Reference documentation for deal.II version 9.4.0
\(\newcommand{\dealvcentcolon}{\mathrel{\mathop{:}}}\) \(\newcommand{\dealcoloneq}{\dealvcentcolon\mathrel{\mkern-1.2mu}=}\) \(\newcommand{\jump}[1]{\left[\!\left[ #1 \right]\!\right]}\) \(\newcommand{\average}[1]{\left\{\!\left\{ #1 \right\}\!\right\}}\)
step-61.h
Go to the documentation of this file.
1 ,
734  * const unsigned int /*component*/) const
735  * {
736  * return 0;
737  * }
738  *
739  *
740  *
741  * template <int dim>
742  * class RightHandSide : public Function<dim>
743  * {
744  * public:
745  * virtual double value(const Point<dim> & p,
746  * const unsigned int component = 0) const override;
747  * };
748  *
749  *
750  *
751  * template <int dim>
752  * double RightHandSide<dim>::value(const Point<dim> &p,
753  * const unsigned int /*component*/) const
754  * {
755  * return (2 * numbers::PI * numbers::PI * std::sin(numbers::PI * p[0]) *
756  * std::sin(numbers::PI * p[1]));
757  * }
758  *
759  *
760  *
761  * @endcode
762  *
763  * The class that implements the exact pressure solution has an
764  * oddity in that we implement it as a vector-valued one with two
765  * components. (We say that it has two components in the constructor
766  * where we call the constructor of the base Function class.) In the
767  * `value()` function, we do not test for the value of the
768  * `component` argument, which implies that we return the same value
769  * for both components of the vector-valued function. We do this
770  * because we describe the finite element in use in this program as
771  * a vector-valued system that contains the interior and the
772  * interface pressures, and when we compute errors, we will want to
773  * use the same pressure solution to test both of these components.
774  *
775  * @code
776  * template <int dim>
777  * class ExactPressure : public Function<dim>
778  * {
779  * public:
780  * ExactPressure()
781  * : Function<dim>(2)
782  * {}
783  *
784  * virtual double value(const Point<dim> & p,
785  * const unsigned int component) const override;
786  * };
787  *
788  *
789  *
790  * template <int dim>
791  * double ExactPressure<dim>::value(const Point<dim> &p,
792  * const unsigned int /*component*/) const
793  * {
794  * return std::sin(numbers::PI * p[0]) * std::sin(numbers::PI * p[1]);
795  * }
796  *
797  *
798  *
799  * template <int dim>
800  * class ExactVelocity : public TensorFunction<1, dim>
801  * {
802  * public:
803  * ExactVelocity()
805  * {}
806  *
807  * virtual Tensor<1, dim> value(const Point<dim> &p) const override;
808  * };
809  *
810  *
811  *
812  * template <int dim>
813  * Tensor<1, dim> ExactVelocity<dim>::value(const Point<dim> &p) const
814  * {
815  * Tensor<1, dim> return_value;
816  * return_value[0] = -numbers::PI * std::cos(numbers::PI * p[0]) *
817  * std::sin(numbers::PI * p[1]);
818  * return_value[1] = -numbers::PI * std::sin(numbers::PI * p[0]) *
819  * std::cos(numbers::PI * p[1]);
820  * return return_value;
821  * }
822  *
823  *
824  *
825  * @endcode
826  *
827  *
828  * <a name="WGDarcyEquationclassimplementation"></a>
829  * <h3>WGDarcyEquation class implementation</h3>
830  *
831 
832  *
833  *
834  * <a name="WGDarcyEquationWGDarcyEquation"></a>
835  * <h4>WGDarcyEquation::WGDarcyEquation</h4>
836  *
837 
838  *
839  * In this constructor, we create a finite element space for vector valued
840  * functions, which will here include the ones used for the interior and
841  * interface pressures, @f$p^\circ@f$ and @f$p^\partial@f$.
842  *
843  * @code
844  * template <int dim>
845  * WGDarcyEquation<dim>::WGDarcyEquation(const unsigned int degree)
846  * : fe(FE_DGQ<dim>(degree), 1, FE_FaceQ<dim>(degree), 1)
847  * , dof_handler(triangulation)
848  * , fe_dgrt(degree)
849  * , dof_handler_dgrt(triangulation)
850  * {}
851  *
852  *
853  *
854  * @endcode
855  *
856  *
857  * <a name="WGDarcyEquationmake_grid"></a>
858  * <h4>WGDarcyEquation::make_grid</h4>
859  *
860 
861  *
862  * We generate a mesh on the unit square domain and refine it.
863  *
864  * @code
865  * template <int dim>
866  * void WGDarcyEquation<dim>::make_grid()
867  * {
869  * triangulation.refine_global(5);
870  *
871  * std::cout << " Number of active cells: " << triangulation.n_active_cells()
872  * << std::endl
873  * << " Total number of cells: " << triangulation.n_cells()
874  * << std::endl;
875  * }
876  *
877  *
878  *
879  * @endcode
880  *
881  *
882  * <a name="WGDarcyEquationsetup_system"></a>
883  * <h4>WGDarcyEquation::setup_system</h4>
884  *
885 
886  *
887  * After we have created the mesh above, we distribute degrees of
888  * freedom and resize matrices and vectors. The only piece of
889  * interest in this function is how we interpolate the boundary
890  * values for the pressure. Since the pressure consists of interior
891  * and interface components, we need to make sure that we only
892  * interpolate onto that component of the vector-valued solution
893  * space that corresponds to the interface pressures (as these are
894  * the only ones that are defined on the boundary of the domain). We
895  * do this via a component mask object for only the interface
896  * pressures.
897  *
898  * @code
899  * template <int dim>
900  * void WGDarcyEquation<dim>::setup_system()
901  * {
902  * dof_handler.distribute_dofs(fe);
903  * dof_handler_dgrt.distribute_dofs(fe_dgrt);
904  *
905  * std::cout << " Number of pressure degrees of freedom: "
906  * << dof_handler.n_dofs() << std::endl;
907  *
908  * solution.reinit(dof_handler.n_dofs());
909  * system_rhs.reinit(dof_handler.n_dofs());
910  *
911  *
912  * {
913  * constraints.clear();
914  * const FEValuesExtractors::Scalar interface_pressure(1);
915  * const ComponentMask interface_pressure_mask =
916  * fe.component_mask(interface_pressure);
918  * 0,
919  * BoundaryValues<dim>(),
920  * constraints,
921  * interface_pressure_mask);
922  * constraints.close();
923  * }
924  *
925  *
926  * @endcode
927  *
928  * In the bilinear form, there is no integration term over faces
929  * between two neighboring cells, so we can just use
930  * <code>DoFTools::make_sparsity_pattern</code> to calculate the sparse
931  * matrix.
932  *
933  * @code
934  * DynamicSparsityPattern dsp(dof_handler.n_dofs());
935  * DoFTools::make_sparsity_pattern(dof_handler, dsp, constraints);
936  * sparsity_pattern.copy_from(dsp);
937  *
938  * system_matrix.reinit(sparsity_pattern);
939  * }
940  *
941  *
942  *
943  * @endcode
944  *
945  *
946  * <a name="WGDarcyEquationassemble_system"></a>
947  * <h4>WGDarcyEquation::assemble_system</h4>
948  *
949 
950  *
951  * This function is more interesting. As detailed in the
952  * introduction, the assembly of the linear system requires us to
953  * evaluate the weak gradient of the shape functions, which is an
954  * element in the Raviart-Thomas space. As a consequence, we need to
955  * define a Raviart-Thomas finite element object, and have FEValues
956  * objects that evaluate it at quadrature points. We then need to
957  * compute the matrix @f$C^K@f$ on every cell @f$K@f$, for which we need the
958  * matrices @f$M^K@f$ and @f$G^K@f$ mentioned in the introduction.
959  *
960 
961  *
962  * A point that may not be obvious is that in all previous tutorial
963  * programs, we have always called FEValues::reinit() with a cell
964  * iterator from a DoFHandler. This is so that one can call
965  * functions such as FEValuesBase::get_function_values() that
966  * extract the values of a finite element function (represented by a
967  * vector of DoF values) on the quadrature points of a cell. For
968  * this operation to work, one needs to know which vector elements
969  * correspond to the degrees of freedom on a given cell -- i.e.,
970  * exactly the kind of information and operation provided by the
971  * DoFHandler class.
972  *
973 
974  *
975  * We could create a DoFHandler object for the "broken" Raviart-Thomas space
976  * (using the FE_DGRT class), but we really don't want to here: At
977  * least in the current function, we have no need for any globally defined
978  * degrees of freedom associated with this broken space, but really only
979  * need to reference the shape functions of such a space on the current
980  * cell. As a consequence, we use the fact that one can call
981  * FEValues::reinit() also with cell iterators into Triangulation
982  * objects (rather than DoFHandler objects). In this case, FEValues
983  * can of course only provide us with information that only
984  * references information about cells, rather than degrees of freedom
985  * enumerated on these cells. So we can't use
986  * FEValuesBase::get_function_values(), but we can use
987  * FEValues::shape_value() to obtain the values of shape functions
988  * at quadrature points on the current cell. It is this kind of
989  * functionality we will make use of below. The variable that will
990  * give us this information about the Raviart-Thomas functions below
991  * is then the `fe_values_rt` (and corresponding `fe_face_values_rt`)
992  * object.
993  *
994 
995  *
996  * Given this introduction, the following declarations should be
997  * pretty obvious:
998  *
999  * @code
1000  * template <int dim>
1001  * void WGDarcyEquation<dim>::assemble_system()
1002  * {
1003  * const QGauss<dim> quadrature_formula(fe_dgrt.degree + 1);
1004  * const QGauss<dim - 1> face_quadrature_formula(fe_dgrt.degree + 1);
1005  *
1006  * FEValues<dim> fe_values(fe,
1007  * quadrature_formula,
1009  * update_JxW_values);
1010  * FEFaceValues<dim> fe_face_values(fe,
1011  * face_quadrature_formula,
1014  * update_JxW_values);
1015  *
1016  * FEValues<dim> fe_values_dgrt(fe_dgrt,
1017  * quadrature_formula,
1020  * update_JxW_values);
1021  * FEFaceValues<dim> fe_face_values_dgrt(fe_dgrt,
1022  * face_quadrature_formula,
1023  * update_values |
1026  * update_JxW_values);
1027  *
1028  * const unsigned int dofs_per_cell = fe.n_dofs_per_cell();
1029  * const unsigned int dofs_per_cell_dgrt = fe_dgrt.n_dofs_per_cell();
1030  *
1031  * const unsigned int n_q_points = fe_values.get_quadrature().size();
1032  * const unsigned int n_q_points_dgrt = fe_values_dgrt.get_quadrature().size();
1033  *
1034  * const unsigned int n_face_q_points = fe_face_values.get_quadrature().size();
1035  *
1036  * RightHandSide<dim> right_hand_side;
1037  * std::vector<double> right_hand_side_values(n_q_points);
1038  *
1039  * const Coefficient<dim> coefficient;
1040  * std::vector<Tensor<2, dim>> coefficient_values(n_q_points);
1041  *
1042  * std::vector<types::global_dof_index> local_dof_indices(dofs_per_cell);
1043  *
1044  *
1045  * @endcode
1046  *
1047  * Next, let us declare the various cell matrices discussed in the
1048  * introduction:
1049  *
1050  * @code
1051  * FullMatrix<double> cell_matrix_M(dofs_per_cell_dgrt, dofs_per_cell_dgrt);
1052  * FullMatrix<double> cell_matrix_G(dofs_per_cell_dgrt, dofs_per_cell);
1053  * FullMatrix<double> cell_matrix_C(dofs_per_cell, dofs_per_cell_dgrt);
1055  * Vector<double> cell_rhs(dofs_per_cell);
1056  * Vector<double> cell_solution(dofs_per_cell);
1057  *
1058  * @endcode
1059  *
1060  * We need <code>FEValuesExtractors</code> to access the @p interior and
1061  * @p face component of the shape functions.
1062  *
1063  * @code
1064  * const FEValuesExtractors::Vector velocities(0);
1065  * const FEValuesExtractors::Scalar pressure_interior(0);
1066  * const FEValuesExtractors::Scalar pressure_face(1);
1067  *
1068  * @endcode
1069  *
1070  * This finally gets us in position to loop over all cells. On
1071  * each cell, we will first calculate the various cell matrices
1072  * used to construct the local matrix -- as they depend on the
1073  * cell in question, they need to be re-computed on each cell. We
1074  * need shape functions for the Raviart-Thomas space as well, for
1075  * which we need to create first an iterator to the cell of the
1076  * triangulation, which we can obtain by assignment from the cell
1077  * pointing into the DoFHandler.
1078  *
1079  * @code
1080  * for (const auto &cell : dof_handler.active_cell_iterators())
1081  * {
1082  * fe_values.reinit(cell);
1083  *
1084  * const typename Triangulation<dim>::active_cell_iterator cell_dgrt =
1085  * cell;
1086  * fe_values_dgrt.reinit(cell_dgrt);
1087  *
1088  * right_hand_side.value_list(fe_values.get_quadrature_points(),
1089  * right_hand_side_values);
1090  * coefficient.value_list(fe_values.get_quadrature_points(),
1091  * coefficient_values);
1092  *
1093  * @endcode
1094  *
1095  * The first cell matrix we will compute is the mass matrix
1096  * for the Raviart-Thomas space. Hence, we need to loop over
1097  * all the quadrature points for the velocity FEValues object.
1098  *
1099  * @code
1100  * cell_matrix_M = 0;
1101  * for (unsigned int q = 0; q < n_q_points_dgrt; ++q)
1102  * for (unsigned int i = 0; i < dofs_per_cell_dgrt; ++i)
1103  * {
1104  * const Tensor<1, dim> v_i = fe_values_dgrt[velocities].value(i, q);
1105  * for (unsigned int k = 0; k < dofs_per_cell_dgrt; ++k)
1106  * {
1107  * const Tensor<1, dim> v_k =
1108  * fe_values_dgrt[velocities].value(k, q);
1109  * cell_matrix_M(i, k) += (v_i * v_k * fe_values_dgrt.JxW(q));
1110  * }
1111  * }
1112  * @endcode
1113  *
1114  * Next we take the inverse of this matrix by using
1115  * FullMatrix::gauss_jordan(). It will be used to calculate
1116  * the coefficient matrix @f$C^K@f$ later. It is worth recalling
1117  * later that `cell_matrix_M` actually contains the *inverse*
1118  * of @f$M^K@f$ after this call.
1119  *
1120  * @code
1121  * cell_matrix_M.gauss_jordan();
1122  *
1123  * @endcode
1124  *
1125  * From the introduction, we know that the right hand side
1126  * @f$G^K@f$ of the equation that defines @f$C^K@f$ is the difference
1127  * between a face integral and a cell integral. Here, we
1128  * approximate the negative of the contribution in the
1129  * interior. Each component of this matrix is the integral of
1130  * a product between a basis function of the polynomial space
1131  * and the divergence of a basis function of the
1132  * Raviart-Thomas space. These basis functions are defined in
1133  * the interior.
1134  *
1135  * @code
1136  * cell_matrix_G = 0;
1137  * for (unsigned int q = 0; q < n_q_points; ++q)
1138  * for (unsigned int i = 0; i < dofs_per_cell_dgrt; ++i)
1139  * {
1140  * const double div_v_i =
1141  * fe_values_dgrt[velocities].divergence(i, q);
1142  * for (unsigned int j = 0; j < dofs_per_cell; ++j)
1143  * {
1144  * const double phi_j_interior =
1145  * fe_values[pressure_interior].value(j, q);
1146  *
1147  * cell_matrix_G(i, j) -=
1148  * (div_v_i * phi_j_interior * fe_values.JxW(q));
1149  * }
1150  * }
1151  *
1152  *
1153  * @endcode
1154  *
1155  * Next, we approximate the integral on faces by quadrature.
1156  * Each component is the integral of a product between a basis function
1157  * of the polynomial space and the dot product of a basis function of
1158  * the Raviart-Thomas space and the normal vector. So we loop over all
1159  * the faces of the element and obtain the normal vector.
1160  *
1161  * @code
1162  * for (const auto &face : cell->face_iterators())
1163  * {
1164  * fe_face_values.reinit(cell, face);
1165  * fe_face_values_dgrt.reinit(cell_dgrt, face);
1166  *
1167  * for (unsigned int q = 0; q < n_face_q_points; ++q)
1168  * {
1169  * const Tensor<1, dim> &normal = fe_face_values.normal_vector(q);
1170  *
1171  * for (unsigned int i = 0; i < dofs_per_cell_dgrt; ++i)
1172  * {
1173  * const Tensor<1, dim> v_i =
1174  * fe_face_values_dgrt[velocities].value(i, q);
1175  * for (unsigned int j = 0; j < dofs_per_cell; ++j)
1176  * {
1177  * const double phi_j_face =
1178  * fe_face_values[pressure_face].value(j, q);
1179  *
1180  * cell_matrix_G(i, j) +=
1181  * ((v_i * normal) * phi_j_face * fe_face_values.JxW(q));
1182  * }
1183  * }
1184  * }
1185  * }
1186  *
1187  * @endcode
1188  *
1189  * @p cell_matrix_C is then the matrix product between the
1190  * transpose of @f$G^K@f$ and the inverse of the mass matrix
1191  * (where this inverse is stored in @p cell_matrix_M):
1192  *
1193  * @code
1194  * cell_matrix_G.Tmmult(cell_matrix_C, cell_matrix_M);
1195  *
1196  * @endcode
1197  *
1198  * Finally we can compute the local matrix @f$A^K@f$. Element
1199  * @f$A^K_{ij}@f$ is given by @f$\int_{E} \sum_{k,l} C_{ik} C_{jl}
1200  * (\mathbf{K} \mathbf{v}_k) \cdot \mathbf{v}_l
1201  * \mathrm{d}x@f$. We have calculated the coefficients @f$C@f$ in
1202  * the previous step, and so obtain the following after
1203  * suitably re-arranging the loops:
1204  *
1205  * @code
1206  * local_matrix = 0;
1207  * for (unsigned int q = 0; q < n_q_points_dgrt; ++q)
1208  * {
1209  * for (unsigned int k = 0; k < dofs_per_cell_dgrt; ++k)
1210  * {
1211  * const Tensor<1, dim> v_k =
1212  * fe_values_dgrt[velocities].value(k, q);
1213  * for (unsigned int l = 0; l < dofs_per_cell_dgrt; ++l)
1214  * {
1215  * const Tensor<1, dim> v_l =
1216  * fe_values_dgrt[velocities].value(l, q);
1217  *
1218  * for (unsigned int i = 0; i < dofs_per_cell; ++i)
1219  * for (unsigned int j = 0; j < dofs_per_cell; ++j)
1220  * local_matrix(i, j) +=
1221  * (coefficient_values[q] * cell_matrix_C[i][k] * v_k) *
1222  * cell_matrix_C[j][l] * v_l * fe_values_dgrt.JxW(q);
1223  * }
1224  * }
1225  * }
1226  *
1227  * @endcode
1228  *
1229  * Next, we calculate the right hand side, @f$\int_{K} f q \mathrm{d}x@f$:
1230  *
1231  * @code
1232  * cell_rhs = 0;
1233  * for (unsigned int q = 0; q < n_q_points; ++q)
1234  * for (unsigned int i = 0; i < dofs_per_cell; ++i)
1235  * {
1236  * cell_rhs(i) += (fe_values[pressure_interior].value(i, q) *
1237  * right_hand_side_values[q] * fe_values.JxW(q));
1238  * }
1239  *
1240  * @endcode
1241  *
1242  * The last step is to distribute components of the local
1243  * matrix into the system matrix and transfer components of
1244  * the cell right hand side into the system right hand side:
1245  *
1246  * @code
1247  * cell->get_dof_indices(local_dof_indices);
1248  * constraints.distribute_local_to_global(
1249  * local_matrix, cell_rhs, local_dof_indices, system_matrix, system_rhs);
1250  * }
1251  * }
1252  *
1253  *
1254  *
1255  * @endcode
1256  *
1257  *
1258  * <a name="WGDarcyEquationdimsolve"></a>
1259  * <h4>WGDarcyEquation<dim>::solve</h4>
1260  *
1261 
1262  *
1263  * This step is rather trivial and the same as in many previous
1264  * tutorial programs:
1265  *
1266  * @code
1267  * template <int dim>
1268  * void WGDarcyEquation<dim>::solve()
1269  * {
1270  * SolverControl solver_control(1000, 1e-8 * system_rhs.l2_norm());
1271  * SolverCG<Vector<double>> solver(solver_control);
1272  * solver.solve(system_matrix, solution, system_rhs, PreconditionIdentity());
1273  * constraints.distribute(solution);
1274  * }
1275  *
1276  *
1277  * @endcode
1278  *
1279  *
1280  * <a name="WGDarcyEquationdimcompute_postprocessed_velocity"></a>
1281  * <h4>WGDarcyEquation<dim>::compute_postprocessed_velocity</h4>
1282  *
1283 
1284  *
1285  * In this function, compute the velocity field from the pressure
1286  * solution previously computed. The
1287  * velocity is defined as @f$\mathbf{u}_h = \mathbf{Q}_h \left(
1288  * -\mathbf{K}\nabla_{w,d}p_h \right)@f$, which requires us to compute
1289  * many of the same terms as in the assembly of the system matrix.
1290  * There are also the matrices @f$E^K,D^K@f$ we need to assemble (see
1291  * the introduction) but they really just follow the same kind of
1292  * pattern.
1293  *
1294 
1295  *
1296  * Computing the same matrices here as we have already done in the
1297  * `assemble_system()` function is of course wasteful in terms of
1298  * CPU time. Likewise, we copy some of the code from there to this
1299  * function, and this is also generally a poor idea. A better
1300  * implementation might provide for a function that encapsulates
1301  * this duplicated code. One could also think of using the classic
1302  * trade-off between computing efficiency and memory efficiency to
1303  * only compute the @f$C^K@f$ matrices once per cell during the
1304  * assembly, storing them somewhere on the side, and re-using them
1305  * here. (This is what @ref step_51 "step-51" does, for example, where the
1306  * `assemble_system()` function takes an argument that determines
1307  * whether the local matrices are recomputed, and a similar approach
1308  * -- maybe with storing local matrices elsewhere -- could be
1309  * adapted for the current program.)
1310  *
1311  * @code
1312  * template <int dim>
1313  * void WGDarcyEquation<dim>::compute_postprocessed_velocity()
1314  * {
1315  * darcy_velocity.reinit(dof_handler_dgrt.n_dofs());
1316  *
1317  * const QGauss<dim> quadrature_formula(fe_dgrt.degree + 1);
1318  * const QGauss<dim - 1> face_quadrature_formula(fe_dgrt.degree + 1);
1319  *
1320  * FEValues<dim> fe_values(fe,
1321  * quadrature_formula,
1323  * update_JxW_values);
1324  *
1325  * FEFaceValues<dim> fe_face_values(fe,
1326  * face_quadrature_formula,
1329  * update_JxW_values);
1330  *
1331  * FEValues<dim> fe_values_dgrt(fe_dgrt,
1332  * quadrature_formula,
1335  * update_JxW_values);
1336  *
1337  * FEFaceValues<dim> fe_face_values_dgrt(fe_dgrt,
1338  * face_quadrature_formula,
1339  * update_values |
1342  * update_JxW_values);
1343  *
1344  * const unsigned int dofs_per_cell = fe.n_dofs_per_cell();
1345  * const unsigned int dofs_per_cell_dgrt = fe_dgrt.n_dofs_per_cell();
1346  *
1347  * const unsigned int n_q_points = fe_values.get_quadrature().size();
1348  * const unsigned int n_q_points_dgrt = fe_values_dgrt.get_quadrature().size();
1349  *
1350  * const unsigned int n_face_q_points = fe_face_values.get_quadrature().size();
1351  *
1352  *
1353  * std::vector<types::global_dof_index> local_dof_indices(dofs_per_cell);
1354  * std::vector<types::global_dof_index> local_dof_indices_dgrt(
1355  * dofs_per_cell_dgrt);
1356  *
1357  * FullMatrix<double> cell_matrix_M(dofs_per_cell_dgrt, dofs_per_cell_dgrt);
1358  * FullMatrix<double> cell_matrix_G(dofs_per_cell_dgrt, dofs_per_cell);
1359  * FullMatrix<double> cell_matrix_C(dofs_per_cell, dofs_per_cell_dgrt);
1360  * FullMatrix<double> cell_matrix_D(dofs_per_cell_dgrt, dofs_per_cell_dgrt);
1361  * FullMatrix<double> cell_matrix_E(dofs_per_cell_dgrt, dofs_per_cell_dgrt);
1362  *
1363  * Vector<double> cell_solution(dofs_per_cell);
1364  * Vector<double> cell_velocity(dofs_per_cell_dgrt);
1365  *
1366  * const Coefficient<dim> coefficient;
1367  * std::vector<Tensor<2, dim>> coefficient_values(n_q_points_dgrt);
1368  *
1369  * const FEValuesExtractors::Vector velocities(0);
1370  * const FEValuesExtractors::Scalar pressure_interior(0);
1371  * const FEValuesExtractors::Scalar pressure_face(1);
1372  *
1373  * @endcode
1374  *
1375  * In the introduction, we explained how to calculate the numerical velocity
1376  * on the cell. We need the pressure solution values on each cell,
1377  * coefficients of the Gram matrix and coefficients of the @f$L_2@f$ projection.
1378  * We have already calculated the global solution, so we will extract the
1379  * cell solution from the global solution. The coefficients of the Gram
1380  * matrix have been calculated when we assembled the system matrix for the
1381  * pressures. We will do the same way here. For the coefficients of the
1382  * projection, we do matrix multiplication, i.e., the inverse of the Gram
1383  * matrix times the matrix with @f$(\mathbf{K} \mathbf{w}, \mathbf{w})@f$ as
1384  * components. Then, we multiply all these coefficients and call them beta.
1385  * The numerical velocity is the product of beta and the basis functions of
1386  * the Raviart-Thomas space.
1387  *
1388  * @code
1390  * cell = dof_handler.begin_active(),
1391  * endc = dof_handler.end(), cell_dgrt = dof_handler_dgrt.begin_active();
1392  * for (; cell != endc; ++cell, ++cell_dgrt)
1393  * {
1394  * fe_values.reinit(cell);
1395  * fe_values_dgrt.reinit(cell_dgrt);
1396  *
1397  * coefficient.value_list(fe_values_dgrt.get_quadrature_points(),
1398  * coefficient_values);
1399  *
1400  * @endcode
1401  *
1402  * The component of this <code>cell_matrix_E</code> is the integral of
1403  * @f$(\mathbf{K} \mathbf{w}, \mathbf{w})@f$. <code>cell_matrix_M</code> is
1404  * the Gram matrix.
1405  *
1406  * @code
1407  * cell_matrix_M = 0;
1408  * cell_matrix_E = 0;
1409  * for (unsigned int q = 0; q < n_q_points_dgrt; ++q)
1410  * for (unsigned int i = 0; i < dofs_per_cell_dgrt; ++i)
1411  * {
1412  * const Tensor<1, dim> v_i = fe_values_dgrt[velocities].value(i, q);
1413  * for (unsigned int k = 0; k < dofs_per_cell_dgrt; ++k)
1414  * {
1415  * const Tensor<1, dim> v_k =
1416  * fe_values_dgrt[velocities].value(k, q);
1417  *
1418  * cell_matrix_E(i, k) +=
1419  * (coefficient_values[q] * v_i * v_k * fe_values_dgrt.JxW(q));
1420  *
1421  * cell_matrix_M(i, k) += (v_i * v_k * fe_values_dgrt.JxW(q));
1422  * }
1423  * }
1424  *
1425  * @endcode
1426  *
1427  * To compute the matrix @f$D@f$ mentioned in the introduction, we
1428  * then need to evaluate @f$D=M^{-1}E@f$ as explained in the
1429  * introduction:
1430  *
1431  * @code
1432  * cell_matrix_M.gauss_jordan();
1433  * cell_matrix_M.mmult(cell_matrix_D, cell_matrix_E);
1434  *
1435  * @endcode
1436  *
1437  * Then we also need, again, to compute the matrix @f$C@f$ that is
1438  * used to evaluate the weak discrete gradient. This is the
1439  * exact same code as used in the assembly of the system
1440  * matrix, so we just copy it from there:
1441  *
1442  * @code
1443  * cell_matrix_G = 0;
1444  * for (unsigned int q = 0; q < n_q_points; ++q)
1445  * for (unsigned int i = 0; i < dofs_per_cell_dgrt; ++i)
1446  * {
1447  * const double div_v_i =
1448  * fe_values_dgrt[velocities].divergence(i, q);
1449  * for (unsigned int j = 0; j < dofs_per_cell; ++j)
1450  * {
1451  * const double phi_j_interior =
1452  * fe_values[pressure_interior].value(j, q);
1453  *
1454  * cell_matrix_G(i, j) -=
1455  * (div_v_i * phi_j_interior * fe_values.JxW(q));
1456  * }
1457  * }
1458  *
1459  * for (const auto &face : cell->face_iterators())
1460  * {
1461  * fe_face_values.reinit(cell, face);
1462  * fe_face_values_dgrt.reinit(cell_dgrt, face);
1463  *
1464  * for (unsigned int q = 0; q < n_face_q_points; ++q)
1465  * {
1466  * const Tensor<1, dim> &normal = fe_face_values.normal_vector(q);
1467  *
1468  * for (unsigned int i = 0; i < dofs_per_cell_dgrt; ++i)
1469  * {
1470  * const Tensor<1, dim> v_i =
1471  * fe_face_values_dgrt[velocities].value(i, q);
1472  * for (unsigned int j = 0; j < dofs_per_cell; ++j)
1473  * {
1474  * const double phi_j_face =
1475  * fe_face_values[pressure_face].value(j, q);
1476  *
1477  * cell_matrix_G(i, j) +=
1478  * ((v_i * normal) * phi_j_face * fe_face_values.JxW(q));
1479  * }
1480  * }
1481  * }
1482  * }
1483  * cell_matrix_G.Tmmult(cell_matrix_C, cell_matrix_M);
1484  *
1485  * @endcode
1486  *
1487  * Finally, we need to extract the pressure unknowns that
1488  * correspond to the current cell:
1489  *
1490  * @code
1491  * cell->get_dof_values(solution, cell_solution);
1492  *
1493  * @endcode
1494  *
1495  * We are now in a position to compute the local velocity
1496  * unknowns (with respect to the Raviart-Thomas space we are
1497  * projecting the term @f$-\mathbf K \nabla_{w,d} p_h@f$ into):
1498  *
1499  * @code
1500  * cell_velocity = 0;
1501  * for (unsigned int k = 0; k < dofs_per_cell_dgrt; ++k)
1502  * for (unsigned int j = 0; j < dofs_per_cell_dgrt; ++j)
1503  * for (unsigned int i = 0; i < dofs_per_cell; ++i)
1504  * cell_velocity(k) +=
1505  * -(cell_solution(i) * cell_matrix_C(i, j) * cell_matrix_D(k, j));
1506  *
1507  * @endcode
1508  *
1509  * We compute Darcy velocity.
1510  * This is same as cell_velocity but used to graph Darcy velocity.
1511  *
1512  * @code
1513  * cell_dgrt->get_dof_indices(local_dof_indices_dgrt);
1514  * for (unsigned int k = 0; k < dofs_per_cell_dgrt; ++k)
1515  * for (unsigned int j = 0; j < dofs_per_cell_dgrt; ++j)
1516  * for (unsigned int i = 0; i < dofs_per_cell; ++i)
1517  * darcy_velocity(local_dof_indices_dgrt[k]) +=
1518  * -(cell_solution(i) * cell_matrix_C(i, j) * cell_matrix_D(k, j));
1519  * }
1520  * }
1521  *
1522  *
1523  *
1524  * @endcode
1525  *
1526  *
1527  * <a name="WGDarcyEquationdimcompute_pressure_error"></a>
1528  * <h4>WGDarcyEquation<dim>::compute_pressure_error</h4>
1529  *
1530 
1531  *
1532  * This part is to calculate the @f$L_2@f$ error of the pressure. We
1533  * define a vector that holds the norm of the error on each cell.
1534  * Next, we use VectorTool::integrate_difference() to compute the
1535  * error in the @f$L_2@f$ norm on each cell. However, we really only
1536  * care about the error in the interior component of the solution
1537  * vector (we can't even evaluate the interface pressures at the
1538  * quadrature points because these are all located in the interior
1539  * of cells) and consequently have to use a weight function that
1540  * ensures that the interface component of the solution variable is
1541  * ignored. This is done by using the ComponentSelectFunction whose
1542  * arguments indicate which component we want to select (component
1543  * zero, i.e., the interior pressures) and how many components there
1544  * are in total (two).
1545  *
1546  * @code
1547  * template <int dim>
1548  * void WGDarcyEquation<dim>::compute_pressure_error()
1549  * {
1550  * Vector<float> difference_per_cell(triangulation.n_active_cells());
1551  * const ComponentSelectFunction<dim> select_interior_pressure(0, 2);
1552  * VectorTools::integrate_difference(dof_handler,
1553  * solution,
1554  * ExactPressure<dim>(),
1555  * difference_per_cell,
1556  * QGauss<dim>(fe.degree + 2),
1558  * &select_interior_pressure);
1559  *
1560  * const double L2_error = difference_per_cell.l2_norm();
1561  * std::cout << "L2_error_pressure " << L2_error << std::endl;
1562  * }
1563  *
1564  *
1565  *
1566  * @endcode
1567  *
1568  *
1569  * <a name="WGDarcyEquationdimcompute_velocity_error"></a>
1570  * <h4>WGDarcyEquation<dim>::compute_velocity_error</h4>
1571  *
1572 
1573  *
1574  * In this function, we evaluate @f$L_2@f$ errors for the velocity on
1575  * each cell, and @f$L_2@f$ errors for the flux on faces. The function
1576  * relies on the `compute_postprocessed_velocity()` function having
1577  * previous computed, which computes the velocity field based on the
1578  * pressure solution that has previously been computed.
1579  *
1580 
1581  *
1582  * We are going to evaluate velocities on each cell and calculate
1583  * the difference between numerical and exact velocities.
1584  *
1585  * @code
1586  * template <int dim>
1587  * void WGDarcyEquation<dim>::compute_velocity_errors()
1588  * {
1589  * const QGauss<dim> quadrature_formula(fe_dgrt.degree + 1);
1590  * const QGauss<dim - 1> face_quadrature_formula(fe_dgrt.degree + 1);
1591  *
1592  * FEValues<dim> fe_values_dgrt(fe_dgrt,
1593  * quadrature_formula,
1596  * update_JxW_values);
1597  *
1598  * FEFaceValues<dim> fe_face_values_dgrt(fe_dgrt,
1599  * face_quadrature_formula,
1600  * update_values |
1603  * update_JxW_values);
1604  *
1605  * const unsigned int n_q_points_dgrt = fe_values_dgrt.get_quadrature().size();
1606  * const unsigned int n_face_q_points_dgrt =
1607  * fe_face_values_dgrt.get_quadrature().size();
1608  *
1609  * std::vector<Tensor<1, dim>> velocity_values(n_q_points_dgrt);
1610  * std::vector<Tensor<1, dim>> velocity_face_values(n_face_q_points_dgrt);
1611  *
1612  * const FEValuesExtractors::Vector velocities(0);
1613  *
1614  * const ExactVelocity<dim> exact_velocity;
1615  *
1616  * double L2_err_velocity_cell_sqr_global = 0;
1617  * double L2_err_flux_sqr = 0;
1618  *
1619  * @endcode
1620  *
1621  * Having previously computed the postprocessed velocity, we here
1622  * only have to extract the corresponding values on each cell and
1623  * face and compare it to the exact values.
1624  *
1625  * @code
1626  * for (const auto &cell_dgrt : dof_handler_dgrt.active_cell_iterators())
1627  * {
1628  * fe_values_dgrt.reinit(cell_dgrt);
1629  *
1630  * @endcode
1631  *
1632  * First compute the @f$L_2@f$ error between the postprocessed velocity
1633  * field and the exact one:
1634  *
1635  * @code
1636  * fe_values_dgrt[velocities].get_function_values(darcy_velocity,
1637  * velocity_values);
1638  * double L2_err_velocity_cell_sqr_local = 0;
1639  * for (unsigned int q = 0; q < n_q_points_dgrt; ++q)
1640  * {
1641  * const Tensor<1, dim> velocity = velocity_values[q];
1642  * const Tensor<1, dim> true_velocity =
1643  * exact_velocity.value(fe_values_dgrt.quadrature_point(q));
1644  *
1645  * L2_err_velocity_cell_sqr_local +=
1646  * ((velocity - true_velocity) * (velocity - true_velocity) *
1647  * fe_values_dgrt.JxW(q));
1648  * }
1649  * L2_err_velocity_cell_sqr_global += L2_err_velocity_cell_sqr_local;
1650  *
1651  * @endcode
1652  *
1653  * For reconstructing the flux we need the size of cells and
1654  * faces. Since fluxes are calculated on faces, we have the
1655  * loop over all four faces of each cell. To calculate the
1656  * face velocity, we extract values at the quadrature points from the
1657  * `darcy_velocity` which we have computed previously. Then, we
1658  * calculate the squared velocity error in normal direction. Finally, we
1659  * calculate the @f$L_2@f$ flux error on the cell by appropriately scaling
1660  * with face and cell areas and add it to the global error.
1661  *
1662  * @code
1663  * const double cell_area = cell_dgrt->measure();
1664  * for (const auto &face_dgrt : cell_dgrt->face_iterators())
1665  * {
1666  * const double face_length = face_dgrt->measure();
1667  * fe_face_values_dgrt.reinit(cell_dgrt, face_dgrt);
1668  * fe_face_values_dgrt[velocities].get_function_values(
1669  * darcy_velocity, velocity_face_values);
1670  *
1671  * double L2_err_flux_face_sqr_local = 0;
1672  * for (unsigned int q = 0; q < n_face_q_points_dgrt; ++q)
1673  * {
1674  * const Tensor<1, dim> velocity = velocity_face_values[q];
1675  * const Tensor<1, dim> true_velocity =
1676  * exact_velocity.value(fe_face_values_dgrt.quadrature_point(q));
1677  *
1678  * const Tensor<1, dim> &normal =
1679  * fe_face_values_dgrt.normal_vector(q);
1680  *
1681  * L2_err_flux_face_sqr_local +=
1682  * ((velocity * normal - true_velocity * normal) *
1683  * (velocity * normal - true_velocity * normal) *
1684  * fe_face_values_dgrt.JxW(q));
1685  * }
1686  * const double err_flux_each_face =
1687  * L2_err_flux_face_sqr_local / face_length * cell_area;
1688  * L2_err_flux_sqr += err_flux_each_face;
1689  * }
1690  * }
1691  *
1692  * @endcode
1693  *
1694  * After adding up errors over all cells and faces, we take the
1695  * square root and get the @f$L_2@f$ errors of velocity and
1696  * flux. These we output to screen.
1697  *
1698  * @code
1699  * const double L2_err_velocity_cell =
1700  * std::sqrt(L2_err_velocity_cell_sqr_global);
1701  * const double L2_err_flux_face = std::sqrt(L2_err_flux_sqr);
1702  *
1703  * std::cout << "L2_error_vel: " << L2_err_velocity_cell << std::endl
1704  * << "L2_error_flux: " << L2_err_flux_face << std::endl;
1705  * }
1706  *
1707  *
1708  * @endcode
1709  *
1710  *
1711  * <a name="WGDarcyEquationoutput_results"></a>
1712  * <h4>WGDarcyEquation::output_results</h4>
1713  *
1714 
1715  *
1716  * We have two sets of results to output: the interior solution and
1717  * the skeleton solution. We use <code>DataOut</code> to visualize
1718  * interior results. The graphical output for the skeleton results
1719  * is done by using the DataOutFaces class.
1720  *
1721 
1722  *
1723  * In both of the output files, both the interior and the face
1724  * variables are stored. For the interface output, the output file
1725  * simply contains the interpolation of the interior pressures onto
1726  * the faces, but because it is undefined which of the two interior
1727  * pressure variables you get from the two adjacent cells, it is
1728  * best to ignore the interior pressure in the interface output
1729  * file. Conversely, for the cell interior output file, it is of
1730  * course impossible to show any interface pressures @f$p^\partial@f$,
1731  * because these are only available on interfaces and not cell
1732  * interiors. Consequently, you will see them shown as an invalid
1733  * value (such as an infinity).
1734  *
1735 
1736  *
1737  * For the cell interior output, we also want to output the velocity
1738  * variables. This is a bit tricky since it lives on the same mesh
1739  * but uses a different DoFHandler object (the pressure variables live
1740  * on the `dof_handler` object, the Darcy velocity on the `dof_handler_dgrt`
1741  * object). Fortunately, there are variations of the
1742  * DataOut::add_data_vector() function that allow specifying which
1743  * DoFHandler a vector corresponds to, and consequently we can visualize
1744  * the data from both DoFHandler objects within the same file.
1745  *
1746  * @code
1747  * template <int dim>
1748  * void WGDarcyEquation<dim>::output_results() const
1749  * {
1750  * {
1751  * DataOut<dim> data_out;
1752  *
1753  * @endcode
1754  *
1755  * First attach the pressure solution to the DataOut object:
1756  *
1757  * @code
1758  * const std::vector<std::string> solution_names = {"interior_pressure",
1759  * "interface_pressure"};
1760  * data_out.add_data_vector(dof_handler, solution, solution_names);
1761  *
1762  * @endcode
1763  *
1764  * Then do the same with the Darcy velocity field, and continue
1765  * with writing everything out into a file.
1766  *
1767  * @code
1768  * const std::vector<std::string> velocity_names(dim, "velocity");
1769  * const std::vector<
1771  * velocity_component_interpretation(
1773  * data_out.add_data_vector(dof_handler_dgrt,
1774  * darcy_velocity,
1775  * velocity_names,
1776  * velocity_component_interpretation);
1777  *
1778  * data_out.build_patches(fe.degree);
1779  * std::ofstream output("solution_interior.vtu");
1780  * data_out.write_vtu(output);
1781  * }
1782  *
1783  * {
1784  * DataOutFaces<dim> data_out_faces(false);
1785  * data_out_faces.attach_dof_handler(dof_handler);
1786  * data_out_faces.add_data_vector(solution, "Pressure_Face");
1787  * data_out_faces.build_patches(fe.degree);
1788  * std::ofstream face_output("solution_interface.vtu");
1789  * data_out_faces.write_vtu(face_output);
1790  * }
1791  * }
1792  *
1793  *
1794  * @endcode
1795  *
1796  *
1797  * <a name="WGDarcyEquationrun"></a>
1798  * <h4>WGDarcyEquation::run</h4>
1799  *
1800 
1801  *
1802  * This is the final function of the main class. It calls the other functions
1803  * of our class.
1804  *
1805  * @code
1806  * template <int dim>
1807  * void WGDarcyEquation<dim>::run()
1808  * {
1809  * std::cout << "Solving problem in " << dim << " space dimensions."
1810  * << std::endl;
1811  * make_grid();
1812  * setup_system();
1813  * assemble_system();
1814  * solve();
1815  * compute_postprocessed_velocity();
1816  * compute_pressure_error();
1817  * compute_velocity_errors();
1818  * output_results();
1819  * }
1820  *
1821  * } // namespace Step61
1822  *
1823  *
1824  * @endcode
1825  *
1826  *
1827  * <a name="Thecodemaincodefunction"></a>
1828  * <h3>The <code>main</code> function</h3>
1829  *
1830 
1831  *
1832  * This is the main function. We can change the dimension here to run in 3d.
1833  *
1834  * @code
1835  * int main()
1836  * {
1837  * try
1838  * {
1839  * Step61::WGDarcyEquation<2> wg_darcy(0);
1840  * wg_darcy.run();
1841  * }
1842  * catch (std::exception &exc)
1843  * {
1844  * std::cerr << std::endl
1845  * << std::endl
1846  * << "----------------------------------------------------"
1847  * << std::endl;
1848  * std::cerr << "Exception on processing: " << std::endl
1849  * << exc.what() << std::endl
1850  * << "Aborting!" << std::endl
1851  * << "----------------------------------------------------"
1852  * << std::endl;
1853  * return 1;
1854  * }
1855  * catch (...)
1856  * {
1857  * std::cerr << std::endl
1858  * << std::endl
1859  * << "----------------------------------------------------"
1860  * << std::endl;
1861  * std::cerr << "Unknown exception!" << std::endl
1862  * << "Aborting!" << std::endl
1863  * << "----------------------------------------------------"
1864  * << std::endl;
1865  * return 1;
1866  * }
1867  *
1868  * return 0;
1869  * }
1870  * @endcode
1871 <a name="Results"></a><h1>Results</h1>
1872 
1873 
1874 We run the program with a right hand side that will produce the
1875 solution @f$p = \sin(\pi x) \sin(\pi y)@f$ and with homogeneous Dirichlet
1876 boundary conditions in the domain @f$\Omega = (0,1)^2@f$. In addition, we
1877 choose the coefficient matrix in the differential operator
1878 @f$\mathbf{K}@f$ as the identity matrix. We test this setup using
1879 @f$\mbox{WG}(Q_0,Q_0;RT_{[0]})@f$, @f$\mbox{WG}(Q_1,Q_1;RT_{[1]})@f$ and
1880 @f$\mbox{WG}(Q_2,Q_2;RT_{[2]})@f$ element combinations, which one can
1881 select by using the appropriate constructor argument for the
1882 `WGDarcyEquation` object in `main()`. We will then visualize pressure
1883 values in interiors of cells and on faces. We want to see that the
1884 pressure maximum is around 1 and the minimum is around 0. With mesh
1885 refinement, the convergence rates of pressure, velocity and flux
1886 should then be around 1 for @f$\mbox{WG}(Q_0,Q_0;RT_{[0]})@f$ , 2 for
1887 @f$\mbox{WG}(Q_1,Q_1;RT_{[1]})@f$, and 3 for
1888 @f$\mbox{WG}(Q_2,Q_2;RT_{[2]})@f$.
1889 
1890 
1891 <a name="TestresultsoniWGQsub0subQsub0subRTsub0subi"></a><h3>Test results on <i>WG(Q<sub>0</sub>,Q<sub>0</sub>;RT<sub>[0]</sub>)</i></h3>
1892 
1893 
1894 The following figures show interior pressures and face pressures using the
1895 @f$\mbox{WG}(Q_0,Q_0;RT_{[0]})@f$ element. The mesh is refined 2 times (top)
1896 and 4 times (bottom), respectively. (This number can be adjusted in the
1897 `make_grid()` function.) When the mesh is coarse, one can see
1898 the face pressures @f$p^\partial@f$ neatly between the values of the interior
1899 pressures @f$p^\circ@f$ on the two adjacent cells.
1900 
1901 <table align="center">
1902  <tr>
1903  <td><img src="https://www.dealii.org/images/steps/developer/step-61.wg000_2d_2.png" alt=""></td>
1904  <td><img src="https://www.dealii.org/images/steps/developer/step-61.wg000_3d_2.png" alt=""></td>
1905  </tr>
1906  <tr>
1907  <td><img src="https://www.dealii.org/images/steps/developer/step-61.wg000_2d_4.png" alt=""></td>
1908  <td><img src="https://www.dealii.org/images/steps/developer/step-61.wg000_3d_4.png" alt=""></td>
1909  </tr>
1910 </table>
1911 
1912 From the figures, we can see that with the mesh refinement, the maximum and
1913 minimum pressure values are approaching the values we expect.
1914 Since the mesh is a rectangular mesh and numbers of cells in each direction is even, we
1915 have symmetric solutions. From the 3d figures on the right,
1916 we can see that on @f$\mbox{WG}(Q_0,Q_0;RT_{[0]})@f$, the pressure is a constant
1917 in the interior of the cell, as expected.
1918 
1919 <a name="Convergencetableforik0i"></a><h4>Convergence table for <i>k=0</i></h4>
1920 
1921 
1922 We run the code with differently refined meshes (chosen in the `make_grid()` function)
1923 and get the following convergence rates of pressure,
1924 velocity, and flux (as defined in the introduction).
1925 
1926 <table align="center" class="doxtable">
1927  <tr>
1928  <th>number of refinements </th><th> @f$\|p-p_h^\circ\|@f$ </th><th> @f$\|\mathbf{u}-\mathbf{u}_h\|@f$ </th><th> @f$\|(\mathbf{u}-\mathbf{u}_h) \cdot \mathbf{n}\|@f$ </th>
1929  </tr>
1930  <tr>
1931  <td> 2 </td><td> 1.587e-01 </td><td> 5.113e-01 </td><td> 7.062e-01 </td>
1932  </tr>
1933  <tr>
1934  <td> 3 </td><td> 8.000e-02 </td><td> 2.529e-01 </td><td> 3.554e-01 </td>
1935  </tr>
1936  <tr>
1937  <td> 4 </td><td> 4.006e-02 </td><td> 1.260e-01 </td><td> 1.780e-01 </td>
1938  </tr>
1939  <tr>
1940  <td> 5 </td><td> 2.004e-02 </td><td> 6.297e-02 </td><td> 8.902e-02 </td>
1941  </tr>
1942  <tr>
1943  <th>Conv.rate </th><th> 1.00 </th><th> 1.00 </th><th> 1.00 </th>
1944  </tr>
1945 </table>
1946 
1947 We can see that the convergence rates of @f$\mbox{WG}(Q_0,Q_0;RT_{[0]})@f$ are around 1.
1948 This, of course, matches our theoretical expectations.
1949 
1950 
1951 <a name="TestresultsoniWGQsub1subQsub1subRTsub1subi"></a><h3>Test results on <i>WG(Q<sub>1</sub>,Q<sub>1</sub>;RT<sub>[1]</sub>)</i></h3>
1952 
1953 
1954 We can repeat the experiment from above using the next higher polynomial
1955 degree:
1956 The following figures are interior pressures and face pressures implemented using
1957 @f$\mbox{WG}(Q_1,Q_1;RT_{[1]})@f$. The mesh is refined 4 times. Compared to the
1958 previous figures using
1959 @f$\mbox{WG}(Q_0,Q_0;RT_{[0]})@f$, on each cell, the solution is no longer constant
1960 on each cell, as we now use bilinear polynomials to do the approximation.
1961 Consequently, there are 4 pressure values in one interior, 2 pressure values on
1962 each face.
1963 
1964 <table align="center">
1965  <tr>
1966  <td><img src="https://www.dealii.org/images/steps/developer/step-61.wg111_2d_4.png" alt=""></td>
1967  <td><img src="https://www.dealii.org/images/steps/developer/step-61.wg111_3d_4.png" alt=""></td>
1968  </tr>
1969 </table>
1970 
1971 Compared to the corresponding image for the @f$\mbox{WG}(Q_0,Q_0;RT_{[0]})@f$
1972 combination, the solution is now substantially more accurate and, in
1973 particular so close to being continuous at the interfaces that we can
1974 no longer distinguish the interface pressures @f$p^\partial@f$ from the
1975 interior pressures @f$p^\circ@f$ on the adjacent cells.
1976 
1977 <a name="Convergencetableforik1i"></a><h4>Convergence table for <i>k=1</i></h4>
1978 
1979 
1980 The following are the convergence rates of pressure, velocity, and flux
1981 we obtain from using the @f$\mbox{WG}(Q_1,Q_1;RT_{[1]})@f$ element combination:
1982 
1983 <table align="center" class="doxtable">
1984  <tr>
1985  <th>number of refinements </th><th> @f$\|p-p_h^\circ\|@f$ </th><th> @f$\|\mathbf{u}-\mathbf{u}_h\|@f$ </th><th> @f$\|(\mathbf{u}-\mathbf{u}_h) \cdot \mathbf{n}\|@f$ </th>
1986  </tr>
1987  <tr>
1988  <td> 2 </td><td> 1.613e-02 </td><td> 5.093e-02 </td><td> 7.167e-02 </td>
1989  </tr>
1990  <tr>
1991  <td> 3 </td><td> 4.056e-03 </td><td> 1.276e-02 </td><td> 1.802e-02 </td>
1992  </tr>
1993  <tr>
1994  <td> 4 </td><td> 1.015e-03 </td><td> 3.191e-03 </td><td> 4.512e-03 </td>
1995  </tr>
1996  <tr>
1997  <td> 5 </td><td> 2.540e-04 </td><td> 7.979e-04 </td><td> 1.128e-03 </td>
1998  </tr>
1999  <tr>
2000  <th>Conv.rate </th><th> 2.00 </th><th> 2.00 </th><th> 2.00 </th>
2001  </tr>
2002 </table>
2003 
2004 The convergence rates of @f$WG(Q_1,Q_1;RT_{[1]})@f$ are around 2, as expected.
2005 
2006 
2007 
2008 <a name="TestresultsoniWGQsub2subQsub2subRTsub2subi"></a><h3>Test results on <i>WG(Q<sub>2</sub>,Q<sub>2</sub>;RT<sub>[2]</sub>)</i></h3>
2009 
2010 
2011 Let us go one polynomial degree higher.
2012 The following are interior pressures and face pressures implemented using
2013 @f$WG(Q_2,Q_2;RT_{[2]})@f$, with mesh size @f$h = 1/32@f$ (i.e., 5 global mesh
2014 refinement steps). In the program, we use
2015 `data_out_face.build_patches(fe.degree)` when generating graphical output
2016 (see the documentation of DataOut::build_patches()), which here implies that
2017 we divide each 2d cell interior into 4 subcells in order to provide a better
2018 visualization of the quadratic polynomials.
2019 <table align="center">
2020  <tr>
2021  <td><img src="https://www.dealii.org/images/steps/developer/step-61.wg222_2d_5.png" alt=""></td>
2022  <td><img src="https://www.dealii.org/images/steps/developer/step-61.wg222_3d_5.png" alt=""></td>
2023  </tr>
2024 </table>
2025 
2026 
2027 <a name="Convergencetableforik2i"></a><h4>Convergence table for <i>k=2</i></h4>
2028 
2029 
2030 As before, we can generate convergence data for the
2031 @f$L_2@f$ errors of pressure, velocity, and flux
2032 using the @f$\mbox{WG}(Q_2,Q_2;RT_{[2]})@f$ combination:
2033 
2034 <table align="center" class="doxtable">
2035  <tr>
2036  <th>number of refinements </th><th> @f$\|p-p_h^\circ\|@f$ </th><th> @f$\|\mathbf{u}-\mathbf{u}_h\|@f$ </th><th> @f$\|(\mathbf{u}-\mathbf{u}_h) \cdot \mathbf{n}\|@f$ </th>
2037  </tr>
2038  <tr>
2039  <td> 2 </td><td> 1.072e-03 </td><td> 3.375e-03 </td><td> 4.762e-03 </td>
2040  </tr>
2041  <tr>
2042  <td> 3 </td><td> 1.347e-04 </td><td> 4.233e-04 </td><td> 5.982e-04 </td>
2043  </tr>
2044  <tr>
2045  <td> 4 </td><td> 1.685e-05 </td><td> 5.295e-05 </td><td> 7.487e-05 </td>
2046  </tr>
2047  <tr>
2048  <td> 5 </td><td> 2.107e-06 </td><td> 6.620e-06 </td><td> 9.362e-06 </td>
2049  </tr>
2050  <tr>
2051  <th>Conv.rate </th><th> 3.00 </th><th> 3.00 </th><th> 3.00 </th>
2052  </tr>
2053 </table>
2054 
2055 Once more, the convergence rates of @f$\mbox{WG}(Q_2,Q_2;RT_{[2]})@f$ is
2056 as expected, with values around 3.
2057  *
2058  *
2059 <a name="PlainProg"></a>
2060 <h1> The plain program</h1>
2061 @include "step-61.cc"
2062 */
std::vector< bool > component_mask
void add_data_vector(const VectorType &data, const std::vector< std::string > &names, const DataVectorType type=type_automatic, const std::vector< DataComponentInterpretation::DataComponentInterpretation > &data_component_interpretation={})
virtual void build_patches(const unsigned int n_subdivisions=0)
Definition: data_out.cc:1064
void reinit(const Triangulation< dim, spacedim > &tria)
const unsigned int dofs_per_cell
Definition: fe_values.h:2450
const SmartPointer< const FiniteElement< dim, spacedim >, FEValuesBase< dim, spacedim > > fe
Definition: fe_values.h:3963
void reinit(const TriaIterator< DoFCellAccessor< dim, spacedim, level_dof_access >> &cell)
const Quadrature< dim > quadrature
Definition: fe_values.h:4154
Definition: fe_dgq.h:111
void gauss_jordan()
virtual RangeNumberType value(const Point< dim > &p, const unsigned int component=0) const
Definition: point.h:111
virtual value_type value(const Point< dim > &p) const
VectorizedArray< Number, width > sin(const ::VectorizedArray< Number, width > &x)
DerivativeForm< 1, spacedim, dim, Number > transpose(const DerivativeForm< 1, dim, spacedim, Number > &DF)
@ update_values
Shape function values.
@ update_normal_vectors
Normal vectors.
@ update_JxW_values
Transformed quadrature weights.
@ update_gradients
Shape function gradients.
@ update_quadrature_points
Transformed quadrature points.
Point< 2 > first
Definition: grid_out.cc:4603
void write_vtu(std::ostream &out) const
typename ActiveSelector::active_cell_iterator active_cell_iterator
Definition: dof_handler.h:438
void loop(ITERATOR begin, typename identity< ITERATOR >::type end, DOFINFO &dinfo, INFOBOX &info, const std::function< void(DOFINFO &, typename INFOBOX::CellInfo &)> &cell_worker, const std::function< void(DOFINFO &, typename INFOBOX::CellInfo &)> &boundary_worker, const std::function< void(DOFINFO &, DOFINFO &, typename INFOBOX::CellInfo &, typename INFOBOX::CellInfo &)> &face_worker, ASSEMBLER &assembler, const LoopControl &lctrl=LoopControl())
Definition: loop.h:439
void make_sparsity_pattern(const DoFHandler< dim, spacedim > &dof_handler, SparsityPatternType &sparsity_pattern, const AffineConstraints< number > &constraints=AffineConstraints< number >(), const bool keep_constrained_dofs=true, const types::subdomain_id subdomain_id=numbers::invalid_subdomain_id)
void approximate(SynchronousIterators< std::tuple< typename DoFHandler< dim, spacedim >::active_cell_iterator, Vector< float >::iterator >> const &cell, const Mapping< dim, spacedim > &mapping, const DoFHandler< dim, spacedim > &dof_handler, const InputVector &solution, const unsigned int component)
void interpolate(const DoFHandler< dim, spacedim > &dof1, const InVector &u1, const DoFHandler< dim, spacedim > &dof2, OutVector &u2)
void hyper_cube(Triangulation< dim, spacedim > &tria, const double left=0., const double right=1., const bool colorize=false)
void refine(Triangulation< dim, spacedim > &tria, const Vector< Number > &criteria, const double threshold, const unsigned int max_to_mark=numbers::invalid_unsigned_int)
static const types::blas_int zero
@ matrix
Contents is actually a matrix.
static const char A
@ symmetric
Matrix is symmetric.
static const types::blas_int one
double norm(const FEValuesBase< dim > &fe, const ArrayView< const std::vector< Tensor< 1, dim >>> &Du)
Definition: divergence.h:472
Point< spacedim > point(const gp_Pnt &p, const double tolerance=1e-10)
Definition: utilities.cc:190
SymmetricTensor< 2, dim, Number > d(const Tensor< 2, dim, Number > &F, const Tensor< 2, dim, Number > &dF_dt)
Tensor< 2, dim, Number > w(const Tensor< 2, dim, Number > &F, const Tensor< 2, dim, Number > &dF_dt)
SymmetricTensor< 2, dim, Number > e(const Tensor< 2, dim, Number > &F)
Tensor< 2, dim, Number > l(const Tensor< 2, dim, Number > &F, const Tensor< 2, dim, Number > &dF_dt)
constexpr ReturnType< rank, T >::value_type & extract(T &t, const ArrayType &indices)
void call(const std::function< RT()> &function, internal::return_value< RT > &ret_val)
void integrate_difference(const Mapping< dim, spacedim > &mapping, const DoFHandler< dim, spacedim > &dof, const InVector &fe_function, const Function< spacedim, typename InVector::value_type > &exact_solution, OutVector &difference, const Quadrature< dim > &q, const NormType &norm, const Function< spacedim, double > *weight=nullptr, const double exponent=2.)
void interpolate_boundary_values(const Mapping< dim, spacedim > &mapping, const DoFHandler< dim, spacedim > &dof, const std::map< types::boundary_id, const Function< spacedim, number > * > &function_map, std::map< types::global_dof_index, number > &boundary_values, const ComponentMask &component_mask=ComponentMask())
void run(const Iterator &begin, const typename identity< Iterator >::type &end, Worker worker, Copier copier, const ScratchData &sample_scratch_data, const CopyData &sample_copy_data, const unsigned int queue_length, const unsigned int chunk_size)
Definition: work_stream.h:474
void copy(const T *begin, const T *end, U *dest)
int(&) functions(const void *v1, const void *v2)
void assemble(const MeshWorker::DoFInfoBox< dim, DOFINFO > &dinfo, A *assembler)
Definition: loop.h:71
void reinit(MatrixBlock< MatrixType > &v, const BlockSparsityPattern &p)
Definition: matrix_block.h:618
static constexpr double E
Definition: numbers.h:208
static constexpr double PI
Definition: numbers.h:233
::VectorizedArray< Number, width > cos(const ::VectorizedArray< Number, width > &)
::VectorizedArray< Number, width > sin(const ::VectorizedArray< Number, width > &)
::VectorizedArray< Number, width > sqrt(const ::VectorizedArray< Number, width > &)
const ::parallel::distributed::Triangulation< dim, spacedim > * triangulation