#include <uniform_refinement_estimator.h>

Inheritance diagram for libMesh::UniformRefinementEstimator:

Public Types
typedef std::map< std::pair< const System , unsigned int >, ErrorVector >	ErrorMap

Public Member Functions
	UniformRefinementEstimator ()

	UniformRefinementEstimator (const UniformRefinementEstimator &)=default

	UniformRefinementEstimator (UniformRefinementEstimator &&)=default

UniformRefinementEstimator &	operator= (const UniformRefinementEstimator &)=default

UniformRefinementEstimator &	operator= (UniformRefinementEstimator &&)=default

virtual	~UniformRefinementEstimator ()=default

virtual void	estimate_error (const System &system, ErrorVector &error_per_cell, const NumericVector< Number > *solution_vector=nullptr, bool estimate_parent_error=false) override

virtual void	estimate_errors (const EquationSystems &equation_systems, ErrorVector &error_per_cell, const std::map< const System , SystemNorm > &error_norms, const std::map< const System , const NumericVector< Number > > solution_vectors=nullptr, bool estimate_parent_error=false) override

virtual void	estimate_errors (const EquationSystems &equation_systems, ErrorMap &errors_per_cell, const std::map< const System , const NumericVector< Number > > *solution_vectors=nullptr, bool estimate_parent_error=false) override

virtual ErrorEstimatorType	type () const override

Public Attributes
unsigned char	number_h_refinements

unsigned char	number_p_refinements

SystemNorm	error_norm

Protected Member Functions
virtual void	_estimate_error (const EquationSystems equation_systems, const System system, ErrorVector error_per_cell, std::map< std::pair< const System , unsigned int >, ErrorVector > errors_per_cell, const std::map< const System , SystemNorm > error_norms, const std::map< const System , const NumericVector< Number > > *solution_vectors=nullptr, bool estimate_parent_error=false)

void	reduce_error (std::vector< ErrorVectorReal > &error_per_cell, const Parallel::Communicator &comm) const

Detailed Description

This class implements a `‘brute force’' error estimator which integrates differences between the current solution and the solution on a uniformly refined (in h and/or p, for an arbitrary number of levels) grid.

Author: Roy H. Stogner

Date: 2006

Definition at line 45 of file uniform_refinement_estimator.h.

Member Typedef Documentation

◆ ErrorMap

typedef std::map<std::pair<const System *, unsigned int>, ErrorVector *> libMesh::ErrorEstimator::ErrorMap

inherited

When calculating many error vectors at once, we need a data structure to hold them all

Definition at line 124 of file error_estimator.h.

Constructor & Destructor Documentation

◆ UniformRefinementEstimator() [1/3]

libMesh::UniformRefinementEstimator::UniformRefinementEstimator ( )

Constructor. Sets the most common default parameter values.

Definition at line 53 of file uniform_refinement_estimator.C.

References libMesh::ErrorEstimator::error_norm, and libMesh::H1.

                                                        :
     ErrorEstimator(),
     number_h_refinements(1),
     number_p_refinements(0)
 {
   error_norm = H1;
 }

◆ UniformRefinementEstimator() [2/3]

libMesh::UniformRefinementEstimator::UniformRefinementEstimator ( const UniformRefinementEstimator & )

default

Copy/move ctor, copy/move assignment operator, and destructor are all explicitly defaulted for this simple class.

◆ UniformRefinementEstimator() [3/3]

libMesh::UniformRefinementEstimator::UniformRefinementEstimator ( UniformRefinementEstimator && )

default

◆ ~UniformRefinementEstimator()

virtual libMesh::UniformRefinementEstimator::~UniformRefinementEstimator ( )

virtualdefault

Member Function Documentation

◆ _estimate_error()

void libMesh::UniformRefinementEstimator::_estimate_error	(	const EquationSystems *	equation_systems,
		const System *	system,
		ErrorVector *	error_per_cell,
		std::map< std::pair< const System , unsigned int >, ErrorVector > *	errors_per_cell,
		const std::map< const System , SystemNorm >	error_norms,
		const std::map< const System , const NumericVector< Number > > *	solution_vectors = `nullptr`,
		bool	estimate_parent_error = `false`
	)

protectedvirtual

The code for estimate_error and both estimate_errors versions is very similar, so we use the same function for all three

Definition at line 117 of file uniform_refinement_estimator.C.

Referenced by estimate_error(), and estimate_errors().

 {
   // Get a vector of the Systems we're going to work on,
   // and set up a error_norms map if necessary
   std::vector<System *> system_list;
   std::unique_ptr<std::map<const System *, SystemNorm>> error_norms =
     std::unique_ptr<std::map<const System *, SystemNorm>>
     (new std::map<const System *, SystemNorm>);
 
   if (_es)
     {
       libmesh_assert(!_system);
       libmesh_assert(_es->n_systems());
       _system = &(_es->get_system(0));
       libmesh_assert_equal_to (&(_system->get_equation_systems()), _es);
 
       libmesh_assert(_es->n_systems());
       for (unsigned int i=0; i != _es->n_systems(); ++i)
         // We have to break the rules here, because we can't refine a const System
         system_list.push_back(const_cast<System *>(&(_es->get_system(i))));
 
       // If we're computing one vector, we need to know how to scale
       // each variable's contributions to it.
       if (_error_norms)
         {
           libmesh_assert(!errors_per_cell);
         }
       else
         // If we're computing many vectors, we just need to know which
         // variables to skip
         {
           libmesh_assert (errors_per_cell);
 
           _error_norms = error_norms.get();
 
           for (unsigned int i=0; i!= _es->n_systems(); ++i)
             {
               const System & sys = _es->get_system(i);
               unsigned int n_vars = sys.n_vars();
 
               std::vector<Real> weights(n_vars, 0.0);
               for (unsigned int v = 0; v != n_vars; ++v)
                 {
                   if (errors_per_cell->find(std::make_pair(&sys, v)) ==
                       errors_per_cell->end())
                     continue;
 
                   weights[v] = 1.0;
                 }
               (*error_norms)[&sys] =
                 SystemNorm(std::vector<FEMNormType>(n_vars, error_norm.type(0)),
                            weights);
             }
         }
     }
   else
     {
       libmesh_assert(_system);
       // We have to break the rules here, because we can't refine a const System
       system_list.push_back(const_cast<System *>(_system));
 
       libmesh_assert(!_error_norms);
       (*error_norms)[_system] = error_norm;
       _error_norms = error_norms.get();
     }
 
   // An EquationSystems reference will be convenient.
   // We have to break the rules here, because we can't refine a const System
   EquationSystems & es =
     const_cast<EquationSystems &>(_system->get_equation_systems());
 
   // The current mesh
   MeshBase & mesh = es.get_mesh();
 
   // The dimensionality of the mesh
   const unsigned int dim = mesh.mesh_dimension();
 
   // Resize the error_per_cell vectors to be
   // the number of elements, initialize them to 0.
   if (error_per_cell)
     {
       error_per_cell->clear();
       error_per_cell->resize (mesh.max_elem_id(), 0.);
     }
   else
     {
       libmesh_assert(errors_per_cell);
       for (const auto & pr : *errors_per_cell)
         {
           ErrorVector * e = pr.second;
           e->clear();
           e->resize(mesh.max_elem_id(), 0.);
         }
     }
 
   // We'll want to back up all coarse grid vectors
   std::vector<std::map<std::string, std::unique_ptr<NumericVector<Number>>>> coarse_vectors(system_list.size());
   std::vector<std::unique_ptr<NumericVector<Number>>> coarse_solutions(system_list.size());
   std::vector<std::unique_ptr<NumericVector<Number>>> coarse_local_solutions(system_list.size());
   // And make copies of projected solutions
   std::vector<std::unique_ptr<NumericVector<Number>>> projected_solutions(system_list.size());
 
   // And we'll need to temporarily change solution projection settings
   std::vector<bool> old_projection_settings(system_list.size());
 
   // And it'll be best to avoid any repartitioning
   std::unique_ptr<Partitioner> old_partitioner(mesh.partitioner().release());
 
   // And we can't allow any renumbering
   const bool old_renumbering_setting = mesh.allow_renumbering();
   mesh.allow_renumbering(false);
 
   for (auto i : index_range(system_list))
     {
       System & system = *system_list[i];
 
       // Check for valid error_norms
       libmesh_assert (_error_norms->find(&system) !=
                       _error_norms->end());
 
       // Back up the solution vector
       coarse_solutions[i] = system.solution->clone();
       coarse_local_solutions[i] = system.current_local_solution->clone();
 
       // Back up all other coarse grid vectors
       for (System::vectors_iterator vec = system.vectors_begin(); vec !=
              system.vectors_end(); ++vec)
         {
           // The (string) name of this vector
           const std::string & var_name = vec->first;
 
           coarse_vectors[i][var_name] = vec->second->clone();
         }
 
       // Use a non-standard solution vector if necessary
       if (solution_vectors &&
           solution_vectors->find(&system) != solution_vectors->end() &&
           solution_vectors->find(&system)->second &&
           solution_vectors->find(&system)->second != system.solution.get())
         {
           NumericVector<Number> * newsol =
             const_cast<NumericVector<Number> *>
             (solution_vectors->find(&system)->second);
           newsol->swap(*system.solution);
           system.update();
         }
 
       // Make sure the solution is projected when we refine the mesh
       old_projection_settings[i] = system.project_solution_on_reinit();
       system.project_solution_on_reinit() = true;
     }
 
   // Find the number of coarse mesh elements, to make it possible
   // to find correct coarse elem ids later
   const dof_id_type max_coarse_elem_id = mesh.max_elem_id();
 #ifndef NDEBUG
   // n_coarse_elem is only used in an assertion later so
   // avoid declaring it unless asserts are active.
   const dof_id_type n_coarse_elem = mesh.n_elem();
 #endif
 
   // Uniformly refine the mesh
   MeshRefinement mesh_refinement(mesh);
 
   libmesh_assert (number_h_refinements > 0 || number_p_refinements > 0);
 
   // FIXME: this may break if there is more than one System
   // on this mesh but estimate_error was still called instead of
   // estimate_errors
   for (unsigned int i = 0; i != number_h_refinements; ++i)
     {
       mesh_refinement.uniformly_refine(1);
       es.reinit();
     }
 
   for (unsigned int i = 0; i != number_p_refinements; ++i)
     {
       mesh_refinement.uniformly_p_refine(1);
       es.reinit();
     }
 
   for (auto i : index_range(system_list))
     {
       System & system = *system_list[i];
 
       // Copy the projected coarse grid solutions, which will be
       // overwritten by solve()
       projected_solutions[i] = NumericVector<Number>::build(system.comm());
       projected_solutions[i]->init(system.solution->size(), true, SERIAL);
       system.solution->localize(*projected_solutions[i],
                                 system.get_dof_map().get_send_list());
     }
 
   // Are we doing a forward or an adjoint solve?
   bool solve_adjoint = false;
   if (solution_vectors)
     {
       System * sys = system_list[0];
       libmesh_assert (solution_vectors->find(sys) !=
                       solution_vectors->end());
       const NumericVector<Number> * vec = solution_vectors->find(sys)->second;
       for (unsigned int j=0; j != sys->n_qois(); ++j)
         {
           std::ostringstream adjoint_name;
           adjoint_name << "adjoint_solution" << j;
 
           if (vec == sys->request_vector(adjoint_name.str()))
             {
               solve_adjoint = true;
               break;
             }
         }
     }
 
   // Get the uniformly refined solution.
 
   if (_es)
     {
       // Even if we had a decent preconditioner, valid matrix etc. before
       // refinement, we don't any more.
       for (unsigned int i=0; i != es.n_systems(); ++i)
         es.get_system(i).disable_cache();
 
       // No specified vectors == forward solve
       if (!solution_vectors)
         es.solve();
       else
         {
           libmesh_assert_equal_to (solution_vectors->size(), es.n_systems());
           libmesh_assert (solution_vectors->find(system_list[0]) !=
                           solution_vectors->end());
           libmesh_assert(solve_adjoint ||
                          (solution_vectors->find(system_list[0])->second ==
                           system_list[0]->solution.get()) ||
                          !solution_vectors->find(system_list[0])->second);
 
 #ifdef DEBUG
           for (const auto & sys : system_list)
             {
               libmesh_assert (solution_vectors->find(sys) !=
                               solution_vectors->end());
               const NumericVector<Number> * vec = solution_vectors->find(sys)->second;
               if (solve_adjoint)
                 {
                   bool found_vec = false;
                   for (unsigned int j=0; j != sys->n_qois(); ++j)
                     {
                       std::ostringstream adjoint_name;
                       adjoint_name << "adjoint_solution" << j;
 
                       if (vec == sys->request_vector(adjoint_name.str()))
                         {
                           found_vec = true;
                           break;
                         }
                     }
                   libmesh_assert(found_vec);
                 }
               else
                 libmesh_assert(vec == sys->solution.get() || !vec);
             }
 #endif
 
           if (solve_adjoint)
             {
               std::vector<unsigned int> adjs(system_list.size(),
                                              libMesh::invalid_uint);
               // Set up proper initial guesses
               for (auto i : index_range(system_list))
                 {
                   System * sys = system_list[i];
                   libmesh_assert (solution_vectors->find(sys) !=
                                   solution_vectors->end());
                   const NumericVector<Number> * vec = solution_vectors->find(sys)->second;
                   for (unsigned int j=0, n_qois = sys->n_qois();
                        j != n_qois; ++j)
                     {
                       std::ostringstream adjoint_name;
                       adjoint_name << "adjoint_solution" << j;
 
                       if (vec == sys->request_vector(adjoint_name.str()))
                         {
                           adjs[i] = j;
                           break;
                         }
                     }
                   libmesh_assert_not_equal_to (adjs[i], libMesh::invalid_uint);
                   sys->get_adjoint_solution(adjs[i]) = *sys->solution;
                 }
 
               es.adjoint_solve();
 
               // Put the adjoint_solution into solution for
               // comparisons
               for (auto i : index_range(system_list))
                 {
                   system_list[i]->get_adjoint_solution(adjs[i]).swap(*system_list[i]->solution);
                   system_list[i]->update();
                 }
             }
           else
             es.solve();
         }
     }
   else
     {
       System * sys = system_list[0];
 
       // Even if we had a decent preconditioner, valid matrix etc. before
       // refinement, we don't any more.
       sys->disable_cache();
 
       // No specified vectors == forward solve
       if (!solution_vectors)
         sys->solve();
       else
         {
           libmesh_assert (solution_vectors->find(sys) !=
                           solution_vectors->end());
 
           const NumericVector<Number> * vec = solution_vectors->find(sys)->second;
 
           libmesh_assert(solve_adjoint ||
                          (solution_vectors->find(sys)->second ==
                           sys->solution.get()) ||
                          !solution_vectors->find(sys)->second);
 
           if (solve_adjoint)
             {
               unsigned int adj = libMesh::invalid_uint;
               for (unsigned int j=0, n_qois = sys->n_qois();
                    j != n_qois; ++j)
                 {
                   std::ostringstream adjoint_name;
                   adjoint_name << "adjoint_solution" << j;
 
                   if (vec == sys->request_vector(adjoint_name.str()))
                     {
                       adj = j;
                       break;
                     }
                 }
               libmesh_assert_not_equal_to (adj, libMesh::invalid_uint);
 
               // Set up proper initial guess
               sys->get_adjoint_solution(adj) = *sys->solution;
               sys->adjoint_solve();
               // Put the adjoint_solution into solution for
               // comparisons
               sys->get_adjoint_solution(adj).swap(*sys->solution);
               sys->update();
             }
           else
             sys->solve();
         }
     }
 
   // Get the error in the uniformly refined solution(s).
   for (auto sysnum : index_range(system_list))
     {
       System & system = *system_list[sysnum];
 
       unsigned int n_vars = system.n_vars();
 
       DofMap & dof_map = system.get_dof_map();
 
       const SystemNorm & system_i_norm =
         _error_norms->find(&system)->second;
 
       NumericVector<Number> * projected_solution = projected_solutions[sysnum].get();
 
       // Loop over all the variables in the system
       for (unsigned int var=0; var<n_vars; var++)
         {
           // Get the error vector to fill for this system and variable
           ErrorVector * err_vec = error_per_cell;
           if (!err_vec)
             {
               libmesh_assert(errors_per_cell);
               err_vec =
                 (*errors_per_cell)[std::make_pair(&system,var)];
             }
 
           // The type of finite element to use for this variable
           const FEType & fe_type = dof_map.variable_type (var);
 
           // Finite element object for each fine element
           std::unique_ptr<FEBase> fe (FEBase::build (dim, fe_type));
 
           // Build and attach an appropriate quadrature rule
           std::unique_ptr<QBase> qrule = fe_type.default_quadrature_rule(dim);
           fe->attach_quadrature_rule (qrule.get());
 
           const std::vector<Real> &  JxW = fe->get_JxW();
           const std::vector<std::vector<Real>> & phi = fe->get_phi();
           const std::vector<std::vector<RealGradient>> & dphi =
             fe->get_dphi();
 #ifdef LIBMESH_ENABLE_SECOND_DERIVATIVES
           const std::vector<std::vector<RealTensor>> & d2phi =
             fe->get_d2phi();
 #endif
 
           // The global DOF indices for the fine element
           std::vector<dof_id_type> dof_indices;
 
           // Iterate over all the active elements in the fine mesh
           // that live on this processor.
           for (const auto & elem : mesh.active_local_element_ptr_range())
             {
               // Find the element id for the corresponding coarse grid element
               const Elem * coarse = elem;
               dof_id_type e_id = coarse->id();
               while (e_id >= max_coarse_elem_id)
                 {
                   libmesh_assert (coarse->parent());
                   coarse = coarse->parent();
                   e_id = coarse->id();
                 }
 
               Real L2normsq = 0., H1seminormsq = 0.;
 #ifdef LIBMESH_ENABLE_SECOND_DERIVATIVES
               Real H2seminormsq = 0.;
 #endif
 
               // reinitialize the element-specific data
               // for the current element
               fe->reinit (elem);
 
               // Get the local to global degree of freedom maps
               dof_map.dof_indices (elem, dof_indices, var);
 
               // The number of quadrature points
               const unsigned int n_qp = qrule->n_points();
 
               // The number of shape functions
               const unsigned int n_sf =
                 cast_int<unsigned int>(dof_indices.size());
 
               //
               // Begin the loop over the Quadrature points.
               //
               for (unsigned int qp=0; qp<n_qp; qp++)
                 {
                   Number u_fine = 0., u_coarse = 0.;
 
                   Gradient grad_u_fine, grad_u_coarse;
 #ifdef LIBMESH_ENABLE_SECOND_DERIVATIVES
                   Tensor grad2_u_fine, grad2_u_coarse;
 #endif
 
                   // Compute solution values at the current
                   // quadrature point.  This requires a sum
                   // over all the shape functions evaluated
                   // at the quadrature point.
                   for (unsigned int i=0; i<n_sf; i++)
                     {
                       u_fine            += phi[i][qp]*system.current_solution (dof_indices[i]);
                       u_coarse          += phi[i][qp]*(*projected_solution) (dof_indices[i]);
                       grad_u_fine       += dphi[i][qp]*system.current_solution (dof_indices[i]);
                       grad_u_coarse     += dphi[i][qp]*(*projected_solution) (dof_indices[i]);
 #ifdef LIBMESH_ENABLE_SECOND_DERIVATIVES
                       grad2_u_fine      += d2phi[i][qp]*system.current_solution (dof_indices[i]);
                       grad2_u_coarse    += d2phi[i][qp]*(*projected_solution) (dof_indices[i]);
 #endif
                     }
 
                   // Compute the value of the error at this quadrature point
                   const Number val_error = u_fine - u_coarse;
 
                   // Add the squares of the error to each contribution
                   if (system_i_norm.type(var) == L2 ||
                       system_i_norm.type(var) == H1 ||
                       system_i_norm.type(var) == H2)
                     {
                       L2normsq += JxW[qp] * system_i_norm.weight_sq(var) *
                         TensorTools::norm_sq(val_error);
                       libmesh_assert_greater_equal (L2normsq, 0.);
                     }
 
 
                   // Compute the value of the error in the gradient at this
                   // quadrature point
                   if (system_i_norm.type(var) == H1 ||
                       system_i_norm.type(var) == H2 ||
                       system_i_norm.type(var) == H1_SEMINORM)
                     {
                       Gradient grad_error = grad_u_fine - grad_u_coarse;
 
                       H1seminormsq += JxW[qp] * system_i_norm.weight_sq(var) *
                         grad_error.norm_sq();
                       libmesh_assert_greater_equal (H1seminormsq, 0.);
                     }
 
                   // Compute the value of the error in the hessian at this
                   // quadrature point
                   if (system_i_norm.type(var) == H2 ||
                       system_i_norm.type(var) == H2_SEMINORM)
                     {
 #ifdef LIBMESH_ENABLE_SECOND_DERIVATIVES
                       Tensor grad2_error = grad2_u_fine - grad2_u_coarse;
 
                       H2seminormsq += JxW[qp] * system_i_norm.weight_sq(var) *
                         grad2_error.norm_sq();
                       libmesh_assert_greater_equal (H2seminormsq, 0.);
 #else
                       libmesh_error_msg
                         ("libMesh was not configured with --enable-second");
 #endif
                     }
                 } // end qp loop
 
               if (system_i_norm.type(var) == L2 ||
                   system_i_norm.type(var) == H1 ||
                   system_i_norm.type(var) == H2)
                 (*err_vec)[e_id] +=
                   static_cast<ErrorVectorReal>(L2normsq);
               if (system_i_norm.type(var) == H1 ||
                   system_i_norm.type(var) == H2 ||
                   system_i_norm.type(var) == H1_SEMINORM)
                 (*err_vec)[e_id] +=
                   static_cast<ErrorVectorReal>(H1seminormsq);
 
 #ifdef LIBMESH_ENABLE_SECOND_DERIVATIVES
               if (system_i_norm.type(var) == H2 ||
                   system_i_norm.type(var) == H2_SEMINORM)
                 (*err_vec)[e_id] +=
                   static_cast<ErrorVectorReal>(H2seminormsq);
 #endif
             } // End loop over active local elements
         } // End loop over variables
 
       // Don't bother projecting the solution; we'll restore from backup
       // after coarsening
       system.project_solution_on_reinit() = false;
     }
 
 
   // Uniformly coarsen the mesh, without projecting the solution
   libmesh_assert (number_h_refinements > 0 || number_p_refinements > 0);
 
   for (unsigned int i = 0; i != number_h_refinements; ++i)
     {
       mesh_refinement.uniformly_coarsen(1);
       // FIXME - should the reinits here be necessary? - RHS
       es.reinit();
     }
 
   for (unsigned int i = 0; i != number_p_refinements; ++i)
     {
       mesh_refinement.uniformly_p_coarsen(1);
       es.reinit();
     }
 
   // We should be back where we started
   libmesh_assert_equal_to (n_coarse_elem, mesh.n_elem());
 
   // Each processor has now computed the error contributions
   // for its local elements.  We need to sum the vector
   // and then take the square-root of each component.  Note
   // that we only need to sum if we are running on multiple
   // processors, and we only need to take the square-root
   // if the value is nonzero.  There will in general be many
   // zeros for the inactive elements.
 
   if (error_per_cell)
     {
       // First sum the vector of estimated error values
       this->reduce_error(*error_per_cell, es.comm());
 
       // Compute the square-root of each component.
       LOG_SCOPE("std::sqrt()", "UniformRefinementEstimator");
       for (auto & val : *error_per_cell)
         if (val != 0.)
           val = std::sqrt(val);
     }
   else
     {
       for (const auto & pr : *errors_per_cell)
         {
           ErrorVector & e = *(pr.second);
           // First sum the vector of estimated error values
           this->reduce_error(e, es.comm());
 
           // Compute the square-root of each component.
           LOG_SCOPE("std::sqrt()", "UniformRefinementEstimator");
           for (auto & val : e)
             if (val != 0.)
               val = std::sqrt(val);
         }
     }
 
   // Restore old solutions and clean up the heap
   for (auto i : index_range(system_list))
     {
       System & system = *system_list[i];
 
       system.project_solution_on_reinit() = old_projection_settings[i];
 
       // Restore the coarse solution vectors and delete their copies
       *system.solution = *coarse_solutions[i];
       *system.current_local_solution = *coarse_local_solutions[i];
 
       for (System::vectors_iterator vec = system.vectors_begin(); vec !=
              system.vectors_end(); ++vec)
         {
           // The (string) name of this vector
           const std::string & var_name = vec->first;
 
           system.get_vector(var_name) = *coarse_vectors[i][var_name];
 
           coarse_vectors[i][var_name]->clear();
         }
     }
 
   // Restore old partitioner and renumbering settings
   mesh.partitioner().reset(old_partitioner.release());
   mesh.allow_renumbering(old_renumbering_setting);
 }

◆ estimate_error()

void libMesh::UniformRefinementEstimator::estimate_error	(	const System &	system,
		ErrorVector &	error_per_cell,
		const NumericVector< Number > *	solution_vector = `nullptr`,
		bool	estimate_parent_error = `false`
	)

overridevirtual

This function does uniform refinements and a solve to get an improved solution on each cell, then estimates the error by integrating differences between the coarse and fine solutions.

system.solve() must be called, and so should have no side effects.

Only the provided system is solved on the refined mesh; for problems decoupled into multiple systems, use of estimate_errors() should be more reliable.

The estimated error is output in the vector error_per_cell

Implements libMesh::ErrorEstimator.

Definition at line 69 of file uniform_refinement_estimator.C.

References _estimate_error().

 {
   LOG_SCOPE("estimate_error()", "UniformRefinementEstimator");
   std::map<const System *, const NumericVector<Number> *> solution_vectors;
   solution_vectors[&_system] = solution_vector;
   this->_estimate_error (nullptr,
                          &_system,
                          &error_per_cell,
                          nullptr,
                          nullptr,
                          &solution_vectors,
                          estimate_parent_error);
 }

◆ estimate_errors() [1/2]

void libMesh::UniformRefinementEstimator::estimate_errors	(	const EquationSystems &	equation_systems,
		ErrorVector &	error_per_cell,
		const std::map< const System *, SystemNorm > &	error_norms,
		const std::map< const System , const NumericVector< Number > > *	solution_vectors = `nullptr`,
		bool	estimate_parent_error = `false`
	)

overridevirtual

Currently this function ignores the error_norm member variable, and uses the function argument error_norms instead.

This function is named estimate_errors instead of estimate_error because otherwise C++ can get confused.

Reimplemented from libMesh::ErrorEstimator.

Definition at line 86 of file uniform_refinement_estimator.C.

References _estimate_error().

 {
   LOG_SCOPE("estimate_errors()", "UniformRefinementEstimator");
   this->_estimate_error (&_es,
                          nullptr,
                          &error_per_cell,
                          nullptr,
                          &error_norms,
                          solution_vectors,
                          estimate_parent_error);
 }

◆ estimate_errors() [2/2]

void libMesh::UniformRefinementEstimator::estimate_errors	(	const EquationSystems &	equation_systems,
		ErrorMap &	errors_per_cell,
		const std::map< const System , const NumericVector< Number > > *	solution_vectors = `nullptr`,
		bool	estimate_parent_error = `false`
	)

overridevirtual

Currently this function ignores the component_scale member variable, because it calculates each error individually, unscaled.

The user selects which errors get computed by filling a map with error vectors: If errors_per_cell[&system][v] exists, it will be filled with the error values in variable v of system

Reimplemented from libMesh::ErrorEstimator.

Definition at line 102 of file uniform_refinement_estimator.C.

References _estimate_error().

 {
   LOG_SCOPE("estimate_errors()", "UniformRefinementEstimator");
   this->_estimate_error (&_es,
                          nullptr,
                          nullptr,
                          &errors_per_cell,
                          nullptr,
                          solution_vectors,
                          estimate_parent_error);
 }

◆ operator=() [1/2]

UniformRefinementEstimator& libMesh::UniformRefinementEstimator::operator= ( const UniformRefinementEstimator & )

default

◆ operator=() [2/2]

UniformRefinementEstimator& libMesh::UniformRefinementEstimator::operator= ( UniformRefinementEstimator && )

default

◆ reduce_error()

void libMesh::ErrorEstimator::reduce_error	(	std::vector< ErrorVectorReal > &	error_per_cell,
		const Parallel::Communicator &	comm
	)		const

protectedinherited

This method takes the local error contributions in error_per_cell from each processor and combines them to get the global error vector.

Definition at line 32 of file error_estimator.C.

References libMesh::Parallel::Communicator::sum().

Referenced by _estimate_error(), libMesh::WeightedPatchRecoveryErrorEstimator::estimate_error(), libMesh::PatchRecoveryErrorEstimator::estimate_error(), libMesh::JumpErrorEstimator::estimate_error(), libMesh::AdjointRefinementEstimator::estimate_error(), and libMesh::ExactErrorEstimator::estimate_error().

 {
   // This function must be run on all processors at once
   // parallel_object_only();
 
   // Each processor has now computed the error contributions
   // for its local elements.  We may need to sum the vector to
   // recover the error for each element.
 
   comm.sum(error_per_cell);
 }

◆ type()

ErrorEstimatorType libMesh::UniformRefinementEstimator::type ( ) const

overridevirtual

Returns: The type for the ErrorEstimator subclass.

Implements libMesh::ErrorEstimator.

Definition at line 63 of file uniform_refinement_estimator.C.

References libMesh::UNIFORM_REFINEMENT.

 {
   return UNIFORM_REFINEMENT;
 }

Member Data Documentation

◆ error_norm

SystemNorm libMesh::ErrorEstimator::error_norm

inherited

When estimating the error in a single system, the error_norm is used to control the scaling and norm choice for each variable. Not all estimators will support all norm choices. The default scaling is for all variables to be weighted equally. The default norm choice depends on the error estimator.

Part of this functionality was supported via component_scale and sobolev_order in older libMesh versions, and a small part was supported via component_mask in even older versions. Hopefully the encapsulation here will allow us to avoid changing this API again.

Definition at line 161 of file error_estimator.h.

Referenced by _estimate_error(), libMesh::AdjointRefinementEstimator::AdjointRefinementEstimator(), libMesh::DiscontinuityMeasure::boundary_side_integration(), libMesh::KellyErrorEstimator::boundary_side_integration(), libMesh::DiscontinuityMeasure::DiscontinuityMeasure(), libMesh::JumpErrorEstimator::estimate_error(), libMesh::AdjointResidualErrorEstimator::estimate_error(), libMesh::ExactErrorEstimator::estimate_error(), libMesh::ErrorEstimator::estimate_errors(), libMesh::ExactErrorEstimator::ExactErrorEstimator(), libMesh::ExactErrorEstimator::find_squared_element_error(), libMesh::LaplacianErrorEstimator::init_context(), libMesh::DiscontinuityMeasure::init_context(), libMesh::KellyErrorEstimator::init_context(), libMesh::LaplacianErrorEstimator::internal_side_integration(), libMesh::DiscontinuityMeasure::internal_side_integration(), libMesh::KellyErrorEstimator::internal_side_integration(), libMesh::KellyErrorEstimator::KellyErrorEstimator(), libMesh::LaplacianErrorEstimator::LaplacianErrorEstimator(), libMesh::WeightedPatchRecoveryErrorEstimator::EstimateError::operator()(), libMesh::PatchRecoveryErrorEstimator::EstimateError::operator()(), libMesh::PatchRecoveryErrorEstimator::PatchRecoveryErrorEstimator(), libMesh::JumpErrorEstimator::reinit_sides(), and UniformRefinementEstimator().