Reference documentation for deal.II version 9.4.1
\(\newcommand{\dealvcentcolon}{\mathrel{\mathop{:}}}\) \(\newcommand{\dealcoloneq}{\dealvcentcolon\mathrel{\mkern-1.2mu}=}\) \(\newcommand{\jump}[1]{\left[\!\left[ #1 \right]\!\right]}\) \(\newcommand{\average}[1]{\left\{\!\left\{ #1 \right\}\!\right\}}\)
Loading...
Searching...
No Matches
fe_q_bubbles.cc
Go to the documentation of this file.
1// ---------------------------------------------------------------------
2//
3// Copyright (C) 2012 - 2021 by the deal.II authors
4//
5// This file is part of the deal.II library.
6//
7// The deal.II library is free software; you can use it, redistribute
8// it, and/or modify it under the terms of the GNU Lesser General
9// Public License as published by the Free Software Foundation; either
10// version 2.1 of the License, or (at your option) any later version.
11// The full text of the license can be found in the file LICENSE.md at
12// the top level directory of deal.II.
13//
14// ---------------------------------------------------------------------
15
16
20
23
24#include <deal.II/fe/fe_dgq.h>
27#include <deal.II/fe/fe_tools.h>
30
32#include <deal.II/grid/tria.h>
33
34#include <memory>
35#include <sstream>
36#include <vector>
37
39
40
41namespace internal
42{
43 namespace FE_Q_Bubbles
44 {
45 namespace
46 {
47 template <int dim, int spacedim>
48 inline void
49 compute_embedding_matrices(
50 const ::FE_Q_Bubbles<dim, spacedim> & fe,
51 std::vector<std::vector<FullMatrix<double>>> &matrices,
52 const bool isotropic_only)
53 {
54 const unsigned int dpc = fe.n_dofs_per_cell();
55 const unsigned int degree = fe.degree;
56
57 // Initialize quadrature formula on fine cells
58 std::unique_ptr<Quadrature<dim>> q_fine;
59 Quadrature<1> q_dummy(std::vector<Point<1>>(1),
60 std::vector<double>(1, 1.));
61 switch (dim)
62 {
63 case 1:
64 if (spacedim == 1)
65 q_fine = std::make_unique<QGauss<dim>>(degree + 1);
66 else if (spacedim == 2)
67 q_fine =
68 std::make_unique<QAnisotropic<dim>>(QGauss<1>(degree + 1),
69 q_dummy);
70 else
71 q_fine =
72 std::make_unique<QAnisotropic<dim>>(QGauss<1>(degree + 1),
73 q_dummy,
74 q_dummy);
75 break;
76 case 2:
77 if (spacedim == 2)
78 q_fine = std::make_unique<QGauss<dim>>(degree + 1);
79 else
80 q_fine =
81 std::make_unique<QAnisotropic<dim>>(QGauss<1>(degree + 1),
82 QGauss<1>(degree + 1),
83 q_dummy);
84 break;
85 case 3:
86 q_fine = std::make_unique<QGauss<dim>>(degree + 1);
87 break;
88 default:
89 Assert(false, ExcInternalError());
90 }
91
92 Assert(q_fine.get() != nullptr, ExcInternalError());
93 const unsigned int nq = q_fine->size();
94
95 // loop over all possible refinement cases
96 unsigned int ref_case = (isotropic_only) ?
99 for (; ref_case <= RefinementCase<dim>::isotropic_refinement;
100 ++ref_case)
101 {
102 const unsigned int nc =
104
105 for (unsigned int i = 0; i < nc; ++i)
106 {
107 Assert(matrices[ref_case - 1][i].n() == dpc,
108 ExcDimensionMismatch(matrices[ref_case - 1][i].n(),
109 dpc));
110 Assert(matrices[ref_case - 1][i].m() == dpc,
111 ExcDimensionMismatch(matrices[ref_case - 1][i].m(),
112 dpc));
113 }
114
115 // create a respective refinement on the triangulation
118 tr.begin_active()->set_refine_flag(RefinementCase<dim>(ref_case));
120
122 dh.distribute_dofs(fe);
123
125 fe,
126 *q_fine,
130
131 const unsigned int n_dofs = dh.n_dofs();
132
133 FullMatrix<double> fine_mass(n_dofs);
134 FullMatrix<double> coarse_rhs_matrix(n_dofs, dpc);
135
136 std::vector<std::vector<types::global_dof_index>> child_ldi(
137 nc, std::vector<types::global_dof_index>(fe.n_dofs_per_cell()));
138
139 // now create the mass matrix and all the right_hand sides
140 unsigned int child_no = 0;
141 typename ::DoFHandler<dim>::active_cell_iterator cell =
142 dh.begin_active();
143 for (; cell != dh.end(); ++cell, ++child_no)
144 {
145 fine.reinit(cell);
146 cell->get_dof_indices(child_ldi[child_no]);
147
148 for (unsigned int q = 0; q < nq; ++q)
149 for (unsigned int i = 0; i < dpc; ++i)
150 for (unsigned int j = 0; j < dpc; ++j)
151 {
152 const unsigned int gdi = child_ldi[child_no][i];
153 const unsigned int gdj = child_ldi[child_no][j];
154 fine_mass(gdi, gdj) += fine.shape_value(i, q) *
155 fine.shape_value(j, q) *
156 fine.JxW(q);
157 Point<dim> quad_tmp;
158 for (unsigned int k = 0; k < dim; ++k)
159 quad_tmp(k) = fine.quadrature_point(q)(k);
160 coarse_rhs_matrix(gdi, j) +=
161 fine.shape_value(i, q) * fe.shape_value(j, quad_tmp) *
162 fine.JxW(q);
163 }
164 }
165
166 // now solve for all right-hand sides simultaneously
167 ::FullMatrix<double> solution(n_dofs, dpc);
168 fine_mass.gauss_jordan();
169 fine_mass.mmult(solution, coarse_rhs_matrix);
170
171 // and distribute to the fine cell matrices
172 for (unsigned int child_no = 0; child_no < nc; ++child_no)
173 for (unsigned int i = 0; i < dpc; ++i)
174 for (unsigned int j = 0; j < dpc; ++j)
175 {
176 const unsigned int gdi = child_ldi[child_no][i];
177 // remove small entries
178 if (std::fabs(solution(gdi, j)) > 1.e-12)
179 matrices[ref_case - 1][child_no](i, j) = solution(gdi, j);
180 }
181 }
182 }
183 } // namespace
184 } // namespace FE_Q_Bubbles
185} // namespace internal
186
187
188template <int dim, int spacedim>
190 : FE_Q_Base<dim, spacedim>(TensorProductPolynomialsBubbles<dim>(
191 Polynomials::generate_complete_Lagrange_basis(
192 QGaussLobatto<1>(q_degree + 1).get_points())),
193 FiniteElementData<dim>(get_dpo_vector(q_degree),
194 1,
195 q_degree + 1,
196 FiniteElementData<dim>::H1),
197 get_riaf_vector(q_degree))
198 , n_bubbles((q_degree <= 1) ? 1 : dim)
199{
200 Assert(q_degree > 0,
201 ExcMessage("This element can only be used for polynomial degrees "
202 "greater than zero"));
203
204 this->initialize(QGaussLobatto<1>(q_degree + 1).get_points());
205
206 // adjust unit support point for discontinuous node
207 Point<dim> point;
208 for (unsigned int d = 0; d < dim; ++d)
209 point[d] = 0.5;
210 for (unsigned int i = 0; i < n_bubbles; ++i)
211 this->unit_support_points.push_back(point);
213
215 if (dim == spacedim)
216 {
217 internal::FE_Q_Bubbles::compute_embedding_matrices(*this,
218 this->prolongation,
219 false);
220 // Fill restriction matrices with L2-projection
222 }
223}
224
225
226
227template <int dim, int spacedim>
229 : FE_Q_Base<dim, spacedim>(
231 Polynomials::generate_complete_Lagrange_basis(points.get_points())),
232 FiniteElementData<dim>(get_dpo_vector(points.size() - 1),
233 1,
234 points.size(),
235 FiniteElementData<dim>::H1),
236 get_riaf_vector(points.size() - 1))
237 , n_bubbles((points.size() - 1 <= 1) ? 1 : dim)
238{
239 Assert(points.size() > 1,
240 ExcMessage("This element can only be used for polynomial degrees "
241 "at least one"));
242
243 this->initialize(points.get_points());
244
245 // adjust unit support point for discontinuous node
246 Point<dim> point;
247 for (unsigned int d = 0; d < dim; ++d)
248 point[d] = 0.5;
249 for (unsigned int i = 0; i < n_bubbles; ++i)
250 this->unit_support_points.push_back(point);
252
254 if (dim == spacedim)
255 {
256 internal::FE_Q_Bubbles::compute_embedding_matrices(*this,
257 this->prolongation,
258 false);
259 // Fill restriction matrices with L2-projection
261 }
262}
263
264
265
266template <int dim, int spacedim>
267std::string
269{
270 // note that the FETools::get_fe_by_name function depends on the
271 // particular format of the string this function returns, so they have to be
272 // kept in synch
273
274 std::ostringstream namebuf;
275 bool type = true;
276 const unsigned int n_points = this->degree;
277 std::vector<double> points(n_points);
278 const unsigned int dofs_per_cell = this->n_dofs_per_cell();
279 const std::vector<Point<dim>> &unit_support_points =
280 this->unit_support_points;
281 unsigned int index = 0;
282
283 // Decode the support points in one coordinate direction.
284 for (unsigned int j = 0; j < dofs_per_cell; ++j)
285 {
286 if ((dim > 1) ? (unit_support_points[j](1) == 0 &&
287 ((dim > 2) ? unit_support_points[j](2) == 0 : true)) :
288 true)
289 {
290 if (index == 0)
291 points[index] = unit_support_points[j](0);
292 else if (index == 1)
293 points[n_points - 1] = unit_support_points[j](0);
294 else
295 points[index - 1] = unit_support_points[j](0);
296
297 index++;
298 }
299 }
300 // Do not consider the discontinuous node for dimension 1
301 Assert(index == n_points || (dim == 1 && index == n_points + n_bubbles),
303 "Could not decode support points in one coordinate direction."));
304
305 // Check whether the support points are equidistant.
306 for (unsigned int j = 0; j < n_points; ++j)
307 if (std::fabs(points[j] - static_cast<double>(j) / (this->degree - 1)) >
308 1e-15)
309 {
310 type = false;
311 break;
312 }
313
314 if (type == true)
315 {
316 if (this->degree > 3)
317 namebuf << "FE_Q_Bubbles<" << Utilities::dim_string(dim, spacedim)
318 << ">(QIterated(QTrapezoid()," << this->degree - 1 << "))";
319 else
320 namebuf << "FE_Q_Bubbles<" << Utilities::dim_string(dim, spacedim)
321 << ">(" << this->degree - 1 << ")";
322 }
323 else
324 {
325 // Check whether the support points come from QGaussLobatto.
326 const QGaussLobatto<1> points_gl(n_points);
327 type = true;
328 for (unsigned int j = 0; j < n_points; ++j)
329 if (points[j] != points_gl.point(j)(0))
330 {
331 type = false;
332 break;
333 }
334 if (type == true)
335 namebuf << "FE_Q_Bubbles<" << dim << ">(" << this->degree - 1 << ")";
336 else
337 namebuf << "FE_Q_Bubbles<" << dim << ">(QUnknownNodes(" << this->degree
338 << "))";
339 }
340 return namebuf.str();
341}
342
343
344
345template <int dim, int spacedim>
346std::unique_ptr<FiniteElement<dim, spacedim>>
348{
349 return std::make_unique<FE_Q_Bubbles<dim, spacedim>>(*this);
350}
351
352
353
354template <int dim, int spacedim>
355void
358 const std::vector<Vector<double>> &support_point_values,
359 std::vector<double> & nodal_values) const
360{
361 Assert(support_point_values.size() == this->unit_support_points.size(),
362 ExcDimensionMismatch(support_point_values.size(),
363 this->unit_support_points.size()));
364 Assert(nodal_values.size() == this->n_dofs_per_cell(),
365 ExcDimensionMismatch(nodal_values.size(), this->n_dofs_per_cell()));
366 Assert(support_point_values[0].size() == this->n_components(),
367 ExcDimensionMismatch(support_point_values[0].size(),
368 this->n_components()));
369
370 for (unsigned int i = 0; i < this->n_dofs_per_cell() - 1; ++i)
371 {
372 const std::pair<unsigned int, unsigned int> index =
373 this->system_to_component_index(i);
374 nodal_values[i] = support_point_values[i](index.first);
375 }
376
377 // We don't use the bubble functions for local interpolation
378 for (unsigned int i = 0; i < n_bubbles; ++i)
379 nodal_values[nodal_values.size() - i - 1] = 0.;
380}
381
382
383
384template <int dim, int spacedim>
385void
387 const FiniteElement<dim, spacedim> &x_source_fe,
388 FullMatrix<double> & interpolation_matrix) const
389{
390 // We don't know how to do this properly, yet.
391 // However, for SolutionTransfer to work we need to provide an implementation
392 // for the case that the x_source_fe is identical to this FE
393 using FEQBUBBLES = FE_Q_Bubbles<dim, spacedim>;
394
396 (x_source_fe.get_name().find("FE_Q_Bubbles<") == 0) ||
397 (dynamic_cast<const FEQBUBBLES *>(&x_source_fe) != nullptr),
399 Assert(interpolation_matrix.m() == this->n_dofs_per_cell(),
400 ExcDimensionMismatch(interpolation_matrix.m(),
401 this->n_dofs_per_cell()));
402 Assert(interpolation_matrix.n() == x_source_fe.n_dofs_per_cell(),
403 ExcDimensionMismatch(interpolation_matrix.m(),
404 x_source_fe.n_dofs_per_cell()));
405
406 // Provide a short cut in case we are just inquiring the identity
407 auto casted_fe = dynamic_cast<const FEQBUBBLES *>(&x_source_fe);
408 if (casted_fe != nullptr && casted_fe->degree == this->degree)
409 for (unsigned int i = 0; i < interpolation_matrix.m(); ++i)
410 interpolation_matrix.set(i, i, 1.);
411 // else we need to do more...
412 else
413 Assert(
414 false,
415 (typename FiniteElement<dim,
416 spacedim>::ExcInterpolationNotImplemented()));
417}
418
419
420
421template <int dim, int spacedim>
422std::vector<bool>
424{
425 const unsigned int n_cont_dofs = Utilities::fixed_power<dim>(q_deg + 1);
426 const unsigned int n_bubbles = (q_deg <= 1 ? 1 : dim);
427 return std::vector<bool>(n_cont_dofs + n_bubbles, true);
428}
429
430
431
432template <int dim, int spacedim>
433std::vector<unsigned int>
435{
436 std::vector<unsigned int> dpo(dim + 1, 1U);
437 for (unsigned int i = 1; i < dpo.size(); ++i)
438 dpo[i] = dpo[i - 1] * (q_deg - 1);
439
440 // Then add the bubble functions; they are all associated with the
441 // cell interior
442 dpo[dim] += (q_deg <= 1 ? 1 : dim);
443 return dpo;
444}
445
446
447
448template <int dim, int spacedim>
449bool
451 const unsigned int shape_index,
452 const unsigned int face_index) const
453{
454 // discontinuous functions have no support on faces
455 if (shape_index >= this->n_dofs_per_cell() - n_bubbles)
456 return false;
457 else
459 face_index);
460}
461
462
463
464template <int dim, int spacedim>
465const FullMatrix<double> &
467 const unsigned int child,
468 const RefinementCase<dim> &refinement_case) const
469{
470 AssertIndexRange(refinement_case,
474 "Prolongation matrices are only available for refined cells!"));
475 AssertIndexRange(child, GeometryInfo<dim>::n_children(refinement_case));
476
477 Assert(this->prolongation[refinement_case - 1][child].n() != 0,
478 ExcMessage("This prolongation matrix has not been computed yet!"));
479 // finally return the matrix
480 return this->prolongation[refinement_case - 1][child];
481}
482
483
484
485template <int dim, int spacedim>
486const FullMatrix<double> &
488 const unsigned int child,
489 const RefinementCase<dim> &refinement_case) const
490{
491 AssertIndexRange(refinement_case,
495 "Restriction matrices are only available for refined cells!"));
496 AssertIndexRange(child, GeometryInfo<dim>::n_children(refinement_case));
497
498 Assert(this->restriction[refinement_case - 1][child].n() != 0,
499 ExcMessage("This restriction matrix has not been computed yet!"));
500
501 // finally return the matrix
502 return this->restriction[refinement_case - 1][child];
503}
504
505
506
507template <int dim, int spacedim>
510 const FiniteElement<dim, spacedim> &fe_other,
511 const unsigned int codim) const
512{
513 Assert(codim <= dim, ExcImpossibleInDim(dim));
514
515 // vertex/line/face domination
516 // (if fe_other is derived from FE_DGQ)
517 // ------------------------------------
518 if (codim > 0)
519 if (dynamic_cast<const FE_DGQ<dim, spacedim> *>(&fe_other) != nullptr)
520 // there are no requirements between continuous and discontinuous elements
522
523 // vertex/line/face domination
524 // (if fe_other is not derived from FE_DGQ)
525 // & cell domination
526 // ----------------------------------------
527 if (const FE_Q_Bubbles<dim, spacedim> *fe_bubbles_other =
528 dynamic_cast<const FE_Q_Bubbles<dim, spacedim> *>(&fe_other))
529 {
530 if (this->degree < fe_bubbles_other->degree)
532 else if (this->degree == fe_bubbles_other->degree)
534 else
536 }
537 else if (const FE_Nothing<dim> *fe_nothing =
538 dynamic_cast<const FE_Nothing<dim> *>(&fe_other))
539 {
540 if (fe_nothing->is_dominating())
542 else
543 // the FE_Nothing has no degrees of freedom and it is typically used
544 // in a context where we don't require any continuity along the
545 // interface
547 }
548
549 Assert(false, ExcNotImplemented());
551}
552
553
554// explicit instantiations
555#include "fe_q_bubbles.inst"
556
cell_iterator end() const
void distribute_dofs(const FiniteElement< dim, spacedim > &fe)
active_cell_iterator begin_active(const unsigned int level=0) const
types::global_dof_index n_dofs() const
double JxW(const unsigned int quadrature_point) const
const Point< spacedim > & quadrature_point(const unsigned int q) const
const double & shape_value(const unsigned int function_no, const unsigned int point_no) const
void reinit(const TriaIterator< DoFCellAccessor< dim, spacedim, level_dof_access > > &cell)
Definition: fe_dgq.h:111
const unsigned int q_degree
Definition: fe_q_base.h:344
void initialize(const std::vector< Point< 1 > > &support_points_1d)
virtual bool has_support_on_face(const unsigned int shape_index, const unsigned int face_index) const override
virtual bool has_support_on_face(const unsigned int shape_index, const unsigned int face_index) const override
static std::vector< bool > get_riaf_vector(const unsigned int degree)
virtual void get_interpolation_matrix(const FiniteElement< dim, spacedim > &source, FullMatrix< double > &matrix) const override
virtual const FullMatrix< double > & get_restriction_matrix(const unsigned int child, const RefinementCase< dim > &refinement_case) const override
virtual FiniteElementDomination::Domination compare_for_domination(const FiniteElement< dim, spacedim > &fe_other, const unsigned int codim=0) const override final
virtual const FullMatrix< double > & get_prolongation_matrix(const unsigned int child, const RefinementCase< dim > &refinement_case) const override
const unsigned int n_bubbles
Definition: fe_q_bubbles.h:193
virtual void convert_generalized_support_point_values_to_dof_values(const std::vector< Vector< double > > &support_point_values, std::vector< double > &nodal_values) const override
virtual std::unique_ptr< FiniteElement< dim, spacedim > > clone() const override
FE_Q_Bubbles(const unsigned int p)
virtual std::string get_name() const override
static std::vector< unsigned int > get_dpo_vector(const unsigned int degree)
unsigned int n_dofs_per_cell() const
virtual std::string get_name() const =0
std::vector< std::vector< FullMatrix< double > > > restriction
Definition: fe.h:2413
void reinit_restriction_and_prolongation_matrices(const bool isotropic_restriction_only=false, const bool isotropic_prolongation_only=false)
std::vector< Point< dim > > unit_support_points
Definition: fe.h:2451
std::vector< std::vector< FullMatrix< double > > > prolongation
Definition: fe.h:2427
void mmult(FullMatrix< number2 > &C, const FullMatrix< number2 > &B, const bool adding=false) const
void set(const size_type i, const size_type j, const number value)
size_type n() const
void gauss_jordan()
size_type m() const
Definition: point.h:111
const Point< dim > & point(const unsigned int i) const
const std::vector< Point< dim > > & get_points() const
unsigned int size() const
virtual void execute_coarsening_and_refinement()
active_cell_iterator begin_active(const unsigned int level=0) const
Definition: vector.h:109
#define DEAL_II_NAMESPACE_OPEN
Definition: config.h:442
#define DEAL_II_NAMESPACE_CLOSE
Definition: config.h:443
@ update_values
Shape function values.
@ update_JxW_values
Transformed quadrature weights.
@ update_quadrature_points
Transformed quadrature points.
static ::ExceptionBase & ExcNotImplemented()
#define Assert(cond, exc)
Definition: exceptions.h:1473
static ::ExceptionBase & ExcImpossibleInDim(int arg1)
#define AssertDimension(dim1, dim2)
Definition: exceptions.h:1667
#define AssertIndexRange(index, range)
Definition: exceptions.h:1732
static ::ExceptionBase & ExcInternalError()
static ::ExceptionBase & ExcDimensionMismatch(std::size_t arg1, std::size_t arg2)
static ::ExceptionBase & ExcMessage(std::string arg1)
#define AssertThrow(cond, exc)
Definition: exceptions.h:1583
const Mapping< dim, spacedim > & get_default_linear_mapping(const Triangulation< dim, spacedim > &triangulation)
Definition: mapping.cc:260
void compute_projection_matrices(const FiniteElement< dim, spacedim > &fe, std::vector< std::vector< FullMatrix< number > > > &matrices, const bool isotropic_only=false)
void hyper_cube(Triangulation< dim, spacedim > &tria, const double left=0., const double right=1., const bool colorize=false)
std::string dim_string(const int dim, const int spacedim)
Definition: utilities.cc:558
static unsigned int n_children(const RefinementCase< dim > &refinement_case)