuredMesh &mesh, const unsigned int nx, const unsigned int ny, const Real xmin=0., const Real xmax=1., const Real ymin=0., const Real ymax=1., const ElemType type=INVALID_ELEM, const bool gauss_lobatto_grid=false)
A specialized build_cube() for 2D meshes.
The LibMeshInit class, when constructed, initializes the dependent libraries (e.g.
Definition: libmesh.h:90
NumericVector< Number > * old_local_solution
All the values I need to compute my contribution to the simulation at hand.
SolverPackage default_solver_package()
Definition: libmesh.C:1050
The PerfLog class allows monitoring of specific events.
Definition: perf_log.h:145
void assemble_stokes(EquationSystems &es, const std::string &system_name)
void print_info(std::ostream &os=libMesh::out, const unsigned int verbosity=0, const bool global=true) const
Prints relevant information about the mesh.
Definition: mesh_base.C:1489
void init(triangulateio &t)
Initializes the fields of t to nullptr/0 as necessary.
DIE A HORRIBLE DEATH HERE typedef LIBMESH_DEFAULT_SCALAR_TYPE Real
OStreamProxy out
The Mesh class is a thin wrapper, around the ReplicatedMesh class by default.
Definition: mesh.h:50

◆ set_lid_driven_bcs()

void set_lid_driven_bcs ( TransientLinearImplicitSystem system)

Definition at line 667 of file systems_of_equations_ex2.C.

References libMesh::DofMap::add_dirichlet_boundary().

Referenced by main().

668 {
669  // This problem *does* require a pressure pin, there are Dirichlet
670  // boundary conditions for u and v on the entire boundary.
671  system.get_equation_systems().parameters.set<bool>("pin_pressure") = true;
672 
673 #ifdef LIBMESH_ENABLE_DIRICHLET
674  unsigned short int
675  u_var = system.variable_number("vel_x"),
676  v_var = system.variable_number("vel_y");
677 
678  // Get a convenient reference to the System's DofMap
679  DofMap & dof_map = system.get_dof_map();
680 
681  // u=v=0 on bottom, left, right
683  {u_var, v_var},
685  // u=1 on top
686  dof_map.add_dirichlet_boundary(DirichletBoundary({2}, {u_var},
687  ConstFunction<Number>(1.)));
688  // v=0 on top
689  dof_map.add_dirichlet_boundary(DirichletBoundary({2}, {v_var},
691 #endif // LIBMESH_ENABLE_DIRICHLET
692 }
ConstFunction that simply returns 0.
Definition: zero_function.h:38
This class allows one to associate Dirichlet boundary values with a given set of mesh boundary ids an...
This class handles the numbering of degrees of freedom on a mesh.
Definition: dof_map.h:169
Function that returns a single value that never changes.
void add_dirichlet_boundary(const DirichletBoundary &dirichlet_boundary)
Adds a copy of the specified Dirichlet boundary to the system.

◆ set_poiseuille_bcs()

void set_poiseuille_bcs ( TransientLinearImplicitSystem system)

Definition at line 749 of file systems_of_equations_ex2.C.

References libMesh::DofMap::add_dirichlet_boundary().

750 {
751  // This problem does not require a pressure pin, the Neumann outlet
752  // BCs are sufficient to set the value of the pressure.
753  system.get_equation_systems().parameters.set<bool>("pin_pressure") = false;
754 
755 #ifdef LIBMESH_ENABLE_DIRICHLET
756  unsigned short int
757  u_var = system.variable_number("vel_x"),
758  v_var = system.variable_number("vel_y");
759 
760  // Get a convenient reference to the System's DofMap
761  DofMap & dof_map = system.get_dof_map();
762 
763  {
764  // u=v=0 on top, bottom
765  std::set<boundary_id_type> boundary_ids;
766  boundary_ids.insert(0);
767  boundary_ids.insert(2);
768 
769  std::vector<unsigned int> variables;
770  variables.push_back(u_var);
771  variables.push_back(v_var);
772 
773  dof_map.add_dirichlet_boundary(DirichletBoundary(boundary_ids,
774  variables,
776  }
777  {
778  // u=quadratic on left
779  std::set<boundary_id_type> boundary_ids;
780  boundary_ids.insert(3);
781 
782  std::vector<unsigned int> variables;
783  variables.push_back(u_var);
784 
785  dof_map.add_dirichlet_boundary(DirichletBoundary(boundary_ids,
786  variables,
787  ParsedFunction<Number>("4*y*(1-y)")));
788  }
789  {
790  // v=0 on left
791  std::set<boundary_id_type> boundary_ids;
792  boundary_ids.insert(3);
793 
794  std::vector<unsigned int> variables;
795  variables.push_back(v_var);
796 
797  dof_map.add_dirichlet_boundary(DirichletBoundary(boundary_ids,
798  variables,
800  }
801 #endif // LIBMESH_ENABLE_DIRICHLET
802 }
ConstFunction that simply returns 0.
Definition: zero_function.h:38
A Function generated (via FParser) by parsing a mathematical expression.
This class allows one to associate Dirichlet boundary values with a given set of mesh boundary ids an...
This class handles the numbering of degrees of freedom on a mesh.
Definition: dof_map.h:169
void add_dirichlet_boundary(const DirichletBoundary &dirichlet_boundary)
Adds a copy of the specified Dirichlet boundary to the system.

◆ set_stagnation_bcs()

void set_stagnation_bcs ( TransientLinearImplicitSystem system)

Definition at line 696 of file systems_of_equations_ex2.C.

References libMesh::DofMap::add_dirichlet_boundary(), and libMesh::LOCAL_VARIABLE_ORDER.

697 {
698  // This problem does not require a pressure pin, the Neumann outlet
699  // BCs are sufficient to set the value of the pressure.
700  system.get_equation_systems().parameters.set<bool>("pin_pressure") = false;
701 
702 #ifdef LIBMESH_ENABLE_DIRICHLET
703  unsigned short int
704  u_var = system.variable_number("vel_x"),
705  v_var = system.variable_number("vel_y");
706 
707  // Get a convenient reference to the System's DofMap
708  DofMap & dof_map = system.get_dof_map();
709 
710  // u=v=0 on bottom (boundary 0)
711  dof_map.add_dirichlet_boundary(DirichletBoundary({0}, {u_var, v_var},
713  // u=0 on left (boundary 3) (symmetry)
714  dof_map.add_dirichlet_boundary(DirichletBoundary({3}, {u_var},
716  {
717  // u = k*x on top (boundary 2)
718 
719  // Set up ParsedFunction parameters
720  std::vector<std::string> additional_vars {"k"};
721  std::vector<Number> initial_vals {1.};
722 
723  dof_map.add_dirichlet_boundary(DirichletBoundary({2}, {u_var},
725  &additional_vars,
726  &initial_vals)));
727  }
728  {
729  // v = -k*y on top (boundary 2)
730 
731  // Set up ParsedFunction parameters
732  std::vector<std::string> additional_vars {"k"};
733  std::vector<Number> initial_vals {1.};
734 
735  // Note: we have to specify LOCAL_VARIABLE_ORDER here, since we're
736  // using a ParsedFunction to set the value of v_var, which is
737  // actually the second variable in the system.
738  dof_map.add_dirichlet_boundary(DirichletBoundary({2}, {v_var},
739  ParsedFunction<Number>("-k*y",
740  &additional_vars,
741  &initial_vals),
743  }
744 #endif // LIBMESH_ENABLE_DIRICHLET
745 }
ConstFunction that simply returns 0.
Definition: zero_function.h:38
A Function generated (via FParser) by parsing a mathematical expression.
This class allows one to associate Dirichlet boundary values with a given set of mesh boundary ids an...
This class handles the numbering of degrees of freedom on a mesh.
Definition: dof_map.h:169
void add_dirichlet_boundary(const DirichletBoundary &dirichlet_boundary)
Adds a copy of the specified Dirichlet boundary to the system.