This class measures discontinuities between elements for debugging purposes. More...

#include <discontinuity_measure.h>

Inheritance diagram for libMesh::DiscontinuityMeasure:

Public Types
typedef std::map< std::pair< const System *, unsigned int >, std::unique_ptr< ErrorVector > >	ErrorMap
	When calculating many error vectors at once, we need a data structure to hold them all. More...

Public Member Functions
	DiscontinuityMeasure ()
	Constructor. More...

	DiscontinuityMeasure (const DiscontinuityMeasure &)=delete
	This class cannot be (default) copy constructed/assigned because its base class has unique_ptr members. More...

DiscontinuityMeasure &	operator= (const DiscontinuityMeasure &)=delete

	DiscontinuityMeasure (DiscontinuityMeasure &&)=default
	Defaulted move ctor, move assignment operator, and destructor. More...

DiscontinuityMeasure &	operator= (DiscontinuityMeasure &&)=default

virtual	~DiscontinuityMeasure ()=default

void	attach_essential_bc_function (std::pair< bool, Real > fptr(const System &system, const Point &p, const std::string &var_name))
	Register a user function to use in computing the essential BCs. More...

virtual ErrorEstimatorType	type () const override

virtual void	estimate_error (const System &system, ErrorVector &error_per_cell, const NumericVector< Number > *solution_vector=nullptr, bool estimate_parent_error=false) override
	This function uses the derived class's jump error estimate formula to estimate the error on each cell. More...

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)
	This virtual function can be redefined in derived classes, but by default computes the sum of the error_per_cell for each system in the equation_systems. More...

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)
	This virtual function can be redefined in derived classes, but by default it calls estimate_error repeatedly to calculate the requested error vectors. More...

Public Attributes
bool	scale_by_n_flux_faces
	This boolean flag allows you to scale the error indicator result for each element by the number of "flux faces" the element actually has. More...

bool	use_unweighted_quadrature_rules
	This boolean flag allows you to use "unweighted" quadrature rules (sized to exactly integrate unweighted shape functions in master element space) rather than "default" quadrature rules (sized to exactly integrate polynomials of one higher degree than mass matrix terms). More...

bool	integrate_slits
	A boolean flag, by default false, to be set to true if integrations should be performed on "slits" where two elements' faces overlap even if those elements are not connected by neighbor links. More...

SystemNorm	error_norm
	When estimating the error in a single system, the `error_norm` is used to control the scaling and norm choice for each variable. More...

Protected Member Functions
virtual void	init_context (FEMContext &c) override
	An initialization function, for requesting specific data from the FE objects. More...

virtual void	internal_side_integration () override
	The function which calculates a normal derivative jump based error term on an internal side. More...

virtual bool	boundary_side_integration () override
	The function which calculates a normal derivative jump based error term on a boundary side. More...

void	reinit_sides ()
	A utility function to reinit the finite element data on elements sharing a side. More...

float	coarse_n_flux_faces_increment ()
	A utility function to correctly increase n_flux_faces for the coarse element. More...

void	reduce_error (std::vector< ErrorVectorReal > &error_per_cell, const Parallel::Communicator &comm) const
	This method takes the local error contributions in `error_per_cell` from each processor and combines them to get the global error vector. More...

Protected Attributes
std::pair< bool, Real >(*	_bc_function )(const System &system, const Point &p, const std::string &var_name)
	Pointer to function that provides BC information. More...

bool	integrate_boundary_sides
	A boolean flag, by default false, to be set to true if integrations with boundary_side_integration() should be performed. More...

std::unique_ptr< FEMContext >	fine_context
	Context objects for integrating on the fine and coarse elements sharing a face. More...

std::unique_ptr< FEMContext >	coarse_context

Real	fine_error
	The fine and coarse error values to be set by each side_integration();. More...

Real	coarse_error

unsigned int	var
	The variable number currently being evaluated. More...

Detailed Description

This class measures discontinuities between elements for debugging purposes.

It derives from ErrorEstimator just in case someone finds it useful in a DG framework.

Author: Roy H. Stogner

Date: 2006

Definition at line 49 of file discontinuity_measure.h.

Member Typedef Documentation

◆ ErrorMap

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

inherited

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

Definition at line 121 of file error_estimator.h.

Constructor & Destructor Documentation

◆ DiscontinuityMeasure() [1/3]

libMesh::DiscontinuityMeasure::DiscontinuityMeasure ( )

Constructor.

Responsible for initializing the _bc_function function pointer to nullptr. Defaults to L2 norm; changes to system norm are ignored.

Definition at line 41 of file discontinuity_measure.C.

References libMesh::ErrorEstimator::error_norm, and libMesh::L2.

                                            :
   JumpErrorEstimator(),
   _bc_function(nullptr)
 {
   error_norm = L2;
 }

◆ DiscontinuityMeasure() [2/3]

libMesh::DiscontinuityMeasure::DiscontinuityMeasure ( const DiscontinuityMeasure & )

delete

This class cannot be (default) copy constructed/assigned because its base class has unique_ptr members.

◆ DiscontinuityMeasure() [3/3]

libMesh::DiscontinuityMeasure::DiscontinuityMeasure ( DiscontinuityMeasure && )

default

Defaulted move ctor, move assignment operator, and destructor.

◆ ~DiscontinuityMeasure()

virtual libMesh::DiscontinuityMeasure::~DiscontinuityMeasure ( )

virtualdefault

Member Function Documentation

◆ attach_essential_bc_function()

void libMesh::DiscontinuityMeasure::attach_essential_bc_function ( std::pair< bool, Real > fptrconst System &system, const Point &p, const std::string &var_name )

Register a user function to use in computing the essential BCs.

Definition at line 202 of file discontinuity_measure.C.

References _bc_function, fptr(), and libMesh::JumpErrorEstimator::integrate_boundary_sides.

 {
   _bc_function = fptr;
 
   // We may be turning boundary side integration on or off
   if (fptr)
     integrate_boundary_sides = true;
   else
     integrate_boundary_sides = false;
 }

◆ boundary_side_integration()

bool libMesh::DiscontinuityMeasure::boundary_side_integration ( )

overrideprotectedvirtual

The function which calculates a normal derivative jump based error term on a boundary side.

Returns: true if the flux bc function is in fact defined on the current side.

Reimplemented from libMesh::JumpErrorEstimator.

Definition at line 132 of file discontinuity_measure.C.

References _bc_function, libMesh::Elem::dim(), libMesh::ErrorEstimator::error_norm, libMesh::JumpErrorEstimator::fine_context, libMesh::JumpErrorEstimator::fine_error, libMesh::FEAbstract::get_JxW(), libMesh::FEAbstract::get_xyz(), libMesh::Elem::hmax(), libMesh::FEAbstract::n_quadrature_points(), libMesh::TensorTools::norm_sq(), libMesh::Real, libMesh::JumpErrorEstimator::var, and libMesh::SystemNorm::weight().

 {
   const Elem & fine_elem = fine_context->get_elem();
 
   FEBase * fe_fine = nullptr;
   fine_context->get_side_fe( var, fe_fine, fine_elem.dim() );
 
   const std::string & var_name =
     fine_context->get_system().variable_name(var);
 
   const std::vector<Real> & JxW_face = fe_fine->get_JxW();
   const std::vector<Point> & qface_point = fe_fine->get_xyz();
 
   // The reinitialization also recomputes the locations of
   // the quadrature points on the side.  By checking if the
   // first quadrature point on the side is on an essential boundary
   // for a particular variable, we will determine if the whole
   // element is on an essential boundary (assuming quadrature points
   // are strictly contained in the side).
   if (this->_bc_function(fine_context->get_system(),
                          qface_point[0], var_name).first)
     {
       const Real h = fine_elem.hmax();
 
       // The number of quadrature points
       const unsigned int n_qp = fe_fine->n_quadrature_points();
 
       // The error contribution from this face
       Real error = 1.e-30;
 
       // loop over the integration points on the face.
       for (unsigned int qp=0; qp<n_qp; qp++)
         {
           // Value of the imposed essential BC at this quadrature point.
           const std::pair<bool,Real> essential_bc =
             this->_bc_function(fine_context->get_system(), qface_point[qp],
                                var_name);
 
           // Be sure the BC function still thinks we're on the
           // essential boundary.
           libmesh_assert_equal_to (essential_bc.first, true);
 
           // The solution value on each point
           Number u_fine = fine_context->side_value(var, qp);
 
           // The difference between the desired BC and the approximate solution.
           const Number jump = essential_bc.second - u_fine;
 
           // The flux jump squared.  If using complex numbers,
           // norm_sq(z) returns |z|^2, where |z| is the modulus of z.
           const Real jump2 = TensorTools::norm_sq(jump);
 
           // Integrate the error on the face.  The error is
           // scaled by an additional power of h, where h is
           // the maximum side length for the element.  This
           // arises in the definition of the indicator.
           error += JxW_face[qp]*jump2;
 
         } // End quadrature point loop
 
       fine_error = error*h*error_norm.weight(var);
 
       return true;
     } // end if side on flux boundary
   return false;
 }

◆ coarse_n_flux_faces_increment()

float libMesh::JumpErrorEstimator::coarse_n_flux_faces_increment ( )

protectedinherited

A utility function to correctly increase n_flux_faces for the coarse element.

Definition at line 592 of file jump_error_estimator.C.

References libMesh::JumpErrorEstimator::coarse_context, dim, and libMesh::JumpErrorEstimator::fine_context.

Referenced by libMesh::JumpErrorEstimator::estimate_error().

 {
   // Keep track of the number of internal flux sides found on each
   // element
   unsigned short dim = coarse_context->get_elem().dim();
 
   const unsigned int divisor =
     1 << (dim-1)*(fine_context->get_elem().level() -
                   coarse_context->get_elem().level());
 
   // With a difference of n levels between fine and coarse elements,
   // we compute a fractional flux face for the coarse element by adding:
   // 1/2^n in 2D
   // 1/4^n in 3D
   // each time.  This code will get hit 2^n times in 2D and 4^n
   // times in 3D so that the final flux face count for the coarse
   // element will be an integer value.
 
   return 1.0f / static_cast<float>(divisor);
 }

◆ estimate_error()

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

overridevirtualinherited

This function uses the derived class's jump error estimate formula to estimate the error on each cell.

The estimated error is output in the vector error_per_cell

Conventions for assigning the direction of the normal:

e & f are global element ids

Case (1.) Elements are at the same level, e<f Compute the flux jump on the face and add it as a contribution to error_per_cell[e] and error_per_cell[f]

| | |

	f
e	—> n

Case (2.) The neighbor is at a higher level. Compute the flux jump on e's face and add it as a contribution to error_per_cell[e] and error_per_cell[f]

| | | | | | e |—> n | | | | | |--------—| f |

Implements libMesh::ErrorEstimator.

Definition at line 56 of file jump_error_estimator.C.

Referenced by assemble_and_solve(), main(), and SlitMeshRefinedSystemTest::testSystem().

 {
   LOG_SCOPE("estimate_error()", "JumpErrorEstimator");
 
   // This parameter is not used when !LIBMESH_ENABLE_AMR.
   libmesh_ignore(estimate_parent_error);
 
   // The current mesh
   const MeshBase & mesh = system.get_mesh();
 
   // The number of variables in the system
   const unsigned int n_vars = system.n_vars();
 
   // The DofMap for this system
 #ifdef LIBMESH_ENABLE_AMR
   const DofMap & dof_map = system.get_dof_map();
 #endif
 
   // Resize the error_per_cell vector according to the
   // maximum element ID because we will be indexing it with IDs.
   // Initialize to 0.
   error_per_cell.resize (mesh.max_elem_id());
   std::fill (error_per_cell.begin(), error_per_cell.end(), 0.);
 
   // Declare a vector of floats which is as long as
   // error_per_cell above, and fill with zeros.  This vector will be
   // used to keep track of the number of edges (faces) on each active
   // element which are either:
   // 1) an internal edge
   // 2) an edge on a Neumann boundary for which a boundary condition
   //    function has been specified.
   // The error estimator can be scaled by the number of flux edges (faces)
   // which the element actually has to obtain a more uniform measure
   // of the error.  Use floats instead of ints since in case 2 (above)
   // f gets 1/2 of a flux face contribution from each of his
   // neighbors
   std::vector<float> n_flux_faces;
   if (scale_by_n_flux_faces)
     n_flux_faces.resize(error_per_cell.size(), 0);
 
   // Prepare current_local_solution to localize a non-standard
   // solution vector if necessary
   if (solution_vector && solution_vector != system.solution.get())
     {
       NumericVector<Number> * newsol =
         const_cast<NumericVector<Number> *>(solution_vector);
       System & sys = const_cast<System &>(system);
       newsol->swap(*sys.solution);
       sys.update();
     }
 
   // We don't use full elem_jacobian or subjacobians here.
   fine_context = std::make_unique<FEMContext>
     (system, nullptr, /* allocate_local_matrices = */ false);
   coarse_context = std::make_unique<FEMContext>
     (system, nullptr, /* allocate_local_matrices = */ false);
 
   // Don't overintegrate - we're evaluating differences of FE values,
   // not products of them.
   if (this->use_unweighted_quadrature_rules)
     fine_context->use_unweighted_quadrature_rules(system.extra_quadrature_order);
 
   // Loop over all the variables we've been requested to find jumps in, to
   // pre-request
   for (var=0; var<n_vars; var++)
     {
       // Skip variables which aren't part of our norm,
       // as well as SCALAR variables, which have no jumps
       if (error_norm.weight(var) == 0.0 ||
           system.variable_type(var).family == SCALAR)
         continue;
 
       // FIXME: Need to generalize this to vector-valued elements. [PB]
       FEBase * side_fe = nullptr;
 
       const std::set<unsigned char> & elem_dims =
         fine_context->elem_dimensions();
 
       for (const auto & dim : elem_dims)
         {
           fine_context->get_side_fe( var, side_fe, dim );
 
           side_fe->get_xyz();
         }
     }
 
   this->init_context(*fine_context);
   this->init_context(*coarse_context);
 
   // If we're integrating jumps across mesh slits, then we'll need a
   // point locator to find slits, and we'll need to integrate point by
   // point on sides.
   std::unique_ptr<PointLocatorBase> point_locator;
   std::unique_ptr<const Elem> side_ptr;
 
   if (integrate_slits)
     point_locator = mesh.sub_point_locator();
 
   // Iterate over all the active elements in the mesh
   // that live on this processor.
   for (const auto & e : mesh.active_local_element_ptr_range())
     {
       const dof_id_type e_id = e->id();
 
 #ifdef LIBMESH_ENABLE_AMR
 
       if (e->infinite())
       {
          libmesh_warning("Warning: Jumps on the border of infinite elements are ignored."
                          << std::endl);
          continue;
       }
 
       // See if the parent of element e has been examined yet;
       // if not, we may want to compute the estimator on it
       const Elem * parent = e->parent();
 
       // We only can compute and only need to compute on
       // parents with all active children
       bool compute_on_parent = true;
       if (!parent || !estimate_parent_error)
         compute_on_parent = false;
       else
         for (auto & child : parent->child_ref_range())
           if (!child.active())
             compute_on_parent = false;
 
       if (compute_on_parent &&
           !error_per_cell[parent->id()])
         {
           // Compute a projection onto the parent
           DenseVector<Number> Uparent;
           FEBase::coarsened_dof_values
             (*(system.solution), dof_map, parent, Uparent, false);
 
           // Loop over the neighbors of the parent
           for (auto n_p : parent->side_index_range())
             {
               if (parent->neighbor_ptr(n_p) != nullptr) // parent has a neighbor here
                 {
                   // Find the active neighbors in this direction
                   std::vector<const Elem *> active_neighbors;
                   parent->neighbor_ptr(n_p)->
                     active_family_tree_by_neighbor(active_neighbors,
                                                    parent);
                   // Compute the flux to each active neighbor
                   for (std::size_t a=0,
                         n_active_neighbors = active_neighbors.size();
                        a != n_active_neighbors; ++a)
                     {
                       const Elem * f = active_neighbors[a];
 
                       if (f ->infinite()) // don't take infinite elements into account
                          continue;
 
                       // FIXME - what about when f->level <
                       // parent->level()??
                       if (f->level() >= parent->level())
                         {
                           fine_context->pre_fe_reinit(system, f);
                           coarse_context->pre_fe_reinit(system, parent);
                           libmesh_assert_equal_to
                             (coarse_context->get_elem_solution().size(),
                              Uparent.size());
                           coarse_context->get_elem_solution() = Uparent;
 
                           this->reinit_sides();
 
                           // Loop over all significant variables in the system
                           for (var=0; var<n_vars; var++)
                             if (error_norm.weight(var) != 0.0 &&
                                 system.variable_type(var).family != SCALAR)
                               {
                                 this->internal_side_integration();
 
                                 error_per_cell[fine_context->get_elem().id()] +=
                                   static_cast<ErrorVectorReal>(fine_error);
                                 error_per_cell[coarse_context->get_elem().id()] +=
                                   static_cast<ErrorVectorReal>(coarse_error);
                               }
 
                           // Keep track of the number of internal flux
                           // sides found on each element
                           if (scale_by_n_flux_faces)
                             {
                               n_flux_faces[fine_context->get_elem().id()]++;
                               n_flux_faces[coarse_context->get_elem().id()] +=
                                 this->coarse_n_flux_faces_increment();
                             }
                         }
                     }
                 }
               else if (integrate_boundary_sides)
                 {
                   fine_context->pre_fe_reinit(system, parent);
                   libmesh_assert_equal_to
                     (fine_context->get_elem_solution().size(),
                      Uparent.size());
                   fine_context->get_elem_solution() = Uparent;
                   fine_context->side = cast_int<unsigned char>(n_p);
                   fine_context->side_fe_reinit();
 
                   // If we find a boundary flux for any variable,
                   // let's just count it as a flux face for all
                   // variables.  Otherwise we'd need to keep track of
                   // a separate n_flux_faces and error_per_cell for
                   // every single var.
                   bool found_boundary_flux = false;
 
                   for (var=0; var<n_vars; var++)
                     if (error_norm.weight(var) != 0.0 &&
                         system.variable_type(var).family != SCALAR)
                       {
                         if (this->boundary_side_integration())
                           {
                             error_per_cell[fine_context->get_elem().id()] +=
                               static_cast<ErrorVectorReal>(fine_error);
                             found_boundary_flux = true;
                           }
                       }
 
                   if (scale_by_n_flux_faces && found_boundary_flux)
                     n_flux_faces[fine_context->get_elem().id()]++;
                 }
             }
         }
 #endif // #ifdef LIBMESH_ENABLE_AMR
 
       // If we do any more flux integration, e will be the fine element
       fine_context->pre_fe_reinit(system, e);
 
       // Loop over the neighbors of element e
       for (auto n_e : e->side_index_range())
         {
           if ((e->neighbor_ptr(n_e) != nullptr) ||
               integrate_boundary_sides)
             {
               fine_context->side = cast_int<unsigned char>(n_e);
               fine_context->side_fe_reinit();
             }
 
           // e is not on the boundary (infinite elements are treated as boundary)
           if (e->neighbor_ptr(n_e) != nullptr
               && !e->neighbor_ptr(n_e) ->infinite())
             {
 
               const Elem * f           = e->neighbor_ptr(n_e);
               const dof_id_type f_id = f->id();
 
               // Compute flux jumps if we are in case 1 or case 2.
               if ((f->active() && (f->level() == e->level()) && (e_id < f_id))
                   || (f->level() < e->level()))
                 {
                   // f is now the coarse element
                   coarse_context->pre_fe_reinit(system, f);
 
                   this->reinit_sides();
 
                   // Loop over all significant variables in the system
                   for (var=0; var<n_vars; var++)
                     if (error_norm.weight(var) != 0.0 &&
                         system.variable_type(var).family != SCALAR)
                       {
                         this->internal_side_integration();
 
                         error_per_cell[fine_context->get_elem().id()] +=
                           static_cast<ErrorVectorReal>(fine_error);
                         error_per_cell[coarse_context->get_elem().id()] +=
                           static_cast<ErrorVectorReal>(coarse_error);
                       }
 
                   // Keep track of the number of internal flux
                   // sides found on each element
                   if (scale_by_n_flux_faces)
                     {
                       n_flux_faces[fine_context->get_elem().id()]++;
                       n_flux_faces[coarse_context->get_elem().id()] +=
                         this->coarse_n_flux_faces_increment();
                     }
                 } // end if (case1 || case2)
             } // if (e->neighbor(n_e) != nullptr)
 
           // e might not have a neighbor_ptr, but might still have
           // another element sharing its side.  This can happen in a
           // mesh where solution continuity is maintained via nodal
           // constraint rows.
           else if (integrate_slits)
             {
               side_ptr = e->build_side_ptr(n_e);
               std::set<const Elem *> candidate_elements;
               (*point_locator)(side_ptr->vertex_average(), candidate_elements);
 
               // We should have at least found ourselves...
               libmesh_assert(candidate_elements.count(e));
 
               // If we only found ourselves, this probably isn't a
               // slit; we don't yet support meshes so non-conforming
               // as to have overlap of part of an element side without
               // overlap of its center.
               if (candidate_elements.size() < 2)
                 continue;
 
               FEType hardest_fe_type = fine_context->find_hardest_fe_type();
 
               auto dim = e->dim();
 
               auto side_qrule =
                 hardest_fe_type.unweighted_quadrature_rule
                 (dim-1, system.extra_quadrature_order);
               auto side_fe = FEAbstract::build(dim, hardest_fe_type);
               side_fe->attach_quadrature_rule(side_qrule.get());
               const std::vector<Point> & qface_point = side_fe->get_xyz();
               side_fe->reinit(e, n_e);
 
               for (auto qp : make_range(side_qrule->n_points()))
                 {
                   const Point p = qface_point[qp];
                   const std::vector<Point> qp_pointvec(1, p);
                   std::set<const Elem *> side_elements;
                   (*point_locator)(side_ptr->vertex_average(), side_elements);
 
                   // If we have multiple neighbors meeting here we'll just
                   // take weighted jumps from all of them.
                   //
                   // We'll also do integrations from both sides of slits,
                   // rather than try to figure out a disambiguation rule
                   // that makes sense for non-conforming slits in general.
                   // This means we want an extra factor of 0.5 on the
                   // integrals to compensate for doubling them.
                   const std::size_t n_neighbors = side_elements.size() - 1;
                   const Real neighbor_frac = Real(1)/n_neighbors;
 
                   const std::vector<Real>
                     qp_weightvec(1, neighbor_frac * side_qrule->w(qp));
 
                   for (const Elem * f : side_elements)
                     {
                       if (f == e)
                         continue;
 
                       coarse_context->pre_fe_reinit(system, f);
                       fine_context->pre_fe_reinit(system, e);
                       std::vector<Point> qp_coarse, qp_fine;
                       for (unsigned int v=0; v<n_vars; v++)
                         if (error_norm.weight(v) != 0.0 &&
                             fine_context->get_system().variable_type(v).family != SCALAR)
                           {
                             FEBase * coarse_fe = coarse_context->get_side_fe(v, dim);
                             if (qp_coarse.empty())
                               FEMap::inverse_map (dim, f, qp_pointvec, qp_coarse);
                             coarse_fe->reinit(f, &qp_coarse, &qp_weightvec);
                             FEBase * fine_fe = fine_context->get_side_fe(v, dim);
                             if (qp_fine.empty())
                               FEMap::inverse_map (dim, e, qp_pointvec, qp_fine);
                             fine_fe->reinit(e, &qp_fine, &qp_weightvec);
                           }
 
                       // Loop over all significant variables in the system
                       for (var=0; var<n_vars; var++)
                         if (error_norm.weight(var) != 0.0 &&
                             system.variable_type(var).family != SCALAR)
                           {
                             this->internal_side_integration();
 
                             error_per_cell[fine_context->get_elem().id()] +=
                               static_cast<ErrorVectorReal>(fine_error);
                             error_per_cell[coarse_context->get_elem().id()] +=
                               static_cast<ErrorVectorReal>(coarse_error);
                           }
                     }
                 }
             }
 
           // Otherwise, e is on the boundary.  If it happens to
           // be on a Dirichlet boundary, we need not do anything.
           // On the other hand, if e is on a Neumann (flux) boundary
           // with grad(u).n = g, we need to compute the additional residual
           // (h * \int |g - grad(u_h).n|^2 dS)^(1/2).
           // We can only do this with some knowledge of the boundary
           // conditions, i.e. the user must have attached an appropriate
           // BC function.
           else if (integrate_boundary_sides)
             {
               if (integrate_slits)
                 libmesh_not_implemented();
 
               bool found_boundary_flux = false;
 
               for (var=0; var<n_vars; var++)
                 if (error_norm.weight(var) != 0.0 &&
                     system.variable_type(var).family != SCALAR)
                   if (this->boundary_side_integration())
                     {
                       error_per_cell[fine_context->get_elem().id()] +=
                         static_cast<ErrorVectorReal>(fine_error);
                       found_boundary_flux = true;
                     }
 
               if (scale_by_n_flux_faces && found_boundary_flux)
                 n_flux_faces[fine_context->get_elem().id()]++;
             } // end if (e->neighbor_ptr(n_e) == nullptr)
         } // end loop over neighbors
     } // End loop over active local elements
 
 
   // 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.
 
   // First sum the vector of estimated error values
   this->reduce_error(error_per_cell, system.comm());
 
   // Compute the square-root of each component.
   for (auto i : index_range(error_per_cell))
     if (error_per_cell[i] != 0.)
       error_per_cell[i] = std::sqrt(error_per_cell[i]);
 
 
   if (this->scale_by_n_flux_faces)
     {
       // Sum the vector of flux face counts
       this->reduce_error(n_flux_faces, system.comm());
 
       // Sanity check: Make sure the number of flux faces is
       // always an integer value
 #ifdef DEBUG
       for (const auto & val : n_flux_faces)
         libmesh_assert_equal_to (val, static_cast<float>(static_cast<unsigned int>(val)));
 #endif
 
       // Scale the error by the number of flux faces for each element
       for (auto i : index_range(n_flux_faces))
         {
           if (n_flux_faces[i] == 0.0) // inactive or non-local element
             continue;
 
           error_per_cell[i] /= static_cast<ErrorVectorReal>(n_flux_faces[i]);
         }
     }
 
   // If we used a non-standard solution before, now is the time to fix
   // the current_local_solution
   if (solution_vector && solution_vector != system.solution.get())
     {
       NumericVector<Number> * newsol =
         const_cast<NumericVector<Number> *>(solution_vector);
       System & sys = const_cast<System &>(system);
       newsol->swap(*sys.solution);
       sys.update();
     }
 }

◆ estimate_errors() [1/2]

void libMesh::ErrorEstimator::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`
	)

virtualinherited

This virtual function can be redefined in derived classes, but by default computes the sum of the error_per_cell for each system in the equation_systems.

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 in libMesh::UniformRefinementEstimator.

Definition at line 47 of file error_estimator.C.

References libMesh::ErrorEstimator::error_norm, libMesh::ErrorEstimator::estimate_error(), libMesh::EquationSystems::get_system(), libMesh::index_range(), libMesh::make_range(), and libMesh::EquationSystems::n_systems().

 {
   SystemNorm old_error_norm = this->error_norm;
 
   // Sum the error values from each system
   for (auto s : make_range(equation_systems.n_systems()))
     {
       ErrorVector system_error_per_cell;
       const System & sys = equation_systems.get_system(s);
       if (const auto it = error_norms.find(&sys);
           it == error_norms.end())
         this->error_norm = old_error_norm;
       else
         this->error_norm = it->second;
 
       const NumericVector<Number> * solution_vector = nullptr;
       if (solution_vectors &&
           solution_vectors->find(&sys) != solution_vectors->end())
         solution_vector = solution_vectors->find(&sys)->second;
 
       this->estimate_error(sys, system_error_per_cell,
                            solution_vector, estimate_parent_error);
 
       if (s)
         {
           libmesh_assert_equal_to (error_per_cell.size(), system_error_per_cell.size());
           for (auto i : index_range(error_per_cell))
             error_per_cell[i] += system_error_per_cell[i];
         }
       else
         error_per_cell = system_error_per_cell;
     }
 
   // Restore our old state before returning
   this->error_norm = old_error_norm;
 }

◆ estimate_errors() [2/2]

void libMesh::ErrorEstimator::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`
	)

virtualinherited

This virtual function can be redefined in derived classes, but by default it calls estimate_error repeatedly to calculate the requested error vectors.

FIXME: This is a default implementation - derived classes should reimplement it for efficiency.

Currently this function ignores the error_norm.weight() values because it calculates each variable's 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 in libMesh::UniformRefinementEstimator.

Definition at line 94 of file error_estimator.C.

References libMesh::ErrorEstimator::error_norm, libMesh::ErrorEstimator::estimate_error(), libMesh::EquationSystems::get_system(), libMesh::make_range(), libMesh::EquationSystems::n_systems(), n_vars, libMesh::System::n_vars(), and libMesh::SystemNorm::type().

 {
   SystemNorm old_error_norm = this->error_norm;
 
   // Find the requested error values from each system
   for (auto s : make_range(equation_systems.n_systems()))
     {
       const System & sys = equation_systems.get_system(s);
 
       unsigned int n_vars = sys.n_vars();
 
       for (unsigned int v = 0; v != n_vars; ++v)
         {
           // Only fill in ErrorVectors the user asks for
           if (!errors_per_cell.count(std::make_pair(&sys, v)))
             continue;
 
           // Calculate error in only one variable
           std::vector<Real> weights(n_vars, 0.0);
           weights[v] = 1.0;
           this->error_norm =
             SystemNorm(std::vector<FEMNormType>(n_vars, old_error_norm.type(v)),
                        weights);
 
           const NumericVector<Number> * solution_vector = nullptr;
           if (solution_vectors &&
               solution_vectors->find(&sys) != solution_vectors->end())
             solution_vector = solution_vectors->find(&sys)->second;
 
           this->estimate_error
             (sys, *errors_per_cell[std::make_pair(&sys, v)],
              solution_vector, estimate_parent_error);
         }
     }
 
   // Restore our old state before returning
   this->error_norm = old_error_norm;
 }

◆ init_context()

void libMesh::DiscontinuityMeasure::init_context ( FEMContext & c )

overrideprotectedvirtual

An initialization function, for requesting specific data from the FE objects.

Reimplemented from libMesh::JumpErrorEstimator.

Definition at line 59 of file discontinuity_measure.C.

References libMesh::JumpErrorEstimator::coarse_context, dim, libMesh::FEMContext::elem_dimensions(), libMesh::ErrorEstimator::error_norm, libMesh::FEAbstract::get_JxW(), libMesh::FEGenericBase< OutputType >::get_phi(), libMesh::FEMContext::get_side_fe(), libMesh::DiffContext::n_vars(), n_vars, and libMesh::SystemNorm::weight().

 {
   const unsigned int n_vars = c.n_vars();
   for (unsigned int v=0; v<n_vars; v++)
     {
       // Possibly skip this variable
       if (error_norm.weight(v) == 0.0) continue;
 
       // FIXME: Need to generalize this to vector-valued elements. [PB]
       FEBase * side_fe = nullptr;
 
       const std::set<unsigned char> & elem_dims =
         c.elem_dimensions();
 
       for (const auto & dim : elem_dims)
         {
           c.get_side_fe( v, side_fe, dim );
 
           // We'll need values and mapping Jacobians on both sides for
           // discontinuity computation
           side_fe->get_phi();
 
           // But we only need mapping Jacobians from one side
           if (&c != coarse_context.get())
             side_fe->get_JxW();
         }
     }
 }

◆ internal_side_integration()

void libMesh::DiscontinuityMeasure::internal_side_integration ( )

overrideprotectedvirtual

The function which calculates a normal derivative jump based error term on an internal side.

Implements libMesh::JumpErrorEstimator.

Definition at line 91 of file discontinuity_measure.C.

References libMesh::JumpErrorEstimator::coarse_context, libMesh::JumpErrorEstimator::coarse_error, libMesh::Elem::dim(), libMesh::ErrorEstimator::error_norm, libMesh::JumpErrorEstimator::fine_context, libMesh::JumpErrorEstimator::fine_error, libMesh::FEAbstract::get_JxW(), libMesh::Elem::hmax(), libMesh::FEAbstract::n_quadrature_points(), libMesh::TensorTools::norm_sq(), libMesh::Real, libMesh::JumpErrorEstimator::var, and libMesh::SystemNorm::weight().

 {
   const Elem & coarse_elem = coarse_context->get_elem();
   const Elem & fine_elem = fine_context->get_elem();
 
   FEBase * fe_fine = nullptr;
   fine_context->get_side_fe( var, fe_fine, fine_elem.dim() );
 
   FEBase * fe_coarse = nullptr;
   coarse_context->get_side_fe( var, fe_coarse, fine_elem.dim() );
 
   Real error = 1.e-30;
   unsigned int n_qp = fe_fine->n_quadrature_points();
 
   const std::vector<Real> & JxW_face = fe_fine->get_JxW();
 
   for (unsigned int qp=0; qp != n_qp; ++qp)
     {
       // Calculate solution values on fine and coarse elements
       // at this quadrature point
       Number
         u_fine   = fine_context->side_value(var, qp),
         u_coarse = coarse_context->side_value(var, qp);
 
       // Find the jump in the value
       // at this quadrature point
       const Number jump = u_fine - u_coarse;
       const Real jump2 = TensorTools::norm_sq(jump);
       // Accumulate the jump integral
       error += JxW_face[qp] * jump2;
     }
 
   // Add the h-weighted jump integral to each error term
   fine_error =
     error * fine_elem.hmax() * error_norm.weight(var);
   coarse_error =
     error * coarse_elem.hmax() * error_norm.weight(var);
 }

◆ operator=() [1/2]

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

delete

◆ operator=() [2/2]

DiscontinuityMeasure& libMesh::DiscontinuityMeasure::operator= ( DiscontinuityMeasure && )

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 TIMPI::Communicator::sum().

Referenced by libMesh::UniformRefinementEstimator::_estimate_error(), libMesh::PatchRecoveryErrorEstimator::estimate_error(), libMesh::WeightedPatchRecoveryErrorEstimator::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);
 }

◆ reinit_sides()

void libMesh::JumpErrorEstimator::reinit_sides ( )

protectedinherited

A utility function to reinit the finite element data on elements sharing a side.

Definition at line 554 of file jump_error_estimator.C.

References libMesh::JumpErrorEstimator::coarse_context, dim, libMesh::ErrorEstimator::error_norm, libMesh::JumpErrorEstimator::fine_context, libMesh::FEMap::inverse_map(), n_vars, libMesh::FEAbstract::reinit(), libMesh::SCALAR, and libMesh::SystemNorm::weight().

Referenced by libMesh::JumpErrorEstimator::estimate_error().

 {
   fine_context->side_fe_reinit();
 
   unsigned short dim = fine_context->get_elem().dim();
   libmesh_assert_equal_to(dim, coarse_context->get_elem().dim());
 
   FEBase * fe_fine = nullptr;
   fine_context->get_side_fe( 0, fe_fine, dim );
 
   // Get the physical locations of the fine element quadrature points
   std::vector<Point> qface_point = fe_fine->get_xyz();
 
   // Find the master quadrature point locations on the coarse element
   FEBase * fe_coarse = nullptr;
   coarse_context->get_side_fe( 0, fe_coarse, dim );
 
   std::vector<Point> qp_coarse;
 
   FEMap::inverse_map (coarse_context->get_elem().dim(),
                       &coarse_context->get_elem(), qface_point,
                       qp_coarse);
 
   // The number of variables in the system
   const unsigned int n_vars = fine_context->n_vars();
 
   // Calculate all coarse element shape functions at those locations
   for (unsigned int v=0; v<n_vars; v++)
     if (error_norm.weight(v) != 0.0 &&
         fine_context->get_system().variable_type(v).family != SCALAR)
       {
         coarse_context->get_side_fe( v, fe_coarse, dim );
         fe_coarse->reinit (&coarse_context->get_elem(), &qp_coarse);
       }
 }

◆ type()

ErrorEstimatorType libMesh::DiscontinuityMeasure::type ( ) const

overridevirtual

Returns: The type for the ErrorEstimator subclass.

Implements libMesh::ErrorEstimator.

Definition at line 51 of file discontinuity_measure.C.

References libMesh::DISCONTINUITY_MEASURE.

 {
   return DISCONTINUITY_MEASURE;
 }

Member Data Documentation

◆ _bc_function

std::pair<bool,Real>(* libMesh::DiscontinuityMeasure::_bc_function) (const System &system, const Point &p, const std::string &var_name)

protected

Pointer to function that provides BC information.

Definition at line 108 of file discontinuity_measure.h.

Referenced by attach_essential_bc_function(), and boundary_side_integration().

◆ coarse_context

std::unique_ptr<FEMContext> libMesh::JumpErrorEstimator::coarse_context

protectedinherited

Definition at line 175 of file jump_error_estimator.h.

Referenced by libMesh::JumpErrorEstimator::coarse_n_flux_faces_increment(), libMesh::JumpErrorEstimator::estimate_error(), init_context(), libMesh::KellyErrorEstimator::init_context(), libMesh::LaplacianErrorEstimator::internal_side_integration(), internal_side_integration(), libMesh::KellyErrorEstimator::internal_side_integration(), and libMesh::JumpErrorEstimator::reinit_sides().

◆ coarse_error

Real libMesh::JumpErrorEstimator::coarse_error

protectedinherited

Definition at line 180 of file jump_error_estimator.h.

Referenced by libMesh::JumpErrorEstimator::estimate_error(), libMesh::LaplacianErrorEstimator::internal_side_integration(), internal_side_integration(), and libMesh::KellyErrorEstimator::internal_side_integration().

◆ 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 158 of file error_estimator.h.

Referenced by libMesh::UniformRefinementEstimator::_estimate_error(), libMesh::AdjointRefinementEstimator::AdjointRefinementEstimator(), boundary_side_integration(), libMesh::KellyErrorEstimator::boundary_side_integration(), 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(), init_context(), libMesh::KellyErrorEstimator::init_context(), libMesh::LaplacianErrorEstimator::internal_side_integration(), 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 libMesh::UniformRefinementEstimator::UniformRefinementEstimator().

◆ fine_context

std::unique_ptr<FEMContext> libMesh::JumpErrorEstimator::fine_context

protectedinherited

Context objects for integrating on the fine and coarse elements sharing a face.

Definition at line 175 of file jump_error_estimator.h.

Referenced by boundary_side_integration(), libMesh::KellyErrorEstimator::boundary_side_integration(), libMesh::JumpErrorEstimator::coarse_n_flux_faces_increment(), libMesh::JumpErrorEstimator::estimate_error(), libMesh::LaplacianErrorEstimator::internal_side_integration(), internal_side_integration(), libMesh::KellyErrorEstimator::internal_side_integration(), and libMesh::JumpErrorEstimator::reinit_sides().

◆ fine_error

Real libMesh::JumpErrorEstimator::fine_error

protectedinherited

The fine and coarse error values to be set by each side_integration();.

Definition at line 180 of file jump_error_estimator.h.

Referenced by boundary_side_integration(), libMesh::KellyErrorEstimator::boundary_side_integration(), libMesh::JumpErrorEstimator::estimate_error(), libMesh::LaplacianErrorEstimator::internal_side_integration(), internal_side_integration(), and libMesh::KellyErrorEstimator::internal_side_integration().

◆ integrate_boundary_sides

bool libMesh::JumpErrorEstimator::integrate_boundary_sides

protectedinherited

A boolean flag, by default false, to be set to true if integrations with boundary_side_integration() should be performed.

Definition at line 169 of file jump_error_estimator.h.

Referenced by attach_essential_bc_function(), libMesh::KellyErrorEstimator::attach_flux_bc_function(), and libMesh::JumpErrorEstimator::estimate_error().

◆ integrate_slits

bool libMesh::JumpErrorEstimator::integrate_slits

inherited

A boolean flag, by default false, to be set to true if integrations should be performed on "slits" where two elements' faces overlap even if those elements are not connected by neighbor links.

This may only be useful for flex-IGA meshes, where highly-nonconforming meshes are given pseudo-conforming solutions via nodal constraints.

Note that, to safely use this option in parallel, it is also necessary to expand the default algebraic ghosting requirements to include elements on the opposite sides of slits from local elements.

Definition at line 130 of file jump_error_estimator.h.

Referenced by libMesh::JumpErrorEstimator::estimate_error(), and SlitMeshRefinedSystemTest::testSystem().

◆ scale_by_n_flux_faces

bool libMesh::JumpErrorEstimator::scale_by_n_flux_faces

inherited

This boolean flag allows you to scale the error indicator result for each element by the number of "flux faces" the element actually has.

This tends to weight more evenly cells which are on the boundaries and thus have fewer contributions to their flux. The value is initialized to false, simply set it to true if you want to use the feature.

Definition at line 100 of file jump_error_estimator.h.

Referenced by libMesh::JumpErrorEstimator::estimate_error().

◆ use_unweighted_quadrature_rules

bool libMesh::JumpErrorEstimator::use_unweighted_quadrature_rules

inherited

This boolean flag allows you to use "unweighted" quadrature rules (sized to exactly integrate unweighted shape functions in master element space) rather than "default" quadrature rules (sized to exactly integrate polynomials of one higher degree than mass matrix terms).

The results with the former, lower-order rules will be somewhat less accurate in many cases but will be much cheaper to compute.

The value is initialized to false, simply set it to true if you want to use the feature.

Definition at line 114 of file jump_error_estimator.h.

Referenced by libMesh::JumpErrorEstimator::estimate_error(), and main().

◆ var

unsigned int libMesh::JumpErrorEstimator::var

protectedinherited

The variable number currently being evaluated.

Definition at line 185 of file jump_error_estimator.h.

Referenced by boundary_side_integration(), libMesh::KellyErrorEstimator::boundary_side_integration(), libMesh::JumpErrorEstimator::estimate_error(), libMesh::LaplacianErrorEstimator::internal_side_integration(), internal_side_integration(), and libMesh::KellyErrorEstimator::internal_side_integration().

The documentation for this class was generated from the following files:

include/error_estimation/discontinuity_measure.h
src/error_estimation/discontinuity_measure.C

Public Types

Public Member Functions

Public Attributes

Protected Member Functions

Protected Attributes

Detailed Description

Member Typedef Documentation

◆ ErrorMap

Constructor & Destructor Documentation

◆ DiscontinuityMeasure() [1/3]

◆ DiscontinuityMeasure() [2/3]

◆ DiscontinuityMeasure() [3/3]

◆ ~DiscontinuityMeasure()

Member Function Documentation

◆ attach_essential_bc_function()

◆ boundary_side_integration()

◆ coarse_n_flux_faces_increment()

◆ estimate_error()

◆ estimate_errors() [1/2]

◆ estimate_errors() [2/2]

◆ init_context()

◆ internal_side_integration()

◆ operator=() [1/2]

◆ operator=() [2/2]

◆ reduce_error()

◆ reinit_sides()

◆ type()

Member Data Documentation

◆ _bc_function

◆ coarse_context

◆ coarse_error

◆ error_norm

◆ fine_context

◆ fine_error

◆ integrate_boundary_sides

◆ integrate_slits

◆ scale_by_n_flux_faces

◆ use_unweighted_quadrature_rules

◆ var