Reference documentation for deal.II version GIT relicensing-437-g81ec864850 2024-04-19 07:30:02+00:00
\(\newcommand{\dealvcentcolon}{\mathrel{\mathop{:}}}\) \(\newcommand{\dealcoloneq}{\dealvcentcolon\mathrel{\mkern-1.2mu}=}\) \(\newcommand{\jump}[1]{\left[\!\left[ #1 \right]\!\right]}\) \(\newcommand{\average}[1]{\left\{\!\left\{ #1 \right\}\!\right\}}\)
Loading...
Searching...
No Matches
The 'TRBDF2-DG projection solver for the incompressible Navier-Stokes equations' code gallery program

This program was contributed by Giuseppe Orlando <giuseppe.orlando@polimi.it>.
It comes without any warranty or support by its authors or the authors of deal.II.

This program is part of the deal.II code gallery and consists of the following files (click to inspect):

Pictures from this code gallery program

Annotated version of README.md

TRBDF2-DG projection solver for the incompressible Navier-Stokes equations

Compiling and Running

To generate a makefile for this code using CMake, type the following command into the terminal from the main directory:

    cmake . -DDEAL_II_DIR=/path/to/deal.II

To compile the code in release mode use:

    make release

This command will create the executable, NS_TRBDF2_DG.

To run the code on N processors type the following command into the terminal from the main directory,

    mpirun -np N ./NS_TRBDF2_DG

The output of the code will be in .vtu format and be written to disk in parallel. The results can be viewed using ParaView. A parameter file called parameter-file.prm has to be present in the same folder of the executable, following the same structure employed in step-35. Two extra fields are present: saving_directory with the name of the folder where the results should be saved (which has therefore to be created before launching the program) and refinement_iterations that specifies how often the remeshing procedure has to be performed.

The Navier-Stokes equations and the time discretization strategy

In this section, we briefly describe the problem and the approach employed. A detailed explanation of the numerical scheme is reported in [1]. We consider the classical unsteady incompressible Navier-Stokes equations, written in non-dimensional form as:

\begin{align*} \frac{\partial \mathbf{u}}{\partial t} + \nabla\cdot\left(\mathbf{u} \otimes\mathbf{u}\right) + \nabla p &= \frac{1}{Re}\Delta\mathbf{u} + \mathbf{f} \\ \nabla\cdot\mathbf{u} &= 0, \end{align*}

where \(Re\) denotes the Reynolds number. In the case of projection methods, difficulties arise in choosing the boundary conditions to be imposed for the Poisson equation which is to be solved at each time step to compute the pressure. An alternative that allows to avoid or reduce some of these problems is the so-called artificial compressibility formulation. In this formulation, the incompressibility constraint is relaxed and a time evolution equation for the pressure is introduced, which is characterized by an artificial sound speed \(c\), so as to obtain:

\begin{align*} \frac{\partial\mathbf{u}}{\partial t} + \nabla\cdot\left(\mathbf{u}\otimes\mathbf{u}\right) + \nabla p &= \frac{1}{Re}\Delta\mathbf{u} + \mathbf{f} \\ \frac{1}{c^2}\frac{\partial p}{\partial t} + \nabla\cdot\mathbf{u} &= 0. \end{align*}

For the sake of simplicity, we shall only consider \(\mathbf{f} = \mathbf{0}\). The numerical scheme is an extension of the projection method introduced in [2] based on the TR-BDF2 method. For a generic time-dependent problem \(\mathbf{u}' = \mathcal{N}(\mathbf{u})\), the TR-BDF2 method can be described in terms of two stages as follows:

\begin{align*} \frac{\mathbf{u}^{n+\gamma} - \mathbf{u}^{n}}{\gamma\Delta t} &= \frac{1}{2}\mathcal{N}\left(\mathbf{u}^{n+\gamma}\right) + \frac{1}{2}\mathcal{N}\left(\mathbf{u}^{n}\right) \\ \frac{\mathbf{u}^{n+1} - \mathbf{u}^{n + \gamma}}{\left(1 - \gamma\right)\Delta t} &= \frac{1}{2 - \gamma}\mathcal{N}\left(\mathbf{u}^{n+1}\right) + \frac{1 - \gamma}{2\left(2 - \gamma\right)}\mathcal{N}\left(\mathbf{u}^{n+\gamma}\right) + \frac{1 - \gamma}{2\left(2 - \gamma\right)}\mathcal{N}\left(\mathbf{u}^{n}\right). \end{align*}

Following then the projection approach described in [2], the momentum predictor equation for the first stage reads:

\begin{align*} &&\frac{\mathbf{u}^{n+\gamma,\ast} - \mathbf{u}^{n}}{\gamma\Delta t} - \frac{1}{2Re}\Delta\mathbf{u}^{n+\gamma,\ast} + \frac{1}{2}\nabla\cdot\left(\mathbf{u}^{n+\gamma,\ast}\otimes\mathbf{u}^{n+\frac{\gamma}{2}}\right) = \nonumber \\ &&\frac{1}{2Re}\Delta\mathbf{u}^{n} - \frac{1}{2}\nabla\cdot\left(\mathbf{u}^{n}\otimes\mathbf{u}^{n+\frac{\gamma}{2}}\right) - \nabla p^n \nonumber \\ &&\mathbf{u}^{n+\gamma,\ast}\rvert_{\partial\Omega} = \mathbf{u}_D^{n+\gamma}. \nonumber \end{align*}

Notice that, in order to avoid solving a nonlinear system at each time step, an approximation is introduced in the nonlinear momentum advection term, so that \(\mathbf{u}^{n + \frac{\gamma}{2}}\) is defined by extrapolation as

\begin{align*} \mathbf{u}^{n + \frac{\gamma}{2}} = \left(1 + \frac{\gamma}{2\left(1-\gamma\right)}\right)\mathbf{u}^{n} - \frac{\gamma}{2\left(1-\gamma\right)}\mathbf{u}^{n-1}. \end{align*}

For what concerns the pressure, we introduce the intermediate update \(\mathbf{u}^{n+\gamma,\ast\ast} = \mathbf{u}^{n+\gamma,\ast} + \gamma\Delta t\nabla p^{n}\), and we solve the following Helmholtz equation

\begin{align*} \frac{1}{c^2}\frac{p^{n+\gamma}}{\gamma^2\Delta t^2} -\Delta p^{n+\gamma} = - \frac{1}{\gamma\Delta t} \nabla\cdot\mathbf{u}^{n+\gamma,\ast\ast} + \frac{1}{c^2}\frac{p^{n }}{\gamma^2\Delta t^2} \end{align*}

and, finally, we set \(\mathbf{u}^{n+\gamma} = \mathbf{u}^{n+\gamma,\ast\ast} - \gamma\Delta t\nabla p^{n+\gamma}\). The second stage of the TR-BDF2 scheme is performed in a similar manner (see [1] for all the details).

Some implementation details

A matrix-free approach was employed like for step-37 or step-50. Another feature of the library which it is possible to employ during the numerical simulations is the mesh adaptation capability. On each element \(K\) we define the quantity

\[ \eta_K = \text{diam}(K)^2\left\|\nabla \times \mathbf{u}\right\|^2_K \]

that acts as local refinement indicator. The preconditioned conjugate gradient method implemented in the function SolverCG was employed to solve the Helmholtz equations, whereas, for the momentum equations, the GMRES solver implemented in the function SolverGMRES was used. A Jacobi preconditioner is used for the two momentum predictors, whereas a Geometric Multigrid preconditioner is employed for the Helmholtz equations (see step-37).

Test case

We test the code with a classical benchmark case, namely the flow past a cylinder in 2D at \(Re = 100\) (see [1] for all the details). The image shows the contour plot of the velocity magnitude at \(t = T_{f} = 400\). The evolution of the lift and drag coefficients from \(t = 385\) to \(t = T_{f}\) are also reported and the expected periodic behaviour is retrieved.

contour

drag lift adaptive grid

References

[1] G. Orlando, A. Della Rocca, P.F. Barbante, L. Bonaventura, and N. Parolini. An efficient and accurate implicit DG solver for the incompressible Navier-Stokes equations. International Journal for Numerical Methods in Fluids, 2022. DOI: 10.1002/FLD.5098

[2] A. Della Rocca. Large-Eddy Simulations of Turbulent Reacting Flows with Industrial Applications. PhD thesis. Politecnico di Milano, 2018. http://hdl.handle.net/10589/137775

Annotated version of equation_data.h

  /* -----------------------------------------------------------------------------
  *
  * SPDX-License-Identifier: LGPL-2.1-or-later
  * Copyright (C) 2022 by Giuseppe Orlando
  *
  * This file is part of the deal.II code gallery.
  *
  * -----------------------------------------------------------------------------
  */
 

We start by including all the necessary deal.II header files.

  #include <deal.II/base/point.h>
  #include <deal.II/base/function.h>
 

@sect{Equation data}

In the next namespace, we declare and implement suitable functions that may be used for the initial and boundary conditions

  namespace EquationData {
  using namespace dealii;
 
  static const unsigned int degree_p = 1; /*--- Polynomial degree for the pressure. The one for the velocity
  will be equal to degree_p + 1, but its value can be easily changed
  in the template parameter list of the classes with weak form ---*/
 

We declare class that describes the boundary conditions and initial one for velocity:

  template<int dim>
  class Velocity: public Function<dim> {
  public:
  Velocity(const double initial_time = 0.0);
 
  virtual double value(const Point<dim>& p,
  const unsigned int component = 0) const override;
 
  virtual void vector_value(const Point<dim>& p,
  Vector<double>& values) const override;
  };
 
 
  template<int dim>
  Velocity<dim>::Velocity(const double initial_time): Function<dim>(dim, initial_time) {}
 
 
  template<int dim>
  double Velocity<dim>::value(const Point<dim>& p, const unsigned int component) const {
  AssertIndexRange(component, 3);
  if(component == 0) {
  const double Um = 1.5;
  const double H = 4.1;
 
  return 4.0*Um*p(1)*(H - p(1))/(H*H);
  }
  else
  return 0.0;
  }
 
 
  template<int dim>
  void Velocity<dim>::vector_value(const Point<dim>& p, Vector<double>& values) const {
  Assert(values.size() == dim, ExcDimensionMismatch(values.size(), dim));
 
  for(unsigned int i = 0; i < dim; ++i)
  values[i] = value(p, i);
  }
 
 
virtual RangeNumberType value(const Point< dim > &p, const unsigned int component=0) const
virtual void vector_value(const Point< dim > &p, Vector< RangeNumberType > &values) const
Definition point.h:111
#define Assert(cond, exc)
#define AssertIndexRange(index, range)
spacedim const Point< spacedim > & p
Definition grid_tools.h:990

We do the same for the pressure

  template<int dim>
  class Pressure: public Function<dim> {
  public:
  Pressure(const double initial_time = 0.0);
 
  virtual double value(const Point<dim>& p,
  const unsigned int component = 0) const override;
  };
 
 
  template<int dim>
  Pressure<dim>::Pressure(const double initial_time): Function<dim>(1, initial_time) {}
 
 
  template<int dim>
  double Pressure<dim>::value(const Point<dim>& p, const unsigned int component) const {
  (void)component;
  AssertIndexRange(component, 1);
 
  return 22.0 - p(0);
  }
 
  } // namespace EquationData

Annotated version of navier_stokes_TRBDF2_DG.cc

  /* -----------------------------------------------------------------------------
  *
  * SPDX-License-Identifier: LGPL-2.1-or-later
  * Copyright (C) 2022 by Giuseppe Orlando
  *
  * This file is part of the deal.II code gallery.
  *
  * -----------------------------------------------------------------------------
  *
  * Author: Giuseppe Orlando, 2022.
  */
 

We start by including all the necessary deal.II header files and some C++ related ones.

  #include <deal.II/base/quadrature_lib.h>
  #include <deal.II/base/multithread_info.h>
  #include <deal.II/base/thread_management.h>
  #include <deal.II/base/work_stream.h>
  #include <deal.II/base/parallel.h>
  #include <deal.II/base/utilities.h>
  #include <deal.II/base/conditional_ostream.h>
 
  #include <deal.II/lac/vector.h>
  #include <deal.II/lac/solver_cg.h>
  #include <deal.II/lac/precondition.h>
  #include <deal.II/lac/solver_gmres.h>
  #include <deal.II/lac/affine_constraints.h>
 
  #include <deal.II/grid/tria.h>
  #include <deal.II/grid/grid_generator.h>
  #include <deal.II/grid/grid_tools.h>
  #include <deal.II/grid/grid_refinement.h>
  #include <deal.II/grid/tria_accessor.h>
  #include <deal.II/grid/tria_iterator.h>
  #include <deal.II/distributed/grid_refinement.h>
 
  #include <deal.II/dofs/dof_handler.h>
  #include <deal.II/dofs/dof_accessor.h>
  #include <deal.II/dofs/dof_tools.h>
 
  #include <deal.II/fe/fe_q.h>
  #include <deal.II/fe/fe_dgq.h>
  #include <deal.II/fe/fe_values.h>
  #include <deal.II/fe/fe_tools.h>
  #include <deal.II/fe/fe_system.h>
 
  #include <deal.II/numerics/matrix_tools.h>
  #include <deal.II/numerics/vector_tools.h>
  #include <deal.II/numerics/data_out.h>
 
  #include <fstream>
  #include <cmath>
  #include <iostream>
 
  #include <deal.II/matrix_free/matrix_free.h>
  #include <deal.II/matrix_free/operators.h>
  #include <deal.II/matrix_free/fe_evaluation.h>
  #include <deal.II/fe/component_mask.h>
 
  #include <deal.II/base/timer.h>
  #include <deal.II/distributed/solution_transfer.h>
  #include <deal.II/numerics/error_estimator.h>
 
  #include <deal.II/multigrid/multigrid.h>
  #include <deal.II/multigrid/mg_transfer_matrix_free.h>
  #include <deal.II/multigrid/mg_tools.h>
  #include <deal.II/multigrid/mg_coarse.h>
  #include <deal.II/multigrid/mg_smoother.h>
  #include <deal.II/multigrid/mg_matrix.h>
 
  #include <deal.II/meshworker/mesh_loop.h>
 
  #include "runtime_parameters.h"
  #include "equation_data.h"
 
  namespace MatrixFreeTools {
  using namespace dealii;
 
  template<int dim, typename Number, typename VectorizedArrayType>
  const std::function<void(const MatrixFree<dim, Number, VectorizedArrayType>&,
  const unsigned int&,
  const std::pair<unsigned int, unsigned int>&)>& cell_operation,
  const std::function<void(const MatrixFree<dim, Number, VectorizedArrayType>&,
  const unsigned int&,
  const std::pair<unsigned int, unsigned int>&)>& face_operation,
  const std::function<void(const MatrixFree<dim, Number, VectorizedArrayType>&,
  const unsigned int&,
  const std::pair<unsigned int, unsigned int>&)>& boundary_operation,
  const unsigned int dof_no = 0) {
void compute_diagonal(const MatrixFree< dim, Number, VectorizedArrayType > &matrix_free, VectorType &diagonal_global, const std::function< void(FEEvaluation< dim, fe_degree, n_q_points_1d, n_components, Number, VectorizedArrayType > &)> &local_vmult, const unsigned int dof_no=0, const unsigned int quad_no=0, const unsigned int first_selected_component=0)

initialize vector

  matrix_free.initialize_dof_vector(diagonal_global, dof_no);
 
  const unsigned int dummy = 0;
 
  matrix_free.loop(cell_operation, face_operation, boundary_operation,
  diagonal_global, dummy, false,
  }
  }
 
void loop(const std::function< void(const MatrixFree< dim, Number, VectorizedArrayType > &, OutVector &, const InVector &, const std::pair< unsigned int, unsigned int > &)> &cell_operation, const std::function< void(const MatrixFree< dim, Number, VectorizedArrayType > &, OutVector &, const InVector &, const std::pair< unsigned int, unsigned int > &)> &face_operation, const std::function< void(const MatrixFree< dim, Number, VectorizedArrayType > &, OutVector &, const InVector &, const std::pair< unsigned int, unsigned int > &)> &boundary_operation, OutVector &dst, const InVector &src, const bool zero_dst_vector=false, const DataAccessOnFaces dst_vector_face_access=DataAccessOnFaces::unspecified, const DataAccessOnFaces src_vector_face_access=DataAccessOnFaces::unspecified) const
void initialize_dof_vector(VectorType &vec, const unsigned int dof_handler_index=0) const

We include the code in a suitable namespace:

  namespace NS_TRBDF2 {
  using namespace dealii;
 

The following class is an auxiliary one for post-processing of the vorticity

  template<int dim>
  class PostprocessorVorticity: public DataPostprocessor<dim> {
  public:
  std::vector<Vector<double>>& computed_quantities) const override;
 
  virtual std::vector<std::string> get_names() const override;
 
  virtual std::vector<DataComponentInterpretation::DataComponentInterpretation>
 
  virtual UpdateFlags get_needed_update_flags() const override;
  };
 
virtual UpdateFlags get_needed_update_flags() const =0
virtual void evaluate_vector_field(const DataPostprocessorInputs::Vector< dim > &input_data, std::vector< Vector< double > > &computed_quantities) const
virtual std::vector< std::string > get_names() const =0
virtual std::vector< DataComponentInterpretation::DataComponentInterpretation > get_data_component_interpretation() const
UpdateFlags

This function evaluates the vorticty in both 2D and 3D cases

  template <int dim>
  void PostprocessorVorticity<dim>::evaluate_vector_field(const DataPostprocessorInputs::Vector<dim>& inputs,
  std::vector<Vector<double>>& computed_quantities) const {
  const unsigned int n_quadrature_points = inputs.solution_values.size();
 
  /*--- Check the correctness of all data structres ---*/
  Assert(inputs.solution_gradients.size() == n_quadrature_points, ExcInternalError());
  Assert(computed_quantities.size() == n_quadrature_points, ExcInternalError());
 
  Assert(inputs.solution_values[0].size() == dim, ExcInternalError());
 
  if(dim == 2) {
  Assert(computed_quantities[0].size() == 1, ExcInternalError());
  }
  else {
  Assert(computed_quantities[0].size() == dim, ExcInternalError());
  }
 
  /*--- Compute the vorticty ---*/
  if(dim == 2) {
  for(unsigned int q = 0; q < n_quadrature_points; ++q)
  computed_quantities[q](0) = inputs.solution_gradients[q][1][0] - inputs.solution_gradients[q][0][1];
  }
  else {
  for(unsigned int q = 0; q < n_quadrature_points; ++q) {
  computed_quantities[q](0) = inputs.solution_gradients[q][2][1] - inputs.solution_gradients[q][1][2];
  computed_quantities[q](1) = inputs.solution_gradients[q][0][2] - inputs.solution_gradients[q][2][0];
  computed_quantities[q](2) = inputs.solution_gradients[q][1][0] - inputs.solution_gradients[q][0][1];
  }
  }
  }
 
std::vector<::Vector< double > > solution_values
std::vector< std::vector< Tensor< 1, spacedim > > > solution_gradients

This auxiliary function is required by the base class DataProcessor and simply sets the name for the output file

  template<int dim>
  std::vector<std::string> PostprocessorVorticity<dim>::get_names() const {
  std::vector<std::string> names;
  names.emplace_back("vorticity");
  if(dim == 3) {
  names.emplace_back("vorticity");
  names.emplace_back("vorticity");
  }
 
  return names;
  }
 

This auxiliary function is required by the base class DataProcessor and simply specifies if the vorticity is a scalar (2D) or a vector (3D)

  template<int dim>
  std::vector<DataComponentInterpretation::DataComponentInterpretation>
  PostprocessorVorticity<dim>::get_data_component_interpretation() const {
  std::vector<DataComponentInterpretation::DataComponentInterpretation> interpretation;
  if(dim == 2)
  else {
  }
 
  return interpretation;
  }
 

This auxiliary function is required by the base class DataProcessor and simply sets which variables have to updated (only the gradients)

  template<int dim>
  UpdateFlags PostprocessorVorticity<dim>::get_needed_update_flags() const {
  return update_gradients;
  }
 
 
@ update_gradients
Shape function gradients.

The following structs are auxiliary objects for mesh refinement. ScratchData simply sets the FEValues object

  template <int dim>
  struct ScratchData {
  ScratchData(const FiniteElement<dim>& fe,
  const unsigned int quadrature_degree,
  const UpdateFlags update_flags): fe_values(fe, QGauss<dim>(quadrature_degree), update_flags) {}
 
  ScratchData(const ScratchData<dim>& scratch_data): fe_values(scratch_data.fe_values.get_fe(),
  scratch_data.fe_values.get_quadrature(),
  scratch_data.fe_values.get_update_flags()) {}
  FEValues<dim> fe_values;
  };
 
 

CopyData simply sets the cell index

  struct CopyData {
  CopyData() : cell_index(numbers::invalid_unsigned_int), value(0.0) {}
 
  CopyData(const CopyData &) = default;
 
  unsigned int cell_index;
  double value;
  };
 
 
unsigned int cell_index

@sect{ NavierStokesProjectionOperator::NavierStokesProjectionOperator }

The following class sets effecively the weak formulation of the problems for the different stages and for both velocity and pressure. The template parameters are the dimnesion of the problem, the polynomial degree for the pressure, the polynomial degree for the velocity, the number of quadrature points for integrals for the pressure step, the number of quadrature points for integrals for the velocity step, the type of vector for storage and the type of floating point data (in general double or float for preconditioners structures if desired).

  template<int dim, int fe_degree_p, int fe_degree_v, int n_q_points_1d_p, int n_q_points_1d_v, typename Vec>
  class NavierStokesProjectionOperator: public MatrixFreeOperators::Base<dim, Vec> {
  public:
  using Number = typename Vec::value_type;
 
  NavierStokesProjectionOperator();
 
  NavierStokesProjectionOperator(RunTimeParameters::Data_Storage& data);
 
  void set_dt(const double time_step);
 
  void set_TR_BDF2_stage(const unsigned int stage);
 
  void set_NS_stage(const unsigned int stage);
 
  void set_u_extr(const Vec& src);
 
  void vmult_rhs_velocity(Vec& dst, const std::vector<Vec>& src) const;
 
  void vmult_rhs_pressure(Vec& dst, const std::vector<Vec>& src) const;
 
  void vmult_grad_p_projection(Vec& dst, const Vec& src) const;
 
  virtual void compute_diagonal() override;
 
  protected:
  double Re;
  double dt;
 
  /*--- Parameters of time-marching scheme ---*/
  double gamma;
  double a31;
  double a32;
  double a33;
 
  unsigned int TR_BDF2_stage; /*--- Flag to denote at which stage of the TR-BDF2 are ---*/
  unsigned int NS_stage; /*--- Flag to denote at which stage of NS solution inside each TR-BDF2 stage we are
  (solution of the velocity or of the pressure)---*/
 
  virtual void apply_add(Vec& dst, const Vec& src) const override;
 
  private:
  /*--- Auxiliary variable for the TR stage
  (just to avoid to report a lot of 0.5 and for my personal choice to be coherent with the article) ---*/
  const double a21 = 0.5;
  const double a22 = 0.5;
 
  /*--- Penalty method parameters, theta = 1 means SIP, while C_p and C_u are the penalization coefficients ---*/
  const double theta_v = 1.0;
  const double theta_p = 1.0;
  const double C_p = 1.0*(fe_degree_p + 1)*(fe_degree_p + 1);
  const double C_u = 1.0*(fe_degree_v + 1)*(fe_degree_v + 1);
 
  Vec u_extr; /*--- Auxiliary variable to update the extrapolated velocity ---*/
 
  EquationData::Velocity<dim> vel_boundary_inflow; /*--- Auxiliary variable to impose velocity boundary conditions ---*/
 
  /*--- The following functions basically assemble the linear and bilinear forms. Their syntax is due to
  the base class MatrixFreeOperators::Base ---*/
  void assemble_rhs_cell_term_velocity(const MatrixFree<dim, Number>& data,
  Vec& dst,
  const std::vector<Vec>& src,
  const std::pair<unsigned int, unsigned int>& cell_range) const;
  void assemble_rhs_face_term_velocity(const MatrixFree<dim, Number>& data,
  Vec& dst,
  const std::vector<Vec>& src,
  const std::pair<unsigned int, unsigned int>& face_range) const;
  void assemble_rhs_boundary_term_velocity(const MatrixFree<dim, Number>& data,
  Vec& dst,
  const std::vector<Vec>& src,
  const std::pair<unsigned int, unsigned int>& face_range) const;
 
  void assemble_rhs_cell_term_pressure(const MatrixFree<dim, Number>& data,
  Vec& dst,
  const std::vector<Vec>& src,
  const std::pair<unsigned int, unsigned int>& cell_range) const;
  void assemble_rhs_face_term_pressure(const MatrixFree<dim, Number>& data,
  Vec& dst,
  const std::vector<Vec>& src,
  const std::pair<unsigned int, unsigned int>& face_range) const;
  void assemble_rhs_boundary_term_pressure(const MatrixFree<dim, Number>& data,
  Vec& dst,
  const std::vector<Vec>& src,
  const std::pair<unsigned int, unsigned int>& face_range) const;
 
  void assemble_cell_term_velocity(const MatrixFree<dim, Number>& data,
  Vec& dst,
  const Vec& src,
  const std::pair<unsigned int, unsigned int>& cell_range) const;
  void assemble_face_term_velocity(const MatrixFree<dim, Number>& data,
  Vec& dst,
  const Vec& src,
  const std::pair<unsigned int, unsigned int>& face_range) const;
  void assemble_boundary_term_velocity(const MatrixFree<dim, Number>& data,
  Vec& dst,
  const Vec& src,
  const std::pair<unsigned int, unsigned int>& face_range) const;
 
  void assemble_cell_term_pressure(const MatrixFree<dim, Number>& data,
  Vec& dst,
  const Vec& src,
  const std::pair<unsigned int, unsigned int>& cell_range) const;
  void assemble_face_term_pressure(const MatrixFree<dim, Number>& data,
  Vec& dst,
  const Vec& src,
  const std::pair<unsigned int, unsigned int>& face_range) const;
  void assemble_boundary_term_pressure(const MatrixFree<dim, Number>& data,
  Vec& dst,
  const Vec& src,
  const std::pair<unsigned int, unsigned int>& face_range) const;
 
  void assemble_cell_term_projection_grad_p(const MatrixFree<dim, Number>& data,
  Vec& dst,
  const Vec& src,
  const std::pair<unsigned int, unsigned int>& cell_range) const;
  void assemble_rhs_cell_term_projection_grad_p(const MatrixFree<dim, Number>& data,
  Vec& dst,
  const Vec& src,
  const std::pair<unsigned int, unsigned int>& cell_range) const;
 
  void assemble_diagonal_cell_term_velocity(const MatrixFree<dim, Number>& data,
  Vec& dst,
  const unsigned int& src,
  const std::pair<unsigned int, unsigned int>& cell_range) const;
  void assemble_diagonal_face_term_velocity(const MatrixFree<dim, Number>& data,
  Vec& dst,
  const unsigned int& src,
  const std::pair<unsigned int, unsigned int>& face_range) const;
  void assemble_diagonal_boundary_term_velocity(const MatrixFree<dim, Number>& data,
  Vec& dst,
  const unsigned int& src,
  const std::pair<unsigned int, unsigned int>& face_range) const;
 
  void assemble_diagonal_cell_term_pressure(const MatrixFree<dim, Number>& data,
  Vec& dst,
  const unsigned int& src,
  const std::pair<unsigned int, unsigned int>& cell_range) const;
  void assemble_diagonal_face_term_pressure(const MatrixFree<dim, Number>& data,
  Vec& dst,
  const unsigned int& src,
  const std::pair<unsigned int, unsigned int>& face_range) const;
  void assemble_diagonal_boundary_term_pressure(const MatrixFree<dim, Number>& data,
  Vec& dst,
  const unsigned int& src,
  const std::pair<unsigned int, unsigned int>& face_range) const;
  };
 
 
virtual void compute_diagonal()=0
virtual void apply_add(VectorType &dst, const VectorType &src) const =0

We start with the default constructor. It is important for MultiGrid, so it is fundamental to properly set the parameters of the time scheme.

  template<int dim, int fe_degree_p, int fe_degree_v, int n_q_points_1d_p, int n_q_points_1d_v, typename Vec>
  NavierStokesProjectionOperator<dim, fe_degree_p, fe_degree_v, n_q_points_1d_p, n_q_points_1d_v, Vec>::
  NavierStokesProjectionOperator():
  MatrixFreeOperators::Base<dim, Vec>(), Re(), dt(), gamma(2.0 - std::sqrt(2.0)), a31((1.0 - gamma)/(2.0*(2.0 - gamma))),
  a32(a31), a33(1.0/(2.0 - gamma)), TR_BDF2_stage(1), NS_stage(1), u_extr() {}
 
 
STL namespace.

We focus now on the constructor with runtime parameters storage

  template<int dim, int fe_degree_p, int fe_degree_v, int n_q_points_1d_p, int n_q_points_1d_v, typename Vec>
  NavierStokesProjectionOperator<dim, fe_degree_p, fe_degree_v, n_q_points_1d_p, n_q_points_1d_v, Vec>::
  NavierStokesProjectionOperator(RunTimeParameters::Data_Storage& data):
  MatrixFreeOperators::Base<dim, Vec>(), Re(data.Reynolds), dt(data.dt),
  gamma(2.0 - std::sqrt(2.0)), a31((1.0 - gamma)/(2.0*(2.0 - gamma))),
  a32(a31), a33(1.0/(2.0 - gamma)), TR_BDF2_stage(1), NS_stage(1), u_extr(),
  vel_boundary_inflow(data.initial_time) {}
 
 

Setter of time-step (called by Multigrid and in case a smaller time-step towards the end is needed)

  template<int dim, int fe_degree_p, int fe_degree_v, int n_q_points_1d_p, int n_q_points_1d_v, typename Vec>
  void NavierStokesProjectionOperator<dim, fe_degree_p, fe_degree_v, n_q_points_1d_p, n_q_points_1d_v, Vec>::
  set_dt(const double time_step) {
  dt = time_step;
  }
 
 

Setter of TR-BDF2 stage (this can be known only during the effective execution and so it has to be demanded to the class that really solves the problem)

  template<int dim, int fe_degree_p, int fe_degree_v, int n_q_points_1d_p, int n_q_points_1d_v, typename Vec>
  void NavierStokesProjectionOperator<dim, fe_degree_p, fe_degree_v, n_q_points_1d_p, n_q_points_1d_v, Vec>::
  set_TR_BDF2_stage(const unsigned int stage) {
  AssertIndexRange(stage, 3);
  Assert(stage > 0, ExcInternalError());
 
  TR_BDF2_stage = stage;
  }
 
 

Setter of NS stage (this can be known only during the effective execution and so it has to be demanded to the class that really solves the problem)

  template<int dim, int fe_degree_p, int fe_degree_v, int n_q_points_1d_p, int n_q_points_1d_v, typename Vec>
  void NavierStokesProjectionOperator<dim, fe_degree_p, fe_degree_v, n_q_points_1d_p, n_q_points_1d_v, Vec>::
  set_NS_stage(const unsigned int stage) {
  AssertIndexRange(stage, 4);
  Assert(stage > 0, ExcInternalError());
 
  NS_stage = stage;
  }
 
 

Setter of extrapolated velocity for different stages

  template<int dim, int fe_degree_p, int fe_degree_v, int n_q_points_1d_p, int n_q_points_1d_v, typename Vec>
  void NavierStokesProjectionOperator<dim, fe_degree_p, fe_degree_v, n_q_points_1d_p, n_q_points_1d_v, Vec>::
  set_u_extr(const Vec& src) {
  u_extr = src;
  u_extr.update_ghost_values();
  }
 
 

We are in a DG-MatrixFree framework, so it is convenient to compute separately cell contribution, internal faces contributions and boundary faces contributions. We start by assembling the rhs cell term for the velocity.

  template<int dim, int fe_degree_p, int fe_degree_v, int n_q_points_1d_p, int n_q_points_1d_v, typename Vec>
  void NavierStokesProjectionOperator<dim, fe_degree_p, fe_degree_v, n_q_points_1d_p, n_q_points_1d_v, Vec>::
  assemble_rhs_cell_term_velocity(const MatrixFree<dim, Number>& data,
  Vec& dst,
  const std::vector<Vec>& src,
  const std::pair<unsigned int, unsigned int>& cell_range) const {
  if(TR_BDF2_stage == 1) {
  /*--- We first start by declaring the suitable instances to read the old velocity, the
  extrapolated velocity and the old pressure. 'phi' will be used only to submit the result.
  The second argument specifies which dof handler has to be used (in this implementation 0 stands for
  velocity and 1 for pressure). ---*/
  phi_old(data, 0),
  phi_old_extr(data, 0);
 
  /*--- We loop over the cells in the range ---*/
  for(unsigned int cell = cell_range.first; cell < cell_range.second; ++cell) {
  /*--- Now we need to assign the current cell to each FEEvaluation object and then to specify which src vector
  it has to read (the proper order is clearly delegated to the user, which has to pay attention in the function
  call to be coherent). ---*/
  phi_old.reinit(cell);
  phi_old.gather_evaluate(src[0], EvaluationFlags::values | EvaluationFlags::gradients);
  /*--- The 'gather_evaluate' function reads data from the vector.
  The second and third parameter specifies if you want to read
  values and/or derivative related quantities ---*/
  phi_old_extr.reinit(cell);
  phi_old_extr.gather_evaluate(src[1], EvaluationFlags::values);
  phi_old_press.reinit(cell);
  phi_old_press.gather_evaluate(src[2], EvaluationFlags::values);
  phi.reinit(cell);
 
  /*--- Now we loop over all the quadrature points to compute the integrals ---*/
  for(unsigned int q = 0; q < phi.n_q_points; ++q) {
  const auto& u_n = phi_old.get_value(q);
  const auto& grad_u_n = phi_old.get_gradient(q);
  const auto& u_n_gamma_ov_2 = phi_old_extr.get_value(q);
  const auto& tensor_product_u_n = outer_product(u_n, u_n_gamma_ov_2);
  const auto& p_n = phi_old_press.get_value(q);
  auto p_n_times_identity = tensor_product_u_n;
  p_n_times_identity = 0;
  for(unsigned int d = 0; d < dim; ++d)
  p_n_times_identity[d][d] = p_n;
 
  phi.submit_value(1.0/(gamma*dt)*u_n, q); /*--- 'submit_value' contains quantites that we want to test against the
  test function ---*/
  phi.submit_gradient(-a21/Re*grad_u_n + a21*tensor_product_u_n + p_n_times_identity, q);
  /*--- 'submit_gradient' contains quantites that we want to test against the gradient of test function ---*/
  }
  phi.integrate_scatter(EvaluationFlags::values | EvaluationFlags::gradients, dst);
  /*--- 'integrate_scatter' is the responsible of distributing into dst.
  The flag parameter specifies if we are testing against the test function and/or its gradient ---*/
  }
  }
  else {
  /*--- We first start by declaring the suitable instances to read already available quantities. ---*/
  phi_old(data, 0),
  phi_int(data, 0);
 
  /*--- We loop over the cells in the range ---*/
  for(unsigned int cell = cell_range.first; cell < cell_range.second; ++cell) {
  phi_old.reinit(cell);
  phi_old.gather_evaluate(src[0], EvaluationFlags::values | EvaluationFlags::gradients);
  phi_int.reinit(cell);
  phi_int.gather_evaluate(src[1], EvaluationFlags::values | EvaluationFlags::gradients);
  phi_old_press.reinit(cell);
  phi_old_press.gather_evaluate(src[2], EvaluationFlags::values);
  phi.reinit(cell);
 
  /*--- Now we loop over all the quadrature points to compute the integrals ---*/
  for(unsigned int q = 0; q < phi.n_q_points; ++q) {
  const auto& u_n = phi_old.get_value(q);
  const auto& grad_u_n = phi_old.get_gradient(q);
  const auto& u_n_gamma = phi_int.get_value(q);
  const auto& grad_u_n_gamma = phi_int.get_gradient(q);
  const auto& tensor_product_u_n = outer_product(u_n, u_n);
  const auto& tensor_product_u_n_gamma = outer_product(u_n_gamma, u_n_gamma);
  const auto& p_n = phi_old_press.get_value(q);
  auto p_n_times_identity = tensor_product_u_n;
  p_n_times_identity = 0;
  for(unsigned int d = 0; d < dim; ++d)
  p_n_times_identity[d][d] = p_n;
 
  phi.submit_value(1.0/((1.0 - gamma)*dt)*u_n_gamma, q);
  phi.submit_gradient(a32*tensor_product_u_n_gamma + a31*tensor_product_u_n -
  a32/Re*grad_u_n_gamma - a31/Re*grad_u_n + p_n_times_identity, q);
  }
  phi.integrate_scatter(EvaluationFlags::values | EvaluationFlags::gradients, dst);
  }
  }
  }
 
 
SymmetricTensor< 2, dim, Number > d(const Tensor< 2, dim, Number > &F, const Tensor< 2, dim, Number > &dF_dt)
DEAL_II_HOST constexpr SymmetricTensor< 4, dim, Number > outer_product(const SymmetricTensor< 2, dim, Number > &t1, const SymmetricTensor< 2, dim, Number > &t2)

The followinf function assembles rhs face term for the velocity

  template<int dim, int fe_degree_p, int fe_degree_v, int n_q_points_1d_p, int n_q_points_1d_v, typename Vec>
  void NavierStokesProjectionOperator<dim, fe_degree_p, fe_degree_v, n_q_points_1d_p, n_q_points_1d_v, Vec>::
  assemble_rhs_face_term_velocity(const MatrixFree<dim, Number>& data,
  Vec& dst,
  const std::vector<Vec>& src,
  const std::pair<unsigned int, unsigned int>& face_range) const {
  if(TR_BDF2_stage == 1) {
  /*--- We first start by declaring the suitable instances to read already available quantities. In this case
  we are at the face between two elements and this is the reason of 'FEFaceEvaluation'. It contains an extra
  input argument, the second one, that specifies if it is from 'interior' or not---*/
  phi_m(data, false, 0),
  phi_old_p(data, true, 0),
  phi_old_m(data, false, 0),
  phi_old_extr_p(data, true, 0),
  phi_old_extr_m(data, false, 0);
  phi_old_press_m(data, false, 1);
 
  /*--- We loop over the faces in the range ---*/
  for(unsigned int face = face_range.first; face < face_range.second; ++face) {
  phi_old_p.reinit(face);
  phi_old_p.gather_evaluate(src[0], EvaluationFlags::values | EvaluationFlags::gradients);
  phi_old_m.reinit(face);
  phi_old_m.gather_evaluate(src[0], EvaluationFlags::values | EvaluationFlags::gradients);
  phi_old_extr_p.reinit(face);
  phi_old_extr_p.gather_evaluate(src[1], EvaluationFlags::values);
  phi_old_extr_m.reinit(face);
  phi_old_extr_m.gather_evaluate(src[1], EvaluationFlags::values);
  phi_old_press_p.reinit(face);
  phi_old_press_p.gather_evaluate(src[2], EvaluationFlags::values);
  phi_old_press_m.reinit(face);
  phi_old_press_m.gather_evaluate(src[2], EvaluationFlags::values);
  phi_p.reinit(face);
  phi_m.reinit(face);
 
  /*--- Now we loop over all the quadrature points to compute the integrals ---*/
  for(unsigned int q = 0; q < phi_p.n_q_points; ++q) {
  const auto& n_plus = phi_p.get_normal_vector(q); /*--- The normal vector is the same
  for both phi_p and phi_m. If the face is interior,
  it correspond to the outer normal ---*/
 
  const auto& avg_grad_u_old = 0.5*(phi_old_p.get_gradient(q) + phi_old_m.get_gradient(q));
  const auto& avg_tensor_product_u_n = 0.5*(outer_product(phi_old_p.get_value(q), phi_old_extr_p.get_value(q)) +
  outer_product(phi_old_m.get_value(q), phi_old_extr_m.get_value(q)));
  const auto& avg_p_old = 0.5*(phi_old_press_p.get_value(q) + phi_old_press_m.get_value(q));
 
  phi_p.submit_value((a21/Re*avg_grad_u_old - a21*avg_tensor_product_u_n)*n_plus - avg_p_old*n_plus, q);
  phi_m.submit_value(-(a21/Re*avg_grad_u_old - a21*avg_tensor_product_u_n)*n_plus + avg_p_old*n_plus, q);
  }
  phi_p.integrate_scatter(EvaluationFlags::values, dst);
  phi_m.integrate_scatter(EvaluationFlags::values, dst);
  }
  }
  else {
  /*--- We first start by declaring the suitable instances to read already available quantities. ---*/
  phi_m(data, false, 0),
  phi_old_p(data, true, 0),
  phi_old_m(data, false, 0),
  phi_int_p(data, true, 0),
  phi_int_m(data, false, 0);
  phi_old_press_m(data, false, 1);
 
  /*--- We loop over the faces in the range ---*/
  for(unsigned int face = face_range.first; face < face_range.second; ++ face) {
  phi_old_p.reinit(face);
  phi_old_p.gather_evaluate(src[0], EvaluationFlags::values | EvaluationFlags::gradients);
  phi_old_m.reinit(face);
  phi_old_m.gather_evaluate(src[0], EvaluationFlags::values | EvaluationFlags::gradients);
  phi_int_p.reinit(face);
  phi_int_p.gather_evaluate(src[1], EvaluationFlags::values | EvaluationFlags::gradients);
  phi_int_m.reinit(face);
  phi_int_m.gather_evaluate(src[1], EvaluationFlags::values | EvaluationFlags::gradients);
  phi_old_press_p.reinit(face);
  phi_old_press_p.gather_evaluate(src[2], EvaluationFlags::values);
  phi_old_press_m.reinit(face);
  phi_old_press_m.gather_evaluate(src[2], EvaluationFlags::values);
  phi_p.reinit(face);
  phi_m.reinit(face);
 
  /*--- Now we loop over all the quadrature points to compute the integrals ---*/
  for(unsigned int q = 0; q < phi_p.n_q_points; ++q) {
  const auto& n_plus = phi_p.get_normal_vector(q);
 
  const auto& avg_grad_u_old = 0.5*(phi_old_p.get_gradient(q) + phi_old_m.get_gradient(q));
  const auto& avg_grad_u_int = 0.5*(phi_int_p.get_gradient(q) + phi_int_m.get_gradient(q));
  const auto& avg_tensor_product_u_n = 0.5*(outer_product(phi_old_p.get_value(q), phi_old_p.get_value(q)) +
  outer_product(phi_old_m.get_value(q), phi_old_m.get_value(q)));
  const auto& avg_tensor_product_u_n_gamma = 0.5*(outer_product(phi_int_p.get_value(q), phi_int_p.get_value(q)) +
  outer_product(phi_int_m.get_value(q), phi_int_m.get_value(q)));
  const auto& avg_p_old = 0.5*(phi_old_press_p.get_value(q) + phi_old_press_m.get_value(q));
 
  phi_p.submit_value((a31/Re*avg_grad_u_old + a32/Re*avg_grad_u_int -
  a31*avg_tensor_product_u_n - a32*avg_tensor_product_u_n_gamma)*n_plus - avg_p_old*n_plus, q);
  phi_m.submit_value(-(a31/Re*avg_grad_u_old + a32/Re*avg_grad_u_int -
  a31*avg_tensor_product_u_n - a32*avg_tensor_product_u_n_gamma)*n_plus + avg_p_old*n_plus, q);
  }
  phi_p.integrate_scatter(EvaluationFlags::values, dst);
  phi_m.integrate_scatter(EvaluationFlags::values, dst);
  }
  }
  }
 
 

The followinf function assembles rhs boundary term for the velocity

  template<int dim, int fe_degree_p, int fe_degree_v, int n_q_points_1d_p, int n_q_points_1d_v, typename Vec>
  void NavierStokesProjectionOperator<dim, fe_degree_p, fe_degree_v, n_q_points_1d_p, n_q_points_1d_v, Vec>::
  assemble_rhs_boundary_term_velocity(const MatrixFree<dim, Number>& data,
  Vec& dst,
  const std::vector<Vec>& src,
  const std::pair<unsigned int, unsigned int>& face_range) const {
  if(TR_BDF2_stage == 1) {
  /*--- We first start by declaring the suitable instances to read already available quantities. Clearly on the boundary
  the second argument has to be true. ---*/
  phi_old(data, true, 0),
  phi_old_extr(data, true, 0);
 
  /*--- We loop over the faces in the range ---*/
  for(unsigned int face = face_range.first; face < face_range.second; ++face) {
  phi_old.reinit(face);
  phi_old.gather_evaluate(src[0], EvaluationFlags::values | EvaluationFlags::gradients);
  phi_old_extr.reinit(face);
  phi_old_extr.gather_evaluate(src[1], EvaluationFlags::values);
  phi_old_press.reinit(face);
  phi_old_press.gather_evaluate(src[2], EvaluationFlags::values);
  phi.reinit(face);
 
  const auto boundary_id = data.get_boundary_id(face); /*--- Get the id in order to impose the proper boundary condition ---*/
  const auto coef_jump = (boundary_id == 1) ? 0.0 : C_u*std::abs((phi.get_normal_vector(0) * phi.inverse_jacobian(0))[dim - 1]);
  const double aux_coeff = (boundary_id == 1) ? 0.0 : 1.0;
 
  /*--- Now we loop over all the quadrature points to compute the integrals ---*/
  for(unsigned int q = 0; q < phi.n_q_points; ++q) {
  const auto& n_plus = phi.get_normal_vector(q);
 
  const auto& grad_u_old = phi_old.get_gradient(q);
  const auto& tensor_product_u_n = outer_product(phi_old.get_value(q), phi_old_extr.get_value(q));
  const auto& p_old = phi_old_press.get_value(q);
  const auto& point_vectorized = phi.quadrature_point(q);
  if(boundary_id == 0) {
  for(unsigned int v = 0; v < VectorizedArray<Number>::size(); ++v) {
  Point<dim> point; /*--- The point returned by the 'quadrature_point' function is not an instance of Point
  and so it is not ready to be directly used. We need to pay attention to the
  vectorization ---*/
  for(unsigned int d = 0; d < dim; ++d)
  point[d] = point_vectorized[d][v];
  for(unsigned int d = 0; d < dim; ++d)
  u_int_m[d][v] = vel_boundary_inflow.value(point, d);
  }
  }
  const auto tensor_product_u_int_m = outer_product(u_int_m, phi_old_extr.get_value(q));
  const auto lambda = (boundary_id == 1) ? 0.0 : std::abs(scalar_product(phi_old_extr.get_value(q), n_plus));
 
  phi.submit_value((a21/Re*grad_u_old - a21*tensor_product_u_n)*n_plus - p_old*n_plus +
  a22/Re*2.0*coef_jump*u_int_m -
  aux_coeff*a22*tensor_product_u_int_m*n_plus + a22*lambda*u_int_m, q);
  phi.submit_normal_derivative(-aux_coeff*theta_v*a22/Re*u_int_m, q); /*--- This is equivalent to multiply to the gradient
  with outer product and use 'submit_gradient' ---*/
  }
  phi.integrate_scatter(EvaluationFlags::values | EvaluationFlags::gradients, dst);
  }
  }
  else {
  /*--- We first start by declaring the suitable instances to read already available quantities. ---*/
  phi_old(data, true, 0),
  phi_int(data, true, 0),
  phi_int_extr(data, true, 0);
 
  /*--- We loop over the faces in the range ---*/
  for(unsigned int face = face_range.first; face < face_range.second; ++ face) {
  phi_old.reinit(face);
  phi_old.gather_evaluate(src[0], EvaluationFlags::values | EvaluationFlags::gradients);
  phi_int.reinit(face);
  phi_int.gather_evaluate(src[1], EvaluationFlags::values | EvaluationFlags::gradients);
  phi_old_press.reinit(face);
  phi_old_press.gather_evaluate(src[2], EvaluationFlags::values);
  phi_int_extr.reinit(face);
  phi_int_extr.gather_evaluate(src[3], EvaluationFlags::values);
  phi.reinit(face);
 
  const auto boundary_id = data.get_boundary_id(face);
  const auto coef_jump = (boundary_id == 1) ? 0.0 : C_u*std::abs((phi.get_normal_vector(0) * phi.inverse_jacobian(0))[dim - 1]);
  const double aux_coeff = (boundary_id == 1) ? 0.0 : 1.0;
 
  /*--- Now we loop over all the quadrature points to compute the integrals ---*/
  for(unsigned int q = 0; q < phi.n_q_points; ++q) {
  const auto& n_plus = phi.get_normal_vector(q);
 
  const auto& grad_u_old = phi_old.get_gradient(q);
  const auto& grad_u_int = phi_int.get_gradient(q);
  const auto& tensor_product_u_n = outer_product(phi_old.get_value(q), phi_old.get_value(q));
  const auto& tensor_product_u_n_gamma = outer_product(phi_int.get_value(q), phi_int.get_value(q));
  const auto& p_old = phi_old_press.get_value(q);
  const auto& point_vectorized = phi.quadrature_point(q);
  if(boundary_id == 0) {
  for(unsigned int v = 0; v < VectorizedArray<Number>::size(); ++v) {
  for(unsigned int d = 0; d < dim; ++d)
  point[d] = point_vectorized[d][v];
  for(unsigned int d = 0; d < dim; ++d)
  u_m[d][v] = vel_boundary_inflow.value(point, d);
  }
  }
  const auto tensor_product_u_m = outer_product(u_m, phi_int_extr.get_value(q));
  const auto lambda = (boundary_id == 1) ? 0.0 : std::abs(scalar_product(phi_int_extr.get_value(q), n_plus));
 
  phi.submit_value((a31/Re*grad_u_old + a32/Re*grad_u_int -
  a31*tensor_product_u_n - a32*tensor_product_u_n_gamma)*n_plus - p_old*n_plus +
  a33/Re*2.0*coef_jump*u_m -
  aux_coeff*a33*tensor_product_u_m*n_plus + a33*lambda*u_m, q);
  phi.submit_normal_derivative(-aux_coeff*theta_v*a33/Re*u_m, q);
  }
  phi.integrate_scatter(EvaluationFlags::values | EvaluationFlags::gradients, dst);
  }
  }
  }
 
 
types::boundary_id get_boundary_id(const unsigned int face_batch_index) const
Point< spacedim > point(const gp_Pnt &p, const double tolerance=1e-10)
Definition utilities.cc:191
::VectorizedArray< Number, width > abs(const ::VectorizedArray< Number, width > &)
unsigned int boundary_id
Definition types.h:144

Put together all the previous steps for velocity. This is done automatically by the loop function of 'MatrixFree' class

  template<int dim, int fe_degree_p, int fe_degree_v, int n_q_points_1d_p, int n_q_points_1d_v, typename Vec>
  void NavierStokesProjectionOperator<dim, fe_degree_p, fe_degree_v, n_q_points_1d_p, n_q_points_1d_v, Vec>::
  vmult_rhs_velocity(Vec& dst, const std::vector<Vec>& src) const {
  for(auto& vec : src)
  vec.update_ghost_values();
 
  this->data->loop(&NavierStokesProjectionOperator::assemble_rhs_cell_term_velocity,
  &NavierStokesProjectionOperator::assemble_rhs_face_term_velocity,
  &NavierStokesProjectionOperator::assemble_rhs_boundary_term_velocity,
  this, dst, src, true,
  }
 
 

Now we focus on computing the rhs for the projection step for the pressure with the same ratio. The following function assembles rhs cell term for the pressure

  template<int dim, int fe_degree_p, int fe_degree_v, int n_q_points_1d_p, int n_q_points_1d_v, typename Vec>
  void NavierStokesProjectionOperator<dim, fe_degree_p, fe_degree_v, n_q_points_1d_p, n_q_points_1d_v, Vec>::
  assemble_rhs_cell_term_pressure(const MatrixFree<dim, Number>& data,
  Vec& dst,
  const std::vector<Vec>& src,
  const std::pair<unsigned int, unsigned int>& cell_range) const {
  /*--- We first start by declaring the suitable instances to read already available quantities.
  The third parameter specifies that we want to use the second quadrature formula stored. ---*/
  phi_old(data, 1, 1);
 
  const double coeff = (TR_BDF2_stage == 1) ? 1.0e6*gamma*dt*gamma*dt : 1.0e6*(1.0 - gamma)*dt*(1.0 - gamma)*dt;
 
  const double coeff_2 = (TR_BDF2_stage == 1) ? gamma*dt : (1.0 - gamma)*dt;
 
  /*--- We loop over cells in the range ---*/
  for(unsigned int cell = cell_range.first; cell < cell_range.second; ++cell) {
  phi_proj.reinit(cell);
  phi_proj.gather_evaluate(src[0], EvaluationFlags::values);
  phi_old.reinit(cell);
  phi_old.gather_evaluate(src[1], EvaluationFlags::values);
  phi.reinit(cell);
 
  /*--- Now we loop over all the quadrature points to compute the integrals ---*/
  for(unsigned int q = 0; q < phi.n_q_points; ++q) {
  const auto& u_star_star = phi_proj.get_value(q);
  const auto& p_old = phi_old.get_value(q);
 
  phi.submit_value(1.0/coeff*p_old, q);
  phi.submit_gradient(1.0/coeff_2*u_star_star, q);
  }
  phi.integrate_scatter(EvaluationFlags::values | EvaluationFlags::gradients, dst);
  }
  }
 
 

The following function assembles rhs face term for the pressure

  template<int dim, int fe_degree_p, int fe_degree_v, int n_q_points_1d_p, int n_q_points_1d_v, typename Vec>
  void NavierStokesProjectionOperator<dim, fe_degree_p, fe_degree_v, n_q_points_1d_p, n_q_points_1d_v, Vec>::
  assemble_rhs_face_term_pressure(const MatrixFree<dim, Number>& data,
  Vec& dst,
  const std::vector<Vec>& src,
  const std::pair<unsigned int, unsigned int>& face_range) const {
  /*--- We first start by declaring the suitable instances to read already available quantities. ---*/
  phi_m(data, false, 1, 1);
  phi_proj_m(data, false, 0, 1);
 
  const double coeff = (TR_BDF2_stage == 1) ? 1.0/(gamma*dt) : 1.0/((1.0 - gamma)*dt);
 
  /*--- We loop over faces in the range ---*/
  for(unsigned int face = face_range.first; face < face_range.second; ++face) {
  phi_proj_p.reinit(face);
  phi_proj_p.gather_evaluate(src[0], EvaluationFlags::values);
  phi_proj_m.reinit(face);
  phi_proj_m.gather_evaluate(src[0], EvaluationFlags::values);
  phi_p.reinit(face);
  phi_m.reinit(face);
 
  /*--- Now we loop over all the quadrature points to compute the integrals ---*/
  for(unsigned int q = 0; q < phi_p.n_q_points; ++q) {
  const auto& n_plus = phi_p.get_normal_vector(q);
  const auto& avg_u_star_star = 0.5*(phi_proj_p.get_value(q) + phi_proj_m.get_value(q));
 
  phi_p.submit_value(-coeff*scalar_product(avg_u_star_star, n_plus), q);
  phi_m.submit_value(coeff*scalar_product(avg_u_star_star, n_plus), q);
  }
  phi_p.integrate_scatter(EvaluationFlags::values, dst);
  phi_m.integrate_scatter(EvaluationFlags::values, dst);
  }
  }
 
 

The following function assembles rhs boundary term for the pressure

  template<int dim, int fe_degree_p, int fe_degree_v, int n_q_points_1d_p, int n_q_points_1d_v, typename Vec>
  void NavierStokesProjectionOperator<dim, fe_degree_p, fe_degree_v, n_q_points_1d_p, n_q_points_1d_v, Vec>::
  assemble_rhs_boundary_term_pressure(const MatrixFree<dim, Number>& data,
  Vec& dst,
  const std::vector<Vec>& src,
  const std::pair<unsigned int, unsigned int>& face_range) const {
  /*--- We first start by declaring the suitable instances to read already available quantities. ---*/
 
  const double coeff = (TR_BDF2_stage == 1) ? 1.0/(gamma*dt) : 1.0/((1.0 - gamma)*dt);
 
  /*--- We loop over faces in the range ---*/
  for(unsigned int face = face_range.first; face < face_range.second; ++face) {
  phi_proj.reinit(face);
  phi_proj.gather_evaluate(src[0], EvaluationFlags::values);
  phi.reinit(face);
 
  /*--- Now we loop over all the quadrature points to compute the integrals ---*/
  for(unsigned int q = 0; q < phi.n_q_points; ++q) {
  const auto& n_plus = phi.get_normal_vector(q);
 
  phi.submit_value(-coeff*scalar_product(phi_proj.get_value(q), n_plus), q);
  }
  phi.integrate_scatter(EvaluationFlags::values, dst);
  }
  }
 
 

Put together all the previous steps for pressure

  template<int dim, int fe_degree_p, int fe_degree_v, int n_q_points_1d_p, int n_q_points_1d_v, typename Vec>
  void NavierStokesProjectionOperator<dim, fe_degree_p, fe_degree_v, n_q_points_1d_p, n_q_points_1d_v, Vec>::
  vmult_rhs_pressure(Vec& dst, const std::vector<Vec>& src) const {
  for(auto& vec : src)
  vec.update_ghost_values();
 
  this->data->loop(&NavierStokesProjectionOperator::assemble_rhs_cell_term_pressure,
  &NavierStokesProjectionOperator::assemble_rhs_face_term_pressure,
  &NavierStokesProjectionOperator::assemble_rhs_boundary_term_pressure,
  this, dst, src, true,
  }
 
 

Now we need to build the 'matrices', i.e. the bilinear forms. We start by assembling the cell term for the velocity

  template<int dim, int fe_degree_p, int fe_degree_v, int n_q_points_1d_p, int n_q_points_1d_v, typename Vec>
  void NavierStokesProjectionOperator<dim, fe_degree_p, fe_degree_v, n_q_points_1d_p, n_q_points_1d_v, Vec>::
  assemble_cell_term_velocity(const MatrixFree<dim, Number>& data,
  Vec& dst,
  const Vec& src,
  const std::pair<unsigned int, unsigned int>& cell_range) const {
  if(TR_BDF2_stage == 1) {
  /*--- We first start by declaring the suitable instances to read already available quantities. Moreover 'phi' in
  this case serves for a bilinear form and so it will not used only to submit but also to read the src ---*/
  phi_old_extr(data, 0);
 
  /*--- We loop over all cells in the range ---*/
  for(unsigned int cell = cell_range.first; cell < cell_range.second; ++cell) {
  phi.reinit(cell);
  phi.gather_evaluate(src, EvaluationFlags::values | EvaluationFlags::gradients);
  phi_old_extr.reinit(cell);
  phi_old_extr.gather_evaluate(u_extr, EvaluationFlags::values);
 
  /*--- Now we loop over all quadrature points ---*/
  for(unsigned int q = 0; q < phi.n_q_points; ++q) {
  const auto& u_int = phi.get_value(q);
  const auto& grad_u_int = phi.get_gradient(q);
  const auto& u_n_gamma_ov_2 = phi_old_extr.get_value(q);
  const auto& tensor_product_u_int = outer_product(u_int, u_n_gamma_ov_2);
 
  phi.submit_value(1.0/(gamma*dt)*u_int, q);
  phi.submit_gradient(-a22*tensor_product_u_int + a22/Re*grad_u_int, q);
  }
  phi.integrate_scatter(EvaluationFlags::values | EvaluationFlags::gradients, dst);
  }
  }
  else {
  /*--- We first start by declaring the suitable instances to read already available quantities. ---*/
  phi_int_extr(data, 0);
 
  /*--- We loop over all cells in the range ---*/
  for(unsigned int cell = cell_range.first; cell < cell_range.second; ++cell) {
  phi.reinit(cell);
  phi.gather_evaluate(src, EvaluationFlags::values | EvaluationFlags::gradients);
  phi_int_extr.reinit(cell);
  phi_int_extr.gather_evaluate(u_extr, EvaluationFlags::values);
 
  /*--- Now we loop over all quadrature points ---*/
  for(unsigned int q = 0; q < phi.n_q_points; ++q) {
  const auto& u_curr = phi.get_value(q);
  const auto& grad_u_curr = phi.get_gradient(q);
  const auto& u_n1_int = phi_int_extr.get_value(q);
  const auto& tensor_product_u_curr = outer_product(u_curr, u_n1_int);
 
  phi.submit_value(1.0/((1.0 - gamma)*dt)*u_curr, q);
  phi.submit_gradient(-a33*tensor_product_u_curr + a33/Re*grad_u_curr, q);
  }
  phi.integrate_scatter(EvaluationFlags::values | EvaluationFlags::gradients, dst);
  }
  }
  }
 
 

The following function assembles face term for the velocity

  template<int dim, int fe_degree_p, int fe_degree_v, int n_q_points_1d_p, int n_q_points_1d_v, typename Vec>
  void NavierStokesProjectionOperator<dim, fe_degree_p, fe_degree_v, n_q_points_1d_p, n_q_points_1d_v, Vec>::
  assemble_face_term_velocity(const MatrixFree<dim, Number>& data,
  Vec& dst,
  const Vec& src,
  const std::pair<unsigned int, unsigned int>& face_range) const {
  if(TR_BDF2_stage == 1) {
  /*--- We first start by declaring the suitable instances to read already available quantities. ---*/
  phi_m(data, false, 0),
  phi_old_extr_p(data, true, 0),
  phi_old_extr_m(data, false, 0);
 
  /*--- We loop over all faces in the range ---*/
  for(unsigned int face = face_range.first; face < face_range.second; ++face) {
  phi_p.reinit(face);
  phi_p.gather_evaluate(src, EvaluationFlags::values | EvaluationFlags::gradients);
  phi_m.reinit(face);
  phi_m.gather_evaluate(src, EvaluationFlags::values | EvaluationFlags::gradients);
  phi_old_extr_p.reinit(face);
  phi_old_extr_p.gather_evaluate(u_extr, EvaluationFlags::values);
  phi_old_extr_m.reinit(face);
  phi_old_extr_m.gather_evaluate(u_extr, EvaluationFlags::values);
 
  const auto coef_jump = C_u*0.5*(std::abs((phi_p.get_normal_vector(0)*phi_p.inverse_jacobian(0))[dim - 1]) +
  std::abs((phi_m.get_normal_vector(0)*phi_m.inverse_jacobian(0))[dim - 1]));
 
  /*--- Now we loop over all quadrature points ---*/
  for(unsigned int q = 0; q < phi_p.n_q_points; ++q) {
  const auto& n_plus = phi_p.get_normal_vector(q);
 
  const auto& avg_grad_u_int = 0.5*(phi_p.get_gradient(q) + phi_m.get_gradient(q));
  const auto& jump_u_int = phi_p.get_value(q) - phi_m.get_value(q);
  const auto& avg_tensor_product_u_int = 0.5*(outer_product(phi_p.get_value(q), phi_old_extr_p.get_value(q)) +
  outer_product(phi_m.get_value(q), phi_old_extr_m.get_value(q)));
  const auto lambda = std::max(std::abs(scalar_product(phi_old_extr_p.get_value(q), n_plus)),
  std::abs(scalar_product(phi_old_extr_m.get_value(q), n_plus)));
 
  phi_p.submit_value(a22/Re*(-avg_grad_u_int*n_plus + coef_jump*jump_u_int) +
  a22*avg_tensor_product_u_int*n_plus + 0.5*a22*lambda*jump_u_int, q);
  phi_m.submit_value(-a22/Re*(-avg_grad_u_int*n_plus + coef_jump*jump_u_int) -
  a22*avg_tensor_product_u_int*n_plus - 0.5*a22*lambda*jump_u_int, q);
  phi_p.submit_normal_derivative(-theta_v*a22/Re*0.5*jump_u_int, q);
  phi_m.submit_normal_derivative(-theta_v*a22/Re*0.5*jump_u_int, q);
  }
  phi_p.integrate_scatter(EvaluationFlags::values | EvaluationFlags::gradients, dst);
  phi_m.integrate_scatter(EvaluationFlags::values | EvaluationFlags::gradients, dst);
  }
  }
  else {
  /*--- We first start by declaring the suitable instances to read already available quantities. ---*/
  phi_m(data, false, 0),
  phi_extr_p(data, true, 0),
  phi_extr_m(data, false, 0);
 
  /*--- We loop over all faces in the range ---*/
  for(unsigned int face = face_range.first; face < face_range.second; ++face) {
  phi_p.reinit(face);
  phi_p.gather_evaluate(src, EvaluationFlags::values | EvaluationFlags::gradients);
  phi_m.reinit(face);
  phi_m.gather_evaluate(src, EvaluationFlags::values | EvaluationFlags::gradients);
  phi_extr_p.reinit(face);
  phi_extr_p.gather_evaluate(u_extr, EvaluationFlags::values);
  phi_extr_m.reinit(face);
  phi_extr_m.gather_evaluate(u_extr, EvaluationFlags::values);
 
  const auto coef_jump = C_u*0.5*(std::abs((phi_p.get_normal_vector(0)*phi_p.inverse_jacobian(0))[dim - 1]) +
  std::abs((phi_m.get_normal_vector(0)*phi_m.inverse_jacobian(0))[dim - 1]));
 
  /*--- Now we loop over all quadrature points ---*/
  for(unsigned int q = 0; q < phi_p.n_q_points; ++q) {
  const auto& n_plus = phi_p.get_normal_vector(q);
 
  const auto& avg_grad_u = 0.5*(phi_p.get_gradient(q) + phi_m.get_gradient(q));
  const auto& jump_u = phi_p.get_value(q) - phi_m.get_value(q);
  const auto& avg_tensor_product_u = 0.5*(outer_product(phi_p.get_value(q), phi_extr_p.get_value(q)) +
  outer_product(phi_m.get_value(q), phi_extr_m.get_value(q)));
  const auto lambda = std::max(std::abs(scalar_product(phi_extr_p.get_value(q), n_plus)),
  std::abs(scalar_product(phi_extr_m.get_value(q), n_plus)));
 
  phi_p.submit_value(a33/Re*(-avg_grad_u*n_plus + coef_jump*jump_u) +
  a33*avg_tensor_product_u*n_plus + 0.5*a33*lambda*jump_u, q);
  phi_m.submit_value(-a33/Re*(-avg_grad_u*n_plus + coef_jump*jump_u) -
  a33*avg_tensor_product_u*n_plus - 0.5*a33*lambda*jump_u, q);
  phi_p.submit_normal_derivative(-theta_v*a33/Re*0.5*jump_u, q);
  phi_m.submit_normal_derivative(-theta_v*a33/Re*0.5*jump_u, q);
  }
  phi_p.integrate_scatter(EvaluationFlags::values | EvaluationFlags::gradients, dst);
  phi_m.integrate_scatter(EvaluationFlags::values | EvaluationFlags::gradients, dst);
  }
  }
  }
 
 
::VectorizedArray< Number, width > max(const ::VectorizedArray< Number, width > &, const ::VectorizedArray< Number, width > &)

The following function assembles boundary term for the velocity

  template<int dim, int fe_degree_p, int fe_degree_v, int n_q_points_1d_p, int n_q_points_1d_v, typename Vec>
  void NavierStokesProjectionOperator<dim, fe_degree_p, fe_degree_v, n_q_points_1d_p, n_q_points_1d_v, Vec>::
  assemble_boundary_term_velocity(const MatrixFree<dim, Number>& data,
  Vec& dst,
  const Vec& src,
  const std::pair<unsigned int, unsigned int>& face_range) const {
  if(TR_BDF2_stage == 1) {
  /*--- We first start by declaring the suitable instances to read already available quantities. ---*/
  phi_old_extr(data, true, 0);
 
  /*--- We loop over all faces in the range ---*/
  for(unsigned int face = face_range.first; face < face_range.second; ++face) {
  phi.reinit(face);
  phi.gather_evaluate(src, EvaluationFlags::values | EvaluationFlags::gradients);
  phi_old_extr.reinit(face);
  phi_old_extr.gather_evaluate(u_extr, EvaluationFlags::values);
 
  const auto boundary_id = data.get_boundary_id(face);
  const auto coef_jump = C_u*std::abs((phi.get_normal_vector(0) * phi.inverse_jacobian(0))[dim - 1]);
 
  /*--- The application of the mirror principle is not so trivial because we have a Dirichlet condition
  on a single component for the outflow; so we distinguish the two cases ---*/
  if(boundary_id != 1) {
  const double coef_trasp = 0.0;
 
  /*--- Now we loop over all quadrature points ---*/
  for(unsigned int q = 0; q < phi.n_q_points; ++q) {
  const auto& n_plus = phi.get_normal_vector(q);
  const auto& grad_u_int = phi.get_gradient(q);
  const auto& u_int = phi.get_value(q);
  const auto& tensor_product_u_int = outer_product(phi.get_value(q), phi_old_extr.get_value(q));
  const auto& lambda = std::abs(scalar_product(phi_old_extr.get_value(q), n_plus));
 
  phi.submit_value(a22/Re*(-grad_u_int*n_plus + 2.0*coef_jump*u_int) +
  a22*coef_trasp*tensor_product_u_int*n_plus + a22*lambda*u_int, q);
  phi.submit_normal_derivative(-theta_v*a22/Re*u_int, q);
  }
  phi.integrate_scatter(EvaluationFlags::values | EvaluationFlags::gradients, dst);
  }
  else {
  /*--- Now we loop over all quadrature points ---*/
  for(unsigned int q = 0; q < phi.n_q_points; ++q) {
  const auto& n_plus = phi.get_normal_vector(q);
  const auto& grad_u_int = phi.get_gradient(q);
  const auto& u_int = phi.get_value(q);
  const auto& lambda = std::abs(scalar_product(phi_old_extr.get_value(q), n_plus));
 
  const auto& point_vectorized = phi.quadrature_point(q);
  auto u_int_m = u_int;
  auto grad_u_int_m = grad_u_int;
  for(unsigned int v = 0; v < VectorizedArray<Number>::size(); ++v) {
  for(unsigned int d = 0; d < dim; ++d)
  point[d] = point_vectorized[d][v];
 
  u_int_m[1][v] = -u_int_m[1][v];
 
  grad_u_int_m[0][0][v] = -grad_u_int_m[0][0][v];
  grad_u_int_m[0][1][v] = -grad_u_int_m[0][1][v];
  }
 
  phi.submit_value(a22/Re*(-(0.5*(grad_u_int + grad_u_int_m))*n_plus + coef_jump*(u_int - u_int_m)) +
  a22*outer_product(0.5*(u_int + u_int_m), phi_old_extr.get_value(q))*n_plus +
  a22*0.5*lambda*(u_int - u_int_m), q);
  phi.submit_normal_derivative(-theta_v*a22/Re*(u_int - u_int_m), q);
  }
  phi.integrate_scatter(EvaluationFlags::values | EvaluationFlags::gradients, dst);
  }
  }
  }
  else {
  /*--- We first start by declaring the suitable instances to read already available quantities. ---*/
  phi_extr(data, true, 0);
 
  /*--- We loop over all faces in the range ---*/
  for(unsigned int face = face_range.first; face < face_range.second; ++face) {
  phi.reinit(face);
  phi.gather_evaluate(src, EvaluationFlags::values | EvaluationFlags::gradients);
  phi_extr.reinit(face);
  phi_extr.gather_evaluate(u_extr, EvaluationFlags::values);
 
  const auto boundary_id = data.get_boundary_id(face);
  const auto coef_jump = C_u*std::abs((phi.get_normal_vector(0) * phi.inverse_jacobian(0))[dim - 1]);
 
  if(boundary_id != 1) {
  const double coef_trasp = 0.0;
 
  /*--- Now we loop over all quadrature points ---*/
  for(unsigned int q = 0; q < phi.n_q_points; ++q) {
  const auto& n_plus = phi.get_normal_vector(q);
  const auto& grad_u = phi.get_gradient(q);
  const auto& u = phi.get_value(q);
  const auto& tensor_product_u = outer_product(phi.get_value(q), phi_extr.get_value(q));
  const auto& lambda = std::abs(scalar_product(phi_extr.get_value(q), n_plus));
 
  phi.submit_value(a33/Re*(-grad_u*n_plus + 2.0*coef_jump*u) +
  a33*coef_trasp*tensor_product_u*n_plus + a33*lambda*u, q);
  phi.submit_normal_derivative(-theta_v*a33/Re*u, q);
  }
  phi.integrate_scatter(EvaluationFlags::values | EvaluationFlags::gradients, dst);
  }
  else {
  /*--- Now we loop over all quadrature points ---*/
  for(unsigned int q = 0; q < phi.n_q_points; ++q) {
  const auto& n_plus = phi.get_normal_vector(q);
  const auto& grad_u = phi.get_gradient(q);
  const auto& u = phi.get_value(q);
  const auto& lambda = std::abs(scalar_product(phi_extr.get_value(q), n_plus));
 
  const auto& point_vectorized = phi.quadrature_point(q);
  auto u_m = u;
  auto grad_u_m = grad_u;
  for(unsigned int v = 0; v < VectorizedArray<Number>::size(); ++v) {
  for(unsigned int d = 0; d < dim; ++d)
  point[d] = point_vectorized[d][v];
 
  u_m[1][v] = -u_m[1][v];
 
  grad_u_m[0][0][v] = -grad_u_m[0][0][v];
  grad_u_m[0][1][v] = -grad_u_m[0][1][v];
  }
 
  phi.submit_value(a33/Re*(-(0.5*(grad_u + grad_u_m))*n_plus + coef_jump*(u - u_m)) +
  a33*outer_product(0.5*(u + u_m), phi_extr.get_value(q))*n_plus + a33*0.5*lambda*(u - u_m), q);
  phi.submit_normal_derivative(-theta_v*a33/Re*(u - u_m), q);
  }
  phi.integrate_scatter(EvaluationFlags::values | EvaluationFlags::gradients, dst);
  }
  }
  }
  }
 
 

Next, we focus on 'matrices' to compute the pressure. We first assemble cell term for the pressure

  template<int dim, int fe_degree_p, int fe_degree_v, int n_q_points_1d_p, int n_q_points_1d_v, typename Vec>
  void NavierStokesProjectionOperator<dim, fe_degree_p, fe_degree_v, n_q_points_1d_p, n_q_points_1d_v, Vec>::
  assemble_cell_term_pressure(const MatrixFree<dim, Number>& data,
  Vec& dst,
  const Vec& src,
  const std::pair<unsigned int, unsigned int>& cell_range) const {
  /*--- We first start by declaring the suitable instances to read already available quantities. ---*/
 
  const double coeff = (TR_BDF2_stage == 1) ? 1.0e6*gamma*dt*gamma*dt : 1.0e6*(1.0 - gamma)*dt*(1.0 - gamma)*dt;
 
  /*--- Loop over all cells in the range ---*/
  for(unsigned int cell = cell_range.first; cell < cell_range.second; ++cell) {
  phi.reinit(cell);
  phi.gather_evaluate(src, EvaluationFlags::values | EvaluationFlags::gradients);
 
  /*--- Now we loop over all quadrature points ---*/
  for(unsigned int q = 0; q < phi.n_q_points; ++q) {
  phi.submit_gradient(phi.get_gradient(q), q);
  phi.submit_value(1.0/coeff*phi.get_value(q), q);
  }
 
  phi.integrate_scatter(EvaluationFlags::values | EvaluationFlags::gradients, dst);
  }
  }
 
 

The following function assembles face term for the pressure

  template<int dim, int fe_degree_p, int fe_degree_v, int n_q_points_1d_p, int n_q_points_1d_v, typename Vec>
  void NavierStokesProjectionOperator<dim, fe_degree_p, fe_degree_v, n_q_points_1d_p, n_q_points_1d_v, Vec>::
  assemble_face_term_pressure(const MatrixFree<dim, Number>& data,
  Vec& dst,
  const Vec& src,
  const std::pair<unsigned int, unsigned int>& face_range) const {
  /*--- We first start by declaring the suitable instances to read already available quantities. ---*/
  phi_m(data, false, 1, 1);
 
  /*--- Loop over all faces in the range ---*/
  for(unsigned int face = face_range.first; face < face_range.second; ++face) {
  phi_p.reinit(face);
  phi_p.gather_evaluate(src, EvaluationFlags::values | EvaluationFlags::gradients);
  phi_m.reinit(face);
  phi_m.gather_evaluate(src, EvaluationFlags::values | EvaluationFlags::gradients);
 
  const auto coef_jump = C_p*0.5*(std::abs((phi_p.get_normal_vector(0)*phi_p.inverse_jacobian(0))[dim - 1]) +
  std::abs((phi_m.get_normal_vector(0)*phi_m.inverse_jacobian(0))[dim - 1]));
 
  /*--- Loop over quadrature points ---*/
  for(unsigned int q = 0; q < phi_p.n_q_points; ++q) {
  const auto& n_plus = phi_p.get_normal_vector(q);
 
  const auto& avg_grad_pres = 0.5*(phi_p.get_gradient(q) + phi_m.get_gradient(q));
  const auto& jump_pres = phi_p.get_value(q) - phi_m.get_value(q);
 
  phi_p.submit_value(-scalar_product(avg_grad_pres, n_plus) + coef_jump*jump_pres, q);
  phi_m.submit_value(scalar_product(avg_grad_pres, n_plus) - coef_jump*jump_pres, q);
  phi_p.submit_gradient(-theta_p*0.5*jump_pres*n_plus, q);
  phi_m.submit_gradient(-theta_p*0.5*jump_pres*n_plus, q);
  }
  phi_p.integrate_scatter(EvaluationFlags::values | EvaluationFlags::gradients, dst);
  phi_m.integrate_scatter(EvaluationFlags::values | EvaluationFlags::gradients, dst);
  }
  }
 
 

The following function assembles boundary term for the pressure

  template<int dim, int fe_degree_p, int fe_degree_v, int n_q_points_1d_p, int n_q_points_1d_v, typename Vec>
  void NavierStokesProjectionOperator<dim, fe_degree_p, fe_degree_v, n_q_points_1d_p, n_q_points_1d_v, Vec>::
  assemble_boundary_term_pressure(const MatrixFree<dim, Number>& data,
  Vec& dst,
  const Vec& src,
  const std::pair<unsigned int, unsigned int>& face_range) const {
 
  for(unsigned int face = face_range.first; face < face_range.second; ++face) {
  const auto boundary_id = data.get_boundary_id(face);
 
  if(boundary_id == 1) {
  phi.reinit(face);
  phi.gather_evaluate(src, EvaluationFlags::values | EvaluationFlags::gradients);
 
  const auto coef_jump = C_p*std::abs((phi.get_normal_vector(0)*phi.inverse_jacobian(0))[dim - 1]);
 
  for(unsigned int q = 0; q < phi.n_q_points; ++q) {
  const auto& n_plus = phi.get_normal_vector(q);
 
  const auto& grad_pres = phi.get_gradient(q);
  const auto& pres = phi.get_value(q);
 
  phi.submit_value(-scalar_product(grad_pres, n_plus) + coef_jump*pres , q);
  phi.submit_normal_derivative(-theta_p*pres, q);
  }
  phi.integrate_scatter(EvaluationFlags::values | EvaluationFlags::gradients, dst);
  }
  }
  }
 
 

Before coding the 'apply_add' function, which is the one that will perform the loop, we focus on the linear system that arises to project the gradient of the pressure into the velocity space. The following function assembles rhs cell term for the projection of gradient of pressure. Since no integration by parts is performed, only a cell term contribution is present.

  template<int dim, int fe_degree_p, int fe_degree_v, int n_q_points_1d_p, int n_q_points_1d_v, typename Vec>
  void NavierStokesProjectionOperator<dim, fe_degree_p, fe_degree_v, n_q_points_1d_p, n_q_points_1d_v, Vec>::
  assemble_rhs_cell_term_projection_grad_p(const MatrixFree<dim, Number>& data,
  Vec& dst,
  const Vec& src,
  const std::pair<unsigned int, unsigned int>& cell_range) const {
  /*--- We first start by declaring the suitable instances to read already available quantities. ---*/
 
  /*--- Loop over all cells in the range ---*/
  for(unsigned int cell = cell_range.first; cell < cell_range.second; ++cell) {
  phi_pres.reinit(cell);
  phi_pres.gather_evaluate(src, EvaluationFlags::gradients);
  phi.reinit(cell);
 
  /*--- Loop over quadrature points ---*/
  for(unsigned int q = 0; q < phi.n_q_points; ++q)
  phi.submit_value(phi_pres.get_gradient(q), q);
 
  phi.integrate_scatter(EvaluationFlags::values, dst);
  }
  }
 
 

Put together all the previous steps for porjection of pressure gradient. Here we loop only over cells

  template<int dim, int fe_degree_p, int fe_degree_v, int n_q_points_1d_p, int n_q_points_1d_v, typename Vec>
  void NavierStokesProjectionOperator<dim, fe_degree_p, fe_degree_v, n_q_points_1d_p, n_q_points_1d_v, Vec>::
  vmult_grad_p_projection(Vec& dst, const Vec& src) const {
  this->data->cell_loop(&NavierStokesProjectionOperator::assemble_rhs_cell_term_projection_grad_p,
  this, dst, src, true);
  }
 
 
void cell_loop(const std::function< void(const MatrixFree< dim, Number, VectorizedArrayType > &, OutVector &, const InVector &, const std::pair< unsigned int, unsigned int > &)> &cell_operation, OutVector &dst, const InVector &src, const bool zero_dst_vector=false) const

Assemble now cell term for the projection of gradient of pressure. This is nothing but a mass matrix

  template<int dim, int fe_degree_p, int fe_degree_v, int n_q_points_1d_p, int n_q_points_1d_v, typename Vec>
  void NavierStokesProjectionOperator<dim, fe_degree_p, fe_degree_v, n_q_points_1d_p, n_q_points_1d_v, Vec>::
  assemble_cell_term_projection_grad_p(const MatrixFree<dim, Number>& data,
  Vec& dst,
  const Vec& src,
  const std::pair<unsigned int, unsigned int>& cell_range) const {
 
  /*--- Loop over all cells in the range ---*/
  for(unsigned int cell = cell_range.first; cell < cell_range.second; ++cell) {
  phi.reinit(cell);
  phi.gather_evaluate(src, EvaluationFlags::values);
 
  /*--- Loop over quadrature points ---*/
  for(unsigned int q = 0; q < phi.n_q_points; ++q)
  phi.submit_value(phi.get_value(q), q);
 
  phi.integrate_scatter(EvaluationFlags::values, dst);
  }
  }
 
 

Put together all previous steps. This is the overriden function that effectively performs the matrix-vector multiplication.

  template<int dim, int fe_degree_p, int fe_degree_v, int n_q_points_1d_p, int n_q_points_1d_v, typename Vec>
  void NavierStokesProjectionOperator<dim, fe_degree_p, fe_degree_v, n_q_points_1d_p, n_q_points_1d_v, Vec>::
  apply_add(Vec& dst, const Vec& src) const {
  if(NS_stage == 1) {
  this->data->loop(&NavierStokesProjectionOperator::assemble_cell_term_velocity,
  &NavierStokesProjectionOperator::assemble_face_term_velocity,
  &NavierStokesProjectionOperator::assemble_boundary_term_velocity,
  this, dst, src, false,
  }
  else if(NS_stage == 2) {
  this->data->loop(&NavierStokesProjectionOperator::assemble_cell_term_pressure,
  &NavierStokesProjectionOperator::assemble_face_term_pressure,
  &NavierStokesProjectionOperator::assemble_boundary_term_pressure,
  this, dst, src, false,
  }
  else if(NS_stage == 3) {
  this->data->cell_loop(&NavierStokesProjectionOperator::assemble_cell_term_projection_grad_p,
  this, dst, src, false); /*--- Since we have only a cell term contribution, we use cell_loop ---*/
  }
  else
  Assert(false, ExcNotImplemented());
  }
 
 

Finally, we focus on computing the diagonal for preconditioners and we start by assembling the diagonal cell term for the velocity. Since we do not have access to the entries of the matrix, in order to compute the element i, we test the matrix against a vector which is equal to 1 in position i and 0 elsewhere. This is why 'src' will result as unused.

  template<int dim, int fe_degree_p, int fe_degree_v, int n_q_points_1d_p, int n_q_points_1d_v, typename Vec>
  void NavierStokesProjectionOperator<dim, fe_degree_p, fe_degree_v, n_q_points_1d_p, n_q_points_1d_v, Vec>::
  assemble_diagonal_cell_term_velocity(const MatrixFree<dim, Number>& data,
  Vec& dst,
  const unsigned int& ,
  const std::pair<unsigned int, unsigned int>& cell_range) const {
  if(TR_BDF2_stage == 1) {
  phi_old_extr(data, 0);
 
  AlignedVector<Tensor<1, dim, VectorizedArray<Number>>> diagonal(phi.dofs_per_component);
  /*--- Build a vector of ones to be tested (here we will see the velocity as a whole vector, since
  dof_handler_velocity is vectorial and so the dof values are vectors). ---*/
  for(unsigned int d = 0; d < dim; ++d)
  tmp[d] = make_vectorized_array<Number>(1.0);
 
  /*--- Loop over cells in the range ---*/
  for(unsigned int cell = cell_range.first; cell < cell_range.second; ++cell) {
  phi_old_extr.reinit(cell);
  phi_old_extr.gather_evaluate(u_extr, EvaluationFlags::values);
  phi.reinit(cell);
 
  /*--- Loop over dofs ---*/
  for(unsigned int i = 0; i < phi.dofs_per_component; ++i) {
  for(unsigned int j = 0; j < phi.dofs_per_component; ++j)
  phi.submit_dof_value(Tensor<1, dim, VectorizedArray<Number>>(), j); /*--- Set all dofs to zero ---*/
  phi.submit_dof_value(tmp, i); /*--- Set dof i equal to one ---*/
 
  /*--- Loop over quadrature points ---*/
  for(unsigned int q = 0; q < phi.n_q_points; ++q) {
  const auto& u_int = phi.get_value(q);
  const auto& grad_u_int = phi.get_gradient(q);
  const auto& u_n_gamma_ov_2 = phi_old_extr.get_value(q);
  const auto& tensor_product_u_int = outer_product(u_int, u_n_gamma_ov_2);
 
  phi.submit_value(1.0/(gamma*dt)*u_int, q);
  phi.submit_gradient(-a22*tensor_product_u_int + a22/Re*grad_u_int, q);
  }
  diagonal[i] = phi.get_dof_value(i);
  }
  for(unsigned int i = 0; i < phi.dofs_per_component; ++i)
  phi.submit_dof_value(diagonal[i], i);
  phi.distribute_local_to_global(dst);
  }
  }
  else {
  phi_int_extr(data, 0);
 
  for(unsigned int d = 0; d < dim; ++d)
  tmp[d] = make_vectorized_array<Number>(1.0);
 
  /*--- Loop over cells in the range ---*/
  for(unsigned int cell = cell_range.first; cell < cell_range.second; ++cell) {
  phi_int_extr.reinit(cell);
  phi_int_extr.gather_evaluate(u_extr, EvaluationFlags::values);
  phi.reinit(cell);
 
  /*--- Loop over dofs ---*/
  for(unsigned int i = 0; i < phi.dofs_per_component; ++i) {
  for(unsigned int j = 0; j < phi.dofs_per_component; ++j)
  phi.submit_dof_value(Tensor<1, dim, VectorizedArray<Number>>(), j);
  phi.submit_dof_value(tmp, i);
 
  /*--- Loop over quadrature points ---*/
  for(unsigned int q = 0; q < phi.n_q_points; ++q) {
  const auto& u_curr = phi.get_value(q);
  const auto& grad_u_curr = phi.get_gradient(q);
  const auto& u_n1_int = phi_int_extr.get_value(q);
  const auto& tensor_product_u_curr = outer_product(u_curr, u_n1_int);
 
  phi.submit_value(1.0/((1.0 - gamma)*dt)*u_curr, q);
  phi.submit_gradient(-a33*tensor_product_u_curr + a33/Re*grad_u_curr, q);
  }
  diagonal[i] = phi.get_dof_value(i);
  }
  for(unsigned int i = 0; i < phi.dofs_per_component; ++i)
  phi.submit_dof_value(diagonal[i], i);
  phi.distribute_local_to_global(dst);
  }
  }
  }
 
 
@ diagonal
Matrix is diagonal.

The following function assembles diagonal face term for the velocity

  template<int dim, int fe_degree_p, int fe_degree_v, int n_q_points_1d_p, int n_q_points_1d_v, typename Vec>
  void NavierStokesProjectionOperator<dim, fe_degree_p, fe_degree_v, n_q_points_1d_p, n_q_points_1d_v, Vec>::
  assemble_diagonal_face_term_velocity(const MatrixFree<dim, Number>& data,
  Vec& dst,
  const unsigned int& ,
  const std::pair<unsigned int, unsigned int>& face_range) const {
  if(TR_BDF2_stage == 1) {
  phi_m(data, false, 0),
  phi_old_extr_p(data, true, 0),
  phi_old_extr_m(data, false, 0);
 
  AssertDimension(phi_p.dofs_per_component, phi_m.dofs_per_component); /*--- We just assert for safety that dimension match,
  in the sense that we have selected the proper
  space ---*/
  AlignedVector<Tensor<1, dim, VectorizedArray<Number>>> diagonal_p(phi_p.dofs_per_component),
  diagonal_m(phi_m.dofs_per_component);
 
  for(unsigned int d = 0; d < dim; ++d)
  tmp[d] = make_vectorized_array<Number>(1.0); /*--- We build the usal vector of ones that we will use as dof value ---*/
 
  /*--- Now we loop over faces ---*/
  for(unsigned int face = face_range.first; face < face_range.second; ++face) {
  phi_old_extr_p.reinit(face);
  phi_old_extr_p.gather_evaluate(u_extr, EvaluationFlags::values);
  phi_old_extr_m.reinit(face);
  phi_old_extr_m.gather_evaluate(u_extr, EvaluationFlags::values);
  phi_p.reinit(face);
  phi_m.reinit(face);
 
  const auto coef_jump = C_u*0.5*(std::abs((phi_p.get_normal_vector(0)*phi_p.inverse_jacobian(0))[dim - 1]) +
  std::abs((phi_m.get_normal_vector(0)*phi_m.inverse_jacobian(0))[dim - 1]));
 
  /*--- Loop over dofs. We will set all equal to zero apart from the current one ---*/
  for(unsigned int i = 0; i < phi_p.dofs_per_component; ++i) {
  for(unsigned int j = 0; j < phi_p.dofs_per_component; ++j) {
  phi_p.submit_dof_value(Tensor<1, dim, VectorizedArray<Number>>(), j);
  phi_m.submit_dof_value(Tensor<1, dim, VectorizedArray<Number>>(), j);
  }
  phi_p.submit_dof_value(tmp, i);
  phi_m.submit_dof_value(tmp, i);
 
  /*--- Loop over quadrature points to compute the integral ---*/
  for(unsigned int q = 0; q < phi_p.n_q_points; ++q) {
  const auto& n_plus = phi_p.get_normal_vector(q);
  const auto& avg_grad_u_int = 0.5*(phi_p.get_gradient(q) + phi_m.get_gradient(q));
  const auto& jump_u_int = phi_p.get_value(q) - phi_m.get_value(q);
  const auto& avg_tensor_product_u_int = 0.5*(outer_product(phi_p.get_value(q), phi_old_extr_p.get_value(q)) +
  outer_product(phi_m.get_value(q), phi_old_extr_m.get_value(q)));
  const auto lambda = std::max(std::abs(scalar_product(phi_old_extr_p.get_value(q), n_plus)),
  std::abs(scalar_product(phi_old_extr_m.get_value(q), n_plus)));
 
  phi_p.submit_value(a22/Re*(-avg_grad_u_int*n_plus + coef_jump*jump_u_int) +
  a22*avg_tensor_product_u_int*n_plus + 0.5*a22*lambda*jump_u_int , q);
  phi_m.submit_value(-a22/Re*(-avg_grad_u_int*n_plus + coef_jump*jump_u_int) -
  a22*avg_tensor_product_u_int*n_plus - 0.5*a22*lambda*jump_u_int, q);
  phi_p.submit_normal_derivative(-theta_v*0.5*a22/Re*jump_u_int, q);
  phi_m.submit_normal_derivative(-theta_v*0.5*a22/Re*jump_u_int, q);
  }
  diagonal_p[i] = phi_p.get_dof_value(i);
  diagonal_m[i] = phi_m.get_dof_value(i);
  }
  for(unsigned int i = 0; i < phi_p.dofs_per_component; ++i) {
  phi_p.submit_dof_value(diagonal_p[i], i);
  phi_m.submit_dof_value(diagonal_m[i], i);
  }
  phi_p.distribute_local_to_global(dst);
  phi_m.distribute_local_to_global(dst);
  }
  }
  else {
  phi_m(data, false, 0),
  phi_extr_p(data, true, 0),
  phi_extr_m(data, false, 0);
 
  AssertDimension(phi_p.dofs_per_component, phi_m.dofs_per_component);
  AlignedVector<Tensor<1, dim, VectorizedArray<Number>>> diagonal_p(phi_p.dofs_per_component),
  diagonal_m(phi_m.dofs_per_component);
  for(unsigned int d = 0; d < dim; ++d)
  tmp[d] = make_vectorized_array<Number>(1.0);
 
  /*--- Now we loop over faces ---*/
  for(unsigned int face = face_range.first; face < face_range.second; ++face) {
  phi_extr_p.reinit(face);
  phi_extr_p.gather_evaluate(u_extr, EvaluationFlags::values);
  phi_extr_m.reinit(face);
  phi_extr_m.gather_evaluate(u_extr, EvaluationFlags::values);
  phi_p.reinit(face);
  phi_m.reinit(face);
 
  const auto coef_jump = C_u*0.5*(std::abs((phi_p.get_normal_vector(0)*phi_p.inverse_jacobian(0))[dim - 1]) +
  std::abs((phi_m.get_normal_vector(0)*phi_m.inverse_jacobian(0))[dim - 1]));
 
  /*--- Loop over dofs. We will set all equal to zero apart from the current one ---*/
  for(unsigned int i = 0; i < phi_p.dofs_per_component; ++i) {
  for(unsigned int j = 0; j < phi_p.dofs_per_component; ++j) {
  phi_p.submit_dof_value(Tensor<1, dim, VectorizedArray<Number>>(), j);
  phi_m.submit_dof_value(Tensor<1, dim, VectorizedArray<Number>>(), j);
  }
  phi_p.submit_dof_value(tmp, i);
  phi_m.submit_dof_value(tmp, i);
 
  /*--- Loop over quadrature points to compute the integral ---*/
  for(unsigned int q = 0; q < phi_p.n_q_points; ++q) {
  const auto& n_plus = phi_p.get_normal_vector(q);
  const auto& avg_grad_u = 0.5*(phi_p.get_gradient(q) + phi_m.get_gradient(q));
  const auto& jump_u = phi_p.get_value(q) - phi_m.get_value(q);
  const auto& avg_tensor_product_u = 0.5*(outer_product(phi_p.get_value(q), phi_extr_p.get_value(q)) +
  outer_product(phi_m.get_value(q), phi_extr_m.get_value(q)));
  const auto lambda = std::max(std::abs(scalar_product(phi_extr_p.get_value(q), n_plus)),
  std::abs(scalar_product(phi_extr_m.get_value(q), n_plus)));
 
  phi_p.submit_value(a33/Re*(-avg_grad_u*n_plus + coef_jump*jump_u) +
  a33*avg_tensor_product_u*n_plus + 0.5*a33*lambda*jump_u, q);
  phi_m.submit_value(-a33/Re*(-avg_grad_u*n_plus + coef_jump*jump_u) -
  a33*avg_tensor_product_u*n_plus - 0.5*a33*lambda*jump_u, q);
  phi_p.submit_normal_derivative(-theta_v*0.5*a33/Re*jump_u, q);
  phi_m.submit_normal_derivative(-theta_v*0.5*a33/Re*jump_u, q);
  }
  diagonal_p[i] = phi_p.get_dof_value(i);
  diagonal_m[i] = phi_m.get_dof_value(i);
  }
  for(unsigned int i = 0; i < phi_p.dofs_per_component; ++i) {
  phi_p.submit_dof_value(diagonal_p[i], i);
  phi_m.submit_dof_value(diagonal_m[i], i);
  }
  phi_p.distribute_local_to_global(dst);
  phi_m.distribute_local_to_global(dst);
  }
  }
  }
 
 
#define AssertDimension(dim1, dim2)

The following function assembles boundary term for the velocity

  template<int dim, int fe_degree_p, int fe_degree_v, int n_q_points_1d_p, int n_q_points_1d_v, typename Vec>
  void NavierStokesProjectionOperator<dim, fe_degree_p, fe_degree_v, n_q_points_1d_p, n_q_points_1d_v, Vec>::
  assemble_diagonal_boundary_term_velocity(const MatrixFree<dim, Number>& data,
  Vec& dst,
  const unsigned int& ,
  const std::pair<unsigned int, unsigned int>& face_range) const {
  if(TR_BDF2_stage == 1) {
  phi_old_extr(data, true, 0);
 
  AlignedVector<Tensor<1, dim, VectorizedArray<Number>>> diagonal(phi.dofs_per_component);
  for(unsigned int d = 0; d < dim; ++d)
  tmp[d] = make_vectorized_array<Number>(1.0);
 
  /*--- Loop over all faces in the range ---*/
  for(unsigned int face = face_range.first; face < face_range.second; ++face) {
  phi_old_extr.reinit(face);
  phi_old_extr.gather_evaluate(u_extr, EvaluationFlags::values);
  phi.reinit(face);
 
  const auto boundary_id = data.get_boundary_id(face);
  const auto coef_jump = C_u*std::abs((phi.get_normal_vector(0) * phi.inverse_jacobian(0))[dim - 1]);
 
  if(boundary_id != 1) {
  const double coef_trasp = 0.0;
 
  /*--- Loop over all dofs ---*/
  for(unsigned int i = 0; i < phi.dofs_per_component; ++i) {
  for(unsigned int j = 0; j < phi.dofs_per_component; ++j)
  phi.submit_dof_value(Tensor<1, dim, VectorizedArray<Number>>(), j);
  phi.submit_dof_value(tmp, i);
 
  /*--- Loop over quadrature points to compute the integral ---*/
  for(unsigned int q = 0; q < phi.n_q_points; ++q) {
  const auto& n_plus = phi.get_normal_vector(q);
  const auto& grad_u_int = phi.get_gradient(q);
  const auto& u_int = phi.get_value(q);
  const auto& tensor_product_u_int = outer_product(phi.get_value(q), phi_old_extr.get_value(q));
  const auto& lambda = std::abs(scalar_product(phi_old_extr.get_value(q), n_plus));
 
  phi.submit_value(a22/Re*(-grad_u_int*n_plus + 2.0*coef_jump*u_int) +
  a22*coef_trasp*tensor_product_u_int*n_plus + a22*lambda*u_int, q);
  phi.submit_normal_derivative(-theta_v*a22/Re*u_int, q);
  }
  diagonal[i] = phi.get_dof_value(i);
  }
  for(unsigned int i = 0; i < phi.dofs_per_component; ++i)
  phi.submit_dof_value(diagonal[i], i);
  phi.distribute_local_to_global(dst);
  }
  else {
  /*--- Loop over all dofs ---*/
  for(unsigned int i = 0; i < phi.dofs_per_component; ++i) {
  for(unsigned int j = 0; j < phi.dofs_per_component; ++j)
  phi.submit_dof_value(Tensor<1, dim, VectorizedArray<Number>>(), j);
  phi.submit_dof_value(tmp, i);
 
  /*--- Loop over quadrature points to compute the integral ---*/
  for(unsigned int q = 0; q < phi.n_q_points; ++q) {
  const auto& n_plus = phi.get_normal_vector(q);
  const auto& grad_u_int = phi.get_gradient(q);
  const auto& u_int = phi.get_value(q);
  const auto& lambda = std::abs(scalar_product(phi_old_extr.get_value(q), n_plus));
 
  const auto& point_vectorized = phi.quadrature_point(q);
  auto u_int_m = u_int;
  auto grad_u_int_m = grad_u_int;
  for(unsigned int v = 0; v < VectorizedArray<Number>::size(); ++v) {
  for(unsigned int d = 0; d < dim; ++d)
  point[d] = point_vectorized[d][v];
 
  u_int_m[1][v] = -u_int_m[1][v];
 
  grad_u_int_m[0][0][v] = -grad_u_int_m[0][0][v];
  grad_u_int_m[0][1][v] = -grad_u_int_m[0][1][v];
  }
 
  phi.submit_value(a22/Re*(-(0.5*(grad_u_int + grad_u_int_m))*n_plus + coef_jump*(u_int - u_int_m)) +
  a22*outer_product(0.5*(u_int + u_int_m), phi_old_extr.get_value(q))*n_plus +
  a22*0.5*lambda*(u_int - u_int_m), q);
  phi.submit_normal_derivative(-theta_v*a22/Re*(u_int - u_int_m), q);
  }
  diagonal[i] = phi.get_dof_value(i);
  }
  for(unsigned int i = 0; i < phi.dofs_per_component; ++i)
  phi.submit_dof_value(diagonal[i], i);
  phi.distribute_local_to_global(dst);
  }
  }
  }
  else {
  phi_extr(data, true, 0);
 
  for(unsigned int d = 0; d < dim; ++d)
  tmp[d] = make_vectorized_array<Number>(1.0);
 
  /*--- Loop over all faces in the range ---*/
  for(unsigned int face = face_range.first; face < face_range.second; ++face) {
  phi_extr.reinit(face);
  phi_extr.gather_evaluate(u_extr, EvaluationFlags::values);
  phi.reinit(face);
 
  const auto boundary_id = data.get_boundary_id(face);
  const auto coef_jump = C_u*std::abs((phi.get_normal_vector(0) * phi.inverse_jacobian(0))[dim - 1]);
 
  if(boundary_id != 1) {
  const double coef_trasp = 0.0;
 
  /*--- Loop over all dofs ---*/
  for(unsigned int i = 0; i < phi.dofs_per_component; ++i) {
  for(unsigned int j = 0; j < phi.dofs_per_component; ++j)
  phi.submit_dof_value(Tensor<1, dim, VectorizedArray<Number>>(), j);
  phi.submit_dof_value(tmp, i);
 
  /*--- Loop over quadrature points to compute the integral ---*/
  for(unsigned int q = 0; q < phi.n_q_points; ++q) {
  const auto& n_plus = phi.get_normal_vector(q);
  const auto& grad_u = phi.get_gradient(q);
  const auto& u = phi.get_value(q);
  const auto& tensor_product_u = outer_product(phi.get_value(q), phi_extr.get_value(q));
  const auto& lambda = std::abs(scalar_product(phi_extr.get_value(q), n_plus));
 
  phi.submit_value(a33/Re*(-grad_u*n_plus + 2.0*coef_jump*u) +
  a33*coef_trasp*tensor_product_u*n_plus + a33*lambda*u, q);
  phi.submit_normal_derivative(-theta_v*a33/Re*u, q);
  }
  diagonal[i] = phi.get_dof_value(i);
  }
  for(unsigned int i = 0; i < phi.dofs_per_component; ++i)
  phi.submit_dof_value(diagonal[i], i);
  phi.distribute_local_to_global(dst);
  }
  else {
  /*--- Loop over all dofs ---*/
  for(unsigned int i = 0; i < phi.dofs_per_component; ++i) {
  for(unsigned int j = 0; j < phi.dofs_per_component; ++j)
  phi.submit_dof_value(Tensor<1, dim, VectorizedArray<Number>>(), j);
  phi.submit_dof_value(tmp, i);
 
  /*--- Loop over quadrature points to compute the integral ---*/
  for(unsigned int q = 0; q < phi.n_q_points; ++q) {
  const auto& n_plus = phi.get_normal_vector(q);
  const auto& grad_u = phi.get_gradient(q);
  const auto& u = phi.get_value(q);
  const auto& lambda = std::abs(scalar_product(phi_extr.get_value(q), n_plus));
 
  const auto& point_vectorized = phi.quadrature_point(q);
  auto u_m = u;
  auto grad_u_m = grad_u;
  for(unsigned int v = 0; v < VectorizedArray<Number>::size(); ++v) {
  for(unsigned int d = 0; d < dim; ++d)
  point[d] = point_vectorized[d][v];
 
  u_m[1][v] = -u_m[1][v];
 
  grad_u_m[0][0][v] = -grad_u_m[0][0][v];
  grad_u_m[0][1][v] = -grad_u_m[0][1][v];
  }
 
  phi.submit_value(a33/Re*(-(0.5*(grad_u + grad_u_m))*n_plus + coef_jump*(u - u_m)) +
  a33*outer_product(0.5*(u + u_m), phi_extr.get_value(q))*n_plus +
  a33*0.5*lambda*(u - u_m), q);
  phi.submit_normal_derivative(-theta_v*a33/Re*(u - u_m), q);
  }
  diagonal[i] = phi.get_dof_value(i);
  }
  for(unsigned int i = 0; i < phi.dofs_per_component; ++i)
  phi.submit_dof_value(diagonal[i], i);
  phi.distribute_local_to_global(dst);
  }
  }
  }
  }
 
 

Now we consider the pressure related bilinear forms. We first assemble diagonal cell term for the pressure

  template<int dim, int fe_degree_p, int fe_degree_v, int n_q_points_1d_p, int n_q_points_1d_v, typename Vec>
  void NavierStokesProjectionOperator<dim, fe_degree_p, fe_degree_v, n_q_points_1d_p, n_q_points_1d_v, Vec>::
  assemble_diagonal_cell_term_pressure(const MatrixFree<dim, Number>& data,
  Vec& dst,
  const unsigned int& ,
  const std::pair<unsigned int, unsigned int>& cell_range) const {
 
  AlignedVector<VectorizedArray<Number>> diagonal(phi.dofs_per_component); /*--- Here we are using dofs_per_component but
  it coincides with dofs_per_cell since it is
  scalar finite element space ---*/
 
  const double coeff = (TR_BDF2_stage == 1) ? 1e6*gamma*dt*gamma*dt : 1e6*(1.0 - gamma)*dt*(1.0 - gamma)*dt;
 
  /*--- Loop over all cells in the range ---*/
  for(unsigned int cell = cell_range.first; cell < cell_range.second; ++cell) {
  phi.reinit(cell);
 
  /*--- Loop over all dofs ---*/
  for(unsigned int i = 0; i < phi.dofs_per_component; ++i) {
  for(unsigned int j = 0; j < phi.dofs_per_component; ++j)
  phi.submit_dof_value(VectorizedArray<Number>(), j); /*--- We set all dofs to zero ---*/
  phi.submit_dof_value(make_vectorized_array<Number>(1.0), i); /*--- Now we set the current one to 1; since it is scalar,
  we can directly use 'make_vectorized_array' without
  relying on 'Tensor' ---*/
 
  /*--- Loop over quadrature points ---*/
  for(unsigned int q = 0; q < phi.n_q_points; ++q) {
  phi.submit_value(1.0/coeff*phi.get_value(q), q);
  phi.submit_gradient(phi.get_gradient(q), q);
  }
  diagonal[i] = phi.get_dof_value(i);
  }
  for(unsigned int i = 0; i < phi.dofs_per_component; ++i)
  phi.submit_dof_value(diagonal[i], i);
 
  phi.distribute_local_to_global(dst);
  }
  }
 
 

The following function assembles diagonal face term for the pressure

  template<int dim, int fe_degree_p, int fe_degree_v, int n_q_points_1d_p, int n_q_points_1d_v, typename Vec>
  void NavierStokesProjectionOperator<dim, fe_degree_p, fe_degree_v, n_q_points_1d_p, n_q_points_1d_v, Vec>::
  assemble_diagonal_face_term_pressure(const MatrixFree<dim, Number>& data,
  Vec& dst,
  const unsigned int& ,
  const std::pair<unsigned int, unsigned int>& face_range) const {
  phi_m(data, false, 1, 1);
 
  AssertDimension(phi_p.dofs_per_component, phi_m.dofs_per_component);
  AlignedVector<VectorizedArray<Number>> diagonal_p(phi_p.dofs_per_component),
  diagonal_m(phi_m.dofs_per_component); /*--- Again, we just assert for safety that dimension
  match, in the sense that we have selected
  the proper space ---*/
 
  /*--- Loop over all faces ---*/
  for(unsigned int face = face_range.first; face < face_range.second; ++face) {
  phi_p.reinit(face);
  phi_m.reinit(face);
 
  const auto coef_jump = C_p*0.5*(std::abs((phi_p.get_normal_vector(0)*phi_p.inverse_jacobian(0))[dim - 1]) +
  std::abs((phi_m.get_normal_vector(0)*phi_m.inverse_jacobian(0))[dim - 1]));
 
  /*--- Loop over all dofs ---*/
  for(unsigned int i = 0; i < phi_p.dofs_per_component; ++i) {
  for(unsigned int j = 0; j < phi_p.dofs_per_component; ++j) {
  phi_p.submit_dof_value(VectorizedArray<Number>(), j);
  phi_m.submit_dof_value(VectorizedArray<Number>(), j);
  }
  phi_p.submit_dof_value(make_vectorized_array<Number>(1.0), i);
  phi_m.submit_dof_value(make_vectorized_array<Number>(1.0), i);
 
  /*--- Loop over all quadrature points to compute the integral ---*/
  for(unsigned int q = 0; q < phi_p.n_q_points; ++q) {
  const auto& n_plus = phi_p.get_normal_vector(q);
 
  const auto& avg_grad_pres = 0.5*(phi_p.get_gradient(q) + phi_m.get_gradient(q));
  const auto& jump_pres = phi_p.get_value(q) - phi_m.get_value(q);
 
  phi_p.submit_value(-scalar_product(avg_grad_pres, n_plus) + coef_jump*jump_pres, q);
  phi_m.submit_value(scalar_product(avg_grad_pres, n_plus) - coef_jump*jump_pres, q);
  phi_p.submit_gradient(-theta_p*0.5*jump_pres*n_plus, q);
  phi_m.submit_gradient(-theta_p*0.5*jump_pres*n_plus, q);
  }
  diagonal_p[i] = phi_p.get_dof_value(i);
  diagonal_m[i] = phi_m.get_dof_value(i);
  }
  for(unsigned int i = 0; i < phi_p.dofs_per_component; ++i) {
  phi_p.submit_dof_value(diagonal_p[i], i);
  phi_m.submit_dof_value(diagonal_m[i], i);
  }
  phi_p.distribute_local_to_global(dst);
  phi_m.distribute_local_to_global(dst);
  }
  }
 
 

Eventually, we assemble diagonal boundary term for the pressure

  template<int dim, int fe_degree_p, int fe_degree_v, int n_q_points_1d_p, int n_q_points_1d_v, typename Vec>
  void NavierStokesProjectionOperator<dim, fe_degree_p, fe_degree_v, n_q_points_1d_p, n_q_points_1d_v, Vec>::
  assemble_diagonal_boundary_term_pressure(const MatrixFree<dim, Number>& data,
  Vec& dst,
  const unsigned int& ,
  const std::pair<unsigned int, unsigned int>& face_range) const {
 
  AlignedVector<VectorizedArray<Number>> diagonal(phi.dofs_per_component);
 
  for(unsigned int face = face_range.first; face < face_range.second; ++face) {
  const auto boundary_id = data.get_boundary_id(face);
 
  if(boundary_id == 1) {
  phi.reinit(face);
 
  const auto coef_jump = C_p*std::abs((phi.get_normal_vector(0)*phi.inverse_jacobian(0))[dim - 1]);
 
  for(unsigned int i = 0; i < phi.dofs_per_component; ++i) {
  for(unsigned int j = 0; j < phi.dofs_per_component; ++j)
  phi.submit_dof_value(VectorizedArray<Number>(), j);
  phi.submit_dof_value(make_vectorized_array<Number>(1.0), i);
 
  for(unsigned int q = 0; q < phi.n_q_points; ++q) {
  const auto& n_plus = phi.get_normal_vector(q);
 
  const auto& grad_pres = phi.get_gradient(q);
  const auto& pres = phi.get_value(q);
 
  phi.submit_value(-scalar_product(grad_pres, n_plus) + 2.0*coef_jump*pres , q);
  phi.submit_normal_derivative(-theta_p*pres, q);
  }
  diagonal[i] = phi.get_dof_value(i);
  }
  for(unsigned int i = 0; i < phi.dofs_per_component; ++i)
  phi.submit_dof_value(diagonal[i], i);
  phi.distribute_local_to_global(dst);
  }
  }
  }
 
 

Put together all previous steps. We create a dummy auxliary vector that serves for the src input argument in the previous functions that as we have seen before is unused. Then everything is done by the 'loop' function and it is saved in the field 'inverse_diagonal_entries' already present in the base class. Anyway since there is only one field, we need to resize properly depending on whether we are considering the velocity or the pressure.

  template<int dim, int fe_degree_p, int fe_degree_v, int n_q_points_1d_p, int n_q_points_1d_v, typename Vec>
  void NavierStokesProjectionOperator<dim, fe_degree_p, fe_degree_v, n_q_points_1d_p, n_q_points_1d_v, Vec>::
  compute_diagonal() {
  Assert(NS_stage == 1 || NS_stage == 2, ExcInternalError());
 
  this->inverse_diagonal_entries.reset(new DiagonalMatrix<Vec>());
  auto& inverse_diagonal = this->inverse_diagonal_entries->get_vector();
 
  if(NS_stage == 1) {
  ::MatrixFreeTools::compute_diagonal<dim, Number, VectorizedArray<Number>>
  (*(this->data),
  inverse_diagonal,
  [&](const auto& data, auto& dst, const auto& src, const auto& cell_range) {
  (this->assemble_diagonal_cell_term_velocity)(data, dst, src, cell_range);
  },
  [&](const auto& data, auto& dst, const auto& src, const auto& face_range) {
  (this->assemble_diagonal_face_term_velocity)(data, dst, src, face_range);
  },
  [&](const auto& data, auto& dst, const auto& src, const auto& boundary_range) {
  (this->assemble_diagonal_boundary_term_velocity)(data, dst, src, boundary_range);
  },
  0);
  }
  else if(NS_stage == 2) {
  ::MatrixFreeTools::compute_diagonal<dim, Number, VectorizedArray<Number>>
  (*(this->data),
  inverse_diagonal,
  [&](const auto& data, auto& dst, const auto& src, const auto& cell_range) {
  (this->assemble_diagonal_cell_term_pressure)(data, dst, src, cell_range);
  },
  [&](const auto& data, auto& dst, const auto& src, const auto& face_range) {
  (this->assemble_diagonal_face_term_pressure)(data, dst, src, face_range);
  },
  [&](const auto& data, auto& dst, const auto& src, const auto& boundary_range) {
  (this->assemble_diagonal_boundary_term_pressure)(data, dst, src, boundary_range);
  },
  1);
  }
 
  for(unsigned int i = 0; i < inverse_diagonal.locally_owned_size(); ++i) {
  Assert(inverse_diagonal.local_element(i) != 0.0,
  ExcMessage("No diagonal entry in a definite operator should be zero"));
  inverse_diagonal.local_element(i) = 1.0/inverse_diagonal.local_element(i);
  }
  }
 
 

@sect{The NavierStokesProjection class}

Now we are ready for the main class of the program. It implements the calls to the various steps of the projection method for Navier-Stokes equations.

  template<int dim>
  class NavierStokesProjection {
  public:
  NavierStokesProjection(RunTimeParameters::Data_Storage& data);
 
  void run(const bool verbose = false, const unsigned int output_interval = 10);
 
  protected:
  const double t_0;
  const double T;
  const double gamma; //--- TR-BDF2 parameter
  unsigned int TR_BDF2_stage; //--- Flag to check at which current stage of TR-BDF2 are
  const double Re;
  double dt;
 
  EquationData::Velocity<dim> vel_init;
  EquationData::Pressure<dim> pres_init; /*--- Instance of 'Velocity' and 'Pressure' classes to initialize. ---*/
 
 
  /*--- Finite Element spaces ---*/
  FESystem<dim> fe_velocity;
  FESystem<dim> fe_pressure;
 
  /*--- Handler for dofs ---*/
  DoFHandler<dim> dof_handler_velocity;
  DoFHandler<dim> dof_handler_pressure;
 
  /*--- Quadrature formulas for velocity and pressure, respectively ---*/
  QGauss<dim> quadrature_pressure;
  QGauss<dim> quadrature_velocity;
 
  /*--- Now we define all the vectors for the solution. We start from the pressure
  with p^n, p^(n+gamma) and a vector for rhs ---*/
 
  /*--- Next, we move to the velocity, with u^n, u^(n-1), u^(n+gamma/2),
  u^(n+gamma) and other two auxiliary vectors as well as the rhs ---*/
 
  Vector<double> Linfty_error_per_cell_vel;
 
  DeclException2(ExcInvalidTimeStep,
  double,
  double,
  << " The time step " << arg1 << " is out of range."
  << std::endl
  << " The permitted range is (0," << arg2 << "]");
 
  void create_triangulation(const unsigned int n_refines);
 
  void setup_dofs();
 
  void initialize();
 
  void interpolate_velocity();
 
  void diffusion_step();
 
  void projection_step();
 
  void project_grad(const unsigned int flag);
 
  double get_maximal_velocity();
 
  double get_maximal_difference_velocity();
 
  void output_results(const unsigned int step);
 
  void refine_mesh();
 
  void interpolate_max_res(const unsigned int level);
 
  void save_max_res();
 
  private:
  void compute_lift_and_drag();
 
  /*--- Technical member to handle the various steps ---*/
  std::shared_ptr<MatrixFree<dim, double>> matrix_free_storage;
 
  /*--- Now we need an instance of the class implemented before with the weak form ---*/
  NavierStokesProjectionOperator<dim, EquationData::degree_p, EquationData::degree_p + 1,
  EquationData::degree_p + 1, EquationData::degree_p + 2,
 
  /*--- This is an instance for geometric multigrid preconditioner ---*/
  MGLevelObject<NavierStokesProjectionOperator<dim, EquationData::degree_p, EquationData::degree_p + 1,
  EquationData::degree_p + 1, EquationData::degree_p + 2,
 
  /*--- Here we define two 'AffineConstraints' instance, one for each finite element space.
  This is just a technical issue, due to MatrixFree requirements. In general
  this class is used to impose boundary conditions (or any kind of constraints), but in this case, since
  we are using a weak imposition of bcs, everything is already in the weak forms and so these instances
  will be default constructed ---*/
  AffineConstraints<double> constraints_velocity,
  constraints_pressure;
 
  /*--- Now a bunch of variables handled by 'ParamHandler' introduced at the beginning of the code ---*/
  unsigned int max_its;
  double eps;
 
  unsigned int max_loc_refinements;
  unsigned int min_loc_refinements;
  unsigned int refinement_iterations;
 
  std::string saving_dir;
 
  /*--- Finally, some output related streams ---*/
 
  std::ofstream time_out;
  ConditionalOStream ptime_out;
  TimerOutput time_table;
 
  std::ofstream output_n_dofs_velocity;
  std::ofstream output_n_dofs_pressure;
 
  std::ofstream output_lift;
  std::ofstream output_drag;
  };
 
 
unsigned int level
Definition grid_out.cc:4626
#define DeclException2(Exception2, type1, type2, outsequence)
Definition exceptions.h:539
const ::parallel::distributed::Triangulation< dim, spacedim > * triangulation

In the constructor, we just read all the data from the Data_Storage object that is passed as an argument, verify that the data we read are reasonable and, finally, create the triangulation and load the initial data.

  template<int dim>
  NavierStokesProjection<dim>::NavierStokesProjection(RunTimeParameters::Data_Storage& data):
  t_0(data.initial_time),
  T(data.final_time),
  gamma(2.0 - std::sqrt(2.0)), //--- Save also in the NavierStokes class the TR-BDF2 parameter value
  TR_BDF2_stage(1), //--- Initialize the flag for the TR_BDF2 stage
  Re(data.Reynolds),
  dt(data.dt),
  vel_init(data.initial_time),
  pres_init(data.initial_time),
  triangulation(MPI_COMM_WORLD, parallel::distributed::Triangulation<dim>::limit_level_difference_at_vertices,
  parallel::distributed::Triangulation<dim>::construct_multigrid_hierarchy),
  fe_velocity(FE_DGQ<dim>(EquationData::degree_p + 1), dim),
  fe_pressure(FE_DGQ<dim>(EquationData::degree_p), 1),
  dof_handler_velocity(triangulation),
  dof_handler_pressure(triangulation),
  quadrature_pressure(EquationData::degree_p + 1),
  quadrature_velocity(EquationData::degree_p + 2),
  navier_stokes_matrix(data),
  max_its(data.max_iterations),
  eps(data.eps),
  max_loc_refinements(data.max_loc_refinements),
  min_loc_refinements(data.min_loc_refinements),
  refinement_iterations(data.refinement_iterations),
  saving_dir(data.dir),
  pcout(std::cout, Utilities::MPI::this_mpi_process(MPI_COMM_WORLD) == 0),
  time_out("./" + data.dir + "/time_analysis_" +
  Utilities::int_to_string(Utilities::MPI::n_mpi_processes(MPI_COMM_WORLD)) + "proc.dat"),
  ptime_out(time_out, Utilities::MPI::this_mpi_process(MPI_COMM_WORLD) == 0),
  time_table(ptime_out, TimerOutput::summary, TimerOutput::cpu_and_wall_times),
  output_n_dofs_velocity("./" + data.dir + "/n_dofs_velocity.dat", std::ofstream::out),
  output_n_dofs_pressure("./" + data.dir + "/n_dofs_pressure.dat", std::ofstream::out),
  output_lift("./" + data.dir + "/lift.dat", std::ofstream::out),
  output_drag("./" + data.dir + "/drag.dat", std::ofstream::out) {
  if(EquationData::degree_p < 1) {
  pcout
  << " WARNING: The chosen pair of finite element spaces is not stable."
  << std::endl
  << " The obtained results will be nonsense" << std::endl;
  }
 
  AssertThrow(!((dt <= 0.0) || (dt > 0.5*T)), ExcInvalidTimeStep(dt, 0.5*T));
 
  matrix_free_storage = std::make_shared<MatrixFree<dim, double>>();
 
  create_triangulation(data.n_refines);
  setup_dofs();
  initialize();
  }
 
 
#define AssertThrow(cond, exc)

The method that creates the triangulation and refines it the needed number of times.

  template<int dim>
  void NavierStokesProjection<dim>::create_triangulation(const unsigned int n_refines) {
  TimerOutput::Scope t(time_table, "Create triangulation");
 
  GridGenerator::plate_with_a_hole(triangulation, 0.5, 1.0, 1.0, 1.1, 1.0, 19.0, Point<2>(2.0, 2.0), 0, 1, 1.0, 2, true);
  /*--- We strongly advice to check the documentation to verify the meaning of all input parameters. ---*/
 
  pcout << "Number of refines = " << n_refines << std::endl;
  triangulation.refine_global(n_refines);
  }
 
 
void plate_with_a_hole(Triangulation< dim > &tria, const double inner_radius=0.4, const double outer_radius=1., const double pad_bottom=2., const double pad_top=2., const double pad_left=1., const double pad_right=1., const Point< dim > &center=Point< dim >(), const types::manifold_id polar_manifold_id=0, const types::manifold_id tfi_manifold_id=1, const double L=1., const unsigned int n_slices=2, const bool colorize=false)
Rectangular plate with an (offset) cylindrical hole.

After creating the triangulation, it creates the mesh dependent data, i.e. it distributes degrees of freedom, and initializes the vectors that we will use.

  template<int dim>
  void NavierStokesProjection<dim>::setup_dofs() {
  pcout << "Number of active cells: " << triangulation.n_global_active_cells() << std::endl;
  pcout << "Number of levels: " << triangulation.n_global_levels() << std::endl;
 
  /*--- Distribute dofs and prepare for multigrid ---*/
  dof_handler_velocity.distribute_dofs(fe_velocity);
  dof_handler_pressure.distribute_dofs(fe_pressure);
 
  pcout << "dim (X_h) = " << dof_handler_velocity.n_dofs()
  << std::endl
  << "dim (M_h) = " << dof_handler_pressure.n_dofs()
  << std::endl
  << "Re = " << Re << std::endl
  << std::endl;
 
  if(Utilities::MPI::this_mpi_process(MPI_COMM_WORLD) == 0) {
  output_n_dofs_velocity << dof_handler_velocity.n_dofs() << std::endl;
  output_n_dofs_pressure << dof_handler_pressure.n_dofs() << std::endl;
  }
 
  typename MatrixFree<dim, double>::AdditionalData additional_data;
  additional_data.mapping_update_flags_inner_faces = (update_gradients | update_JxW_values | update_quadrature_points |
  additional_data.mapping_update_flags_boundary_faces = (update_gradients | update_JxW_values | update_quadrature_points |
  additional_data.tasks_parallel_scheme = MatrixFree<dim, double>::AdditionalData::none;
 
  std::vector<const DoFHandler<dim>*> dof_handlers; /*--- Vector of dof_handlers to feed the 'MatrixFree'. Here the order
  counts and enters into the game as parameter of FEEvaluation and
  FEFaceEvaluation in the previous class ---*/
  dof_handlers.push_back(&dof_handler_velocity);
  dof_handlers.push_back(&dof_handler_pressure);
 
  constraints_velocity.clear();
  constraints_velocity.close();
  constraints_pressure.clear();
  constraints_pressure.close();
  std::vector<const AffineConstraints<double>*> constraints;
  constraints.push_back(&constraints_velocity);
  constraints.push_back(&constraints_pressure);
 
  std::vector<QGauss<1>> quadratures; /*--- We cannot directly use 'quadrature_velocity' and 'quadrature_pressure',
  because the 'MatrixFree' structure wants a quadrature formula for 1D
  (this is way the template parameter of the previous class was called 'n_q_points_1d_p'
  and 'n_q_points_1d_v' and the reason of '1' as QGauss template parameter). ---*/
  quadratures.push_back(QGauss<1>(EquationData::degree_p + 2));
  quadratures.push_back(QGauss<1>(EquationData::degree_p + 1));
 
  /*--- Initialize the matrix-free structure and size properly the vectors. Here again the
  second input argument of the 'initialize_dof_vector' method depends on the order of 'dof_handlers' ---*/
  matrix_free_storage->reinit(MappingQ1<dim>(),dof_handlers, constraints, quadratures, additional_data);
  matrix_free_storage->initialize_dof_vector(u_star, 0);
  matrix_free_storage->initialize_dof_vector(rhs_u, 0);
  matrix_free_storage->initialize_dof_vector(u_n, 0);
  matrix_free_storage->initialize_dof_vector(u_extr, 0);
  matrix_free_storage->initialize_dof_vector(u_n_minus_1, 0);
  matrix_free_storage->initialize_dof_vector(u_n_gamma, 0);
  matrix_free_storage->initialize_dof_vector(u_tmp, 0);
  matrix_free_storage->initialize_dof_vector(grad_pres_int, 0);
 
  matrix_free_storage->initialize_dof_vector(pres_int, 1);
  matrix_free_storage->initialize_dof_vector(pres_n, 1);
  matrix_free_storage->initialize_dof_vector(rhs_p, 1);
 
  /*--- Initialize the multigrid structure. We dedicate ad hoc 'dof_handlers_mg' and 'constraints_mg' because
  we use float as type. Moreover we can initialize already with the index of the finite element of the pressure;
  anyway we need by requirement to declare also structures for the velocity for coherence (basically because
  the index of finite element space has to be the same, so the pressure has to be the second).---*/
  mg_matrices.clear_elements();
  dof_handler_velocity.distribute_mg_dofs();
  dof_handler_pressure.distribute_mg_dofs();
 
  const unsigned int nlevels = triangulation.n_global_levels();
  mg_matrices.resize(0, nlevels - 1);
  for(unsigned int level = 0; level < nlevels; ++level) {
  typename MatrixFree<dim, float>::AdditionalData additional_data_mg;
  additional_data_mg.mapping_update_flags = (update_gradients | update_JxW_values);
  additional_data_mg.mapping_update_flags_inner_faces = (update_gradients | update_JxW_values);
  additional_data_mg.mapping_update_flags_boundary_faces = (update_gradients | update_JxW_values);
  additional_data_mg.mg_level = level;
 
  std::vector<const DoFHandler<dim>*> dof_handlers_mg;
  dof_handlers_mg.push_back(&dof_handler_velocity);
  dof_handlers_mg.push_back(&dof_handler_pressure);
  std::vector<const AffineConstraints<float>*> constraints_mg;
  AffineConstraints<float> constraints_velocity_mg;
  constraints_velocity_mg.clear();
  constraints_velocity_mg.close();
  constraints_mg.push_back(&constraints_velocity_mg);
  AffineConstraints<float> constraints_pressure_mg;
  constraints_pressure_mg.clear();
  constraints_pressure_mg.close();
  constraints_mg.push_back(&constraints_pressure_mg);
 
  std::shared_ptr<MatrixFree<dim, float>> mg_mf_storage_level(new MatrixFree<dim, float>());
  mg_mf_storage_level->reinit(MappingQ1<dim>(),dof_handlers_mg, constraints_mg, quadratures, additional_data_mg);
  const std::vector<unsigned int> tmp = {1};
  mg_matrices[level].initialize(mg_mf_storage_level, tmp, tmp);
  mg_matrices[level].set_dt(dt);
  mg_matrices[level].set_NS_stage(2);
  }
 
  Linfty_error_per_cell_vel.reinit(triangulation.n_active_cells());
  }
 
 
void clear()
unsigned int mg_level
@ update_values
Shape function values.
@ update_normal_vectors
Normal vectors.
@ update_JxW_values
Transformed quadrature weights.
@ update_quadrature_points
Transformed quadrature points.
unsigned int this_mpi_process(const MPI_Comm mpi_communicator)
Definition mpi.cc:107
TasksParallelScheme tasks_parallel_scheme

This method loads the initial data. It simply uses the class Pressure instance for the pressure and the class Velocity instance for the velocity.

  template<int dim>
  void NavierStokesProjection<dim>::initialize() {
  TimerOutput::Scope t(time_table, "Initialize pressure and velocity");
 
  VectorTools::interpolate(dof_handler_pressure, pres_init, pres_n);
 
  VectorTools::interpolate(dof_handler_velocity, vel_init, u_n_minus_1);
  VectorTools::interpolate(dof_handler_velocity, vel_init, u_n);
  }
 
 
void interpolate(const Mapping< dim, spacedim > &mapping, const DoFHandler< dim, spacedim > &dof, const Function< spacedim, typename VectorType::value_type > &function, VectorType &vec, const ComponentMask &component_mask={})

This function computes the extrapolated velocity to be used in the momentum predictor

  template<int dim>
  void NavierStokesProjection<dim>::interpolate_velocity() {
  TimerOutput::Scope t(time_table, "Interpolate velocity");
 

— TR-BDF2 first step

  if(TR_BDF2_stage == 1) {
  u_extr.equ(1.0 + gamma/(2.0*(1.0 - gamma)), u_n);
  u_tmp.equ(gamma/(2.0*(1.0 - gamma)), u_n_minus_1);
  u_extr -= u_tmp;
  }

— TR-BDF2 second step

  else {
  u_extr.equ(1.0 + (1.0 - gamma)/gamma, u_n_gamma);
  u_tmp.equ((1.0 - gamma)/gamma, u_n);
  u_extr -= u_tmp;
  }
  }
 
 

We are finally ready to solve the diffusion step.

  template<int dim>
  void NavierStokesProjection<dim>::diffusion_step() {
  TimerOutput::Scope t(time_table, "Diffusion step");
 
  /*--- We first speicify that we want to deal with velocity dof_handler (index 0, since it is the first one
  in the 'dof_handlers' vector) ---*/
  const std::vector<unsigned int> tmp = {0};
  navier_stokes_matrix.initialize(matrix_free_storage, tmp, tmp);
 
  /*--- Next, we specify at we are at stage 1, namely the diffusion step ---*/
  navier_stokes_matrix.set_NS_stage(1);
 
  /*--- Now, we compute the right-hand side and we set the convective velocity. The necessity of 'set_u_extr' is
  that this quantity is required in the bilinear forms and we can't use a vector of src like on the right-hand side,
  so it has to be available ---*/
  if(TR_BDF2_stage == 1) {
  navier_stokes_matrix.vmult_rhs_velocity(rhs_u, {u_n, u_extr, pres_n});
  navier_stokes_matrix.set_u_extr(u_extr);
  u_star = u_extr;
  }
  else {
  navier_stokes_matrix.vmult_rhs_velocity(rhs_u, {u_n, u_n_gamma, pres_int, u_extr});
  navier_stokes_matrix.set_u_extr(u_extr);
  u_star = u_extr;
  }
 
  /*--- Build the linear solver; in this case we specifiy the maximum number of iterations and residual ---*/
  SolverControl solver_control(max_its, eps*rhs_u.l2_norm());
 
  /*--- Build a Jacobi preconditioner and solve ---*/
  PreconditionJacobi<NavierStokesProjectionOperator<dim,
  EquationData::degree_p,
  EquationData::degree_p + 1,
  EquationData::degree_p + 1,
  EquationData::degree_p + 2,
  navier_stokes_matrix.compute_diagonal();
  preconditioner.initialize(navier_stokes_matrix);
 
  gmres.solve(navier_stokes_matrix, u_star, rhs_u, preconditioner);
  }
 
 

Next, we solve the projection step.

  template<int dim>
  void NavierStokesProjection<dim>::projection_step() {
  TimerOutput::Scope t(time_table, "Projection step pressure");
 
  /*--- We start in the same way of 'diffusion_step': we first reinitialize with the index of FE space,
  we specify that this is the second stage and we compute the right-hand side ---*/
  const std::vector<unsigned int> tmp = {1};
  navier_stokes_matrix.initialize(matrix_free_storage, tmp, tmp);
 
  navier_stokes_matrix.set_NS_stage(2);
 
  if(TR_BDF2_stage == 1)
  navier_stokes_matrix.vmult_rhs_pressure(rhs_p, {u_star, pres_n});
  else
  navier_stokes_matrix.vmult_rhs_pressure(rhs_p, {u_star, pres_int});
 
  /*--- Build the linear solver (Conjugate Gradient in this case) ---*/
  SolverControl solver_control(max_its, eps*rhs_p.l2_norm());
 
  /*--- Build the preconditioner (as in @ref step_37 "step-37") ---*/
  mg_transfer.build(dof_handler_pressure);
 
  using SmootherType = PreconditionChebyshev<NavierStokesProjectionOperator<dim,
  EquationData::degree_p,
  EquationData::degree_p + 1,
  EquationData::degree_p + 1,
  EquationData::degree_p + 2,
  smoother_data.resize(0, triangulation.n_global_levels() - 1);
  for(unsigned int level = 0; level < triangulation.n_global_levels(); ++level) {
  if(level > 0) {
  smoother_data[level].smoothing_range = 15.0;
  smoother_data[level].degree = 3;
  smoother_data[level].eig_cg_n_iterations = 10;
  }
  else {
  smoother_data[0].smoothing_range = 2e-2;
  smoother_data[0].degree = numbers::invalid_unsigned_int;
  smoother_data[0].eig_cg_n_iterations = mg_matrices[0].m();
  }
  mg_matrices[level].compute_diagonal();
  smoother_data[level].preconditioner = mg_matrices[level].get_matrix_diagonal_inverse();
  }
  mg_smoother.initialize(mg_matrices, smoother_data);
 
  NavierStokesProjectionOperator<dim,
  EquationData::degree_p,
  EquationData::degree_p + 1,
  EquationData::degree_p + 1,
  EquationData::degree_p + 2,
  PreconditionIdentity> mg_coarse(cg_mg, mg_matrices[0], identity);
 
 
  Multigrid<LinearAlgebra::distributed::Vector<float>> mg(mg_matrix, mg_coarse, mg_transfer, mg_smoother, mg_smoother);
 
  MGTransferMatrixFree<dim, float>> preconditioner(dof_handler_pressure, mg, mg_transfer);
 
  /*--- Solve the linear system ---*/
  if(TR_BDF2_stage == 1) {
  pres_int = pres_n;
  cg.solve(navier_stokes_matrix, pres_int, rhs_p, preconditioner);
  }
  else {
  pres_n = pres_int;
  cg.solve(navier_stokes_matrix, pres_n, rhs_p, preconditioner);
  }
  }
 
 
void resize(const unsigned int new_minlevel, const unsigned int new_maxlevel, Args &&...args)
void build(const DoFHandler< dim > &dof_handler, const std::vector< std::shared_ptr< const Utilities::MPI::Partitioner > > &external_partitioners=std::vector< std::shared_ptr< const Utilities::MPI::Partitioner > >())
SymmetricTensor< 2, dim, Number > e(const Tensor< 2, dim, Number > &F)
Definition mg.h:81
static const unsigned int invalid_unsigned_int
Definition types.h:220
std_cxx20::type_identity< T > identity

This implements the projection step for the gradient of pressure

  template<int dim>
  void NavierStokesProjection<dim>::project_grad(const unsigned int flag) {
  TimerOutput::Scope t(time_table, "Gradient of pressure projection");
 
  /*--- The input parameter flag is used just to specify where we want to save the result ---*/
  AssertIndexRange(flag, 3);
  Assert(flag > 0, ExcInternalError());
 
  /*--- We need to select the dof handler related to the velocity since the result lives there ---*/
  const std::vector<unsigned int> tmp = {0};
  navier_stokes_matrix.initialize(matrix_free_storage, tmp, tmp);
 
  if(flag == 1)
  navier_stokes_matrix.vmult_grad_p_projection(rhs_u, pres_n);
  else if(flag == 2)
  navier_stokes_matrix.vmult_grad_p_projection(rhs_u, pres_int);
 
  /*--- We conventionally decide that the this corresponds to third stage ---*/
  navier_stokes_matrix.set_NS_stage(3);
 
  /*--- Solve the system ---*/
  SolverControl solver_control(max_its, 1e-12*rhs_u.l2_norm());
  cg.solve(navier_stokes_matrix, u_tmp, rhs_u, PreconditionIdentity());
  }
 
 

The following function is used in determining the maximal velocity in order to compute the Courant number.

  template<int dim>
  double NavierStokesProjection<dim>::get_maximal_velocity() {
  return u_n.linfty_norm();
  }
 
 

The following function is used in determining the maximal nodal difference between old and current velocity value in order to see if we have reched steady-state.

  template<int dim>
  double NavierStokesProjection<dim>::get_maximal_difference_velocity() {
  u_tmp = u_n;
  u_tmp -= u_n_minus_1;
 
  return u_tmp.linfty_norm();
  }
 
 

This method plots the current solution. The main difficulty is that we want to create a single output file that contains the data for all velocity components and the pressure. On the other hand, velocities and the pressure live on separate DoFHandler objects, so we need to pay attention when we use 'add_data_vector' to select the proper space.

  template<int dim>
  void NavierStokesProjection<dim>::output_results(const unsigned int step) {
  TimerOutput::Scope t(time_table, "Output results");
 
  DataOut<dim> data_out;
 
  std::vector<std::string> velocity_names(dim, "v");
  std::vector<DataComponentInterpretation::DataComponentInterpretation>
  component_interpretation_velocity(dim, DataComponentInterpretation::component_is_part_of_vector);
  u_n.update_ghost_values();
  data_out.add_data_vector(dof_handler_velocity, u_n, velocity_names, component_interpretation_velocity);
  pres_n.update_ghost_values();
  data_out.add_data_vector(dof_handler_pressure, pres_n, "p", {DataComponentInterpretation::component_is_scalar});
 
  std::vector<std::string> velocity_names_old(dim, "v_old");
  u_n_minus_1.update_ghost_values();
  data_out.add_data_vector(dof_handler_velocity, u_n_minus_1, velocity_names_old, component_interpretation_velocity);
 
  /*--- Here we rely on the postprocessor we have built ---*/
  PostprocessorVorticity<dim> postprocessor;
  data_out.add_data_vector(dof_handler_velocity, u_n, postprocessor);
 
  data_out.build_patches(MappingQ1<dim>(), 1, DataOut<dim>::curved_inner_cells);
 
  const std::string output = "./" + saving_dir + "/solution-" + Utilities::int_to_string(step, 5) + ".vtu";
  data_out.write_vtu_in_parallel(output, MPI_COMM_WORLD);
  }
 
 
std::string int_to_string(const unsigned int value, const unsigned int digits=numbers::invalid_unsigned_int)
Definition utilities.cc:470

@sect{NavierStokesProjection::compute_lift_and_drag}

This routine computes the lift and the drag forces in a non-dimensional framework (so basically for the classical coefficients, it is necessary to multiply by a factor 2).

  template<int dim>
  void NavierStokesProjection<dim>::compute_lift_and_drag() {
  QGauss<dim - 1> face_quadrature_formula(EquationData::degree_p + 2);
  const int n_q_points = face_quadrature_formula.size();
 
  std::vector<double> pressure_values(n_q_points);
  std::vector<std::vector<Tensor<1, dim>>> velocity_gradients(n_q_points, std::vector<Tensor<1, dim>>(dim));
 
  Tensor<1, dim> normal_vector;
  Tensor<2, dim> fluid_stress;
  Tensor<2, dim> fluid_pressure;
  Tensor<1, dim> forces;
 
  /*--- We need to compute the integral over the cylinder boundary, so we need to use 'FEFaceValues' instances.
  For the velocity we need the gradients, for the pressure the values. ---*/
  FEFaceValues<dim> fe_face_values_velocity(fe_velocity, face_quadrature_formula,
  FEFaceValues<dim> fe_face_values_pressure(fe_pressure, face_quadrature_formula, update_values);
 
  double local_drag = 0.0;
  double local_lift = 0.0;
 
  /*--- We need to perform a unique loop because the whole stress tensor takes into account contributions of
  velocity and pressure obviously. However, the two dof_handlers are different, so we neede to create an ad-hoc
  iterator for the pressure that we update manually. It is guaranteed that the cells are visited in the same order
  (see the documentation) ---*/
  auto tmp_cell = dof_handler_pressure.begin_active();
  for(const auto& cell : dof_handler_velocity.active_cell_iterators()) {
  if(cell->is_locally_owned()) {
  for(unsigned int face = 0; face < GeometryInfo<dim>::faces_per_cell; ++face) {
  if(cell->face(face)->at_boundary() && cell->face(face)->boundary_id() == 4) {
  fe_face_values_velocity.reinit(cell, face);
  fe_face_values_pressure.reinit(tmp_cell, face);
 
  fe_face_values_velocity.get_function_gradients(u_n, velocity_gradients); /*--- velocity gradients ---*/
  fe_face_values_pressure.get_function_values(pres_n, pressure_values); /*--- pressure values ---*/
 
  for(int q = 0; q < n_q_points; q++) {
  normal_vector = -fe_face_values_velocity.normal_vector(q);
 
  for(unsigned int d = 0; d < dim; ++ d) {
  fluid_pressure[d][d] = pressure_values[q];
  for(unsigned int k = 0; k < dim; ++k)
  fluid_stress[d][k] = 1.0/Re*velocity_gradients[q][d][k];
  }
  fluid_stress = fluid_stress - fluid_pressure;
 
  forces = fluid_stress*normal_vector*fe_face_values_velocity.JxW(q);
 
  local_drag += forces[0];
  local_lift += forces[1];
  }
  }
  }
  }
  ++tmp_cell;
  }
 
  /*--- At the end, each processor has computed the contribution to the boundary cells it owns and, therefore,
  we need to sum up all the contributions. ---*/
  const double lift = Utilities::MPI::sum(local_lift, MPI_COMM_WORLD);
  const double drag = Utilities::MPI::sum(local_drag, MPI_COMM_WORLD);
  if(Utilities::MPI::this_mpi_process(MPI_COMM_WORLD) == 0) {
  output_lift << lift << std::endl;
  output_drag << drag << std::endl;
  }
  }
 
 
T sum(const T &t, const MPI_Comm mpi_communicator)

@sect{ NavierStokesProjection::refine_mesh}

After finding a good initial guess on the coarse mesh, we hope to decrease the error through refining the mesh. We also need to transfer the current solution to the next mesh using the SolutionTransfer class.

  template <int dim>
  void NavierStokesProjection<dim>::refine_mesh() {
  TimerOutput::Scope t(time_table, "Refine mesh");
 
  /*--- We first create a proper vector for computing estimator ---*/
  const IndexSet locally_relevant_dofs = DoFTools::extract_locally_relevant_dofs(dof_handler_velocity);
  tmp_velocity.reinit(dof_handler_velocity.locally_owned_dofs(), locally_relevant_dofs, MPI_COMM_WORLD);
  tmp_velocity = u_n;
  tmp_velocity.update_ghost_values();
 
  using Iterator = typename DoFHandler<dim>::active_cell_iterator;
  Vector<float> estimated_error_per_cell(triangulation.n_active_cells());
 
  /*--- This is basically the indicator per cell computation (see @ref step_50 "step-50"). Since it is not so complciated
  we implement it through a lambda expression ---*/
  const auto cell_worker = [&](const Iterator& cell,
  ScratchData<dim>& scratch_data,
  CopyData& copy_data) {
  FEValues<dim>& fe_values = scratch_data.fe_values; /*--- Here we finally use the 'FEValues' inside ScratchData ---*/
  fe_values.reinit(cell);
 
  /*--- Compute the gradients for all quadrature points ---*/
  std::vector<std::vector<Tensor<1, dim>>> gradients(fe_values.n_quadrature_points, std::vector<Tensor<1, dim>>(dim));
  fe_values.get_function_gradients(tmp_velocity, gradients);
  copy_data.cell_index = cell->active_cell_index();
  double vorticity_norm_square = 0.0;
  /*--- Loop over quadrature points and evaluate the integral multiplying the vorticty
  by the weights and the determinant of the Jacobian (which are included in 'JxW') ---*/
  for(unsigned k = 0; k < fe_values.n_quadrature_points; ++k) {
  const double vorticity = gradients[k][1][0] - gradients[k][0][1];
  vorticity_norm_square += vorticity*vorticity*fe_values.JxW(k);
  }
  copy_data.value = cell->diameter()*cell->diameter()*vorticity_norm_square;
  };
 
 
  const auto copier = [&](const CopyData &copy_data) {
  if(copy_data.cell_index != numbers::invalid_unsigned_int)
  estimated_error_per_cell[copy_data.cell_index] += copy_data.value;
  };
 
  /*--- Now everything is 'automagically' handled by 'mesh_loop' ---*/
  ScratchData<dim> scratch_data(fe_velocity, EquationData::degree_p + 2, cell_flags);
  CopyData copy_data;
  MeshWorker::mesh_loop(dof_handler_velocity.begin_active(),
  dof_handler_velocity.end(),
  cell_worker,
  copier,
  scratch_data,
  copy_data,
 
  /*--- Refine grid. In case the refinement level is above a certain value (or the coarsening level is below)
  we clear the flags. ---*/
  for(const auto& cell: triangulation.active_cell_iterators()) {
  if(cell->refine_flag_set() && static_cast<unsigned int>(cell->level()) == max_loc_refinements)
  cell->clear_refine_flag();
  if(cell->coarsen_flag_set() && static_cast<unsigned int>(cell->level()) == min_loc_refinements)
  cell->clear_coarsen_flag();
  }
  triangulation.prepare_coarsening_and_refinement();
 
  /*--- Now we prepare the object for transfering, basically saving the old quantities using SolutionTransfer.
  Since the 'prepare_for_coarsening_and_refinement' method can be called only once, but we have two vectors
  for dof_handler_velocity, we need to put them in an auxiliary vector. ---*/
  std::vector<const LinearAlgebra::distributed::Vector<double>*> velocities;
  velocities.push_back(&u_n);
  velocities.push_back(&u_n_minus_1);
  solution_transfer_velocity(dof_handler_velocity);
  solution_transfer_velocity.prepare_for_coarsening_and_refinement(velocities);
  solution_transfer_pressure(dof_handler_pressure);
  solution_transfer_pressure.prepare_for_coarsening_and_refinement(pres_n);
 
  triangulation.execute_coarsening_and_refinement(); /*--- Effectively perform the remeshing ---*/
 
  /*--- First DoFHandler objects are set up within the new grid ----*/
  setup_dofs();
 
  /*--- Interpolate current solutions to new mesh. This is done using auxliary vectors just for safety,
  but the new u_n or pres_n could be used. Again, the only point is that the function 'interpolate'
  can be called once and so the vectors related to 'dof_handler_velocity' have to collected in an auxiliary vector. ---*/
  transfer_velocity_minus_1,
  transfer_pressure;
  transfer_velocity.reinit(u_n);
  transfer_velocity.zero_out_ghost_values();
  transfer_velocity_minus_1.reinit(u_n_minus_1);
  transfer_velocity_minus_1.zero_out_ghost_values();
  transfer_pressure.reinit(pres_n);
  transfer_pressure.zero_out_ghost_values();
 
  std::vector<LinearAlgebra::distributed::Vector<double>*> transfer_velocities;
  transfer_velocities.push_back(&transfer_velocity);
  transfer_velocities.push_back(&transfer_velocity_minus_1);
  solution_transfer_velocity.interpolate(transfer_velocities);
  transfer_velocity.update_ghost_values();
  transfer_velocity_minus_1.update_ghost_values();
  solution_transfer_pressure.interpolate(transfer_pressure);
  transfer_pressure.update_ghost_values();
 
  u_n = transfer_velocity;
  u_n_minus_1 = transfer_velocity_minus_1;
  pres_n = transfer_pressure;
  }
 
 
void reinit(const TriaIterator< DoFCellAccessor< dim, spacedim, level_dof_access > > &cell)
void reinit(const size_type size, const bool omit_zeroing_entries=false)
typename ActiveSelector::active_cell_iterator active_cell_iterator
void mesh_loop(const CellIteratorType &begin, const CellIteratorType &end, const CellWorkerFunctionType &cell_worker, const CopierType &copier, const ScratchData &sample_scratch_data, const CopyData &sample_copy_data, const AssembleFlags flags=assemble_own_cells, const BoundaryWorkerFunctionType &boundary_worker=BoundaryWorkerFunctionType(), const FaceWorkerFunctionType &face_worker=FaceWorkerFunctionType(), const unsigned int queue_length=2 *MultithreadInfo::n_threads(), const unsigned int chunk_size=8)
Definition mesh_loop.h:281
IndexSet extract_locally_relevant_dofs(const DoFHandler< dim, spacedim > &dof_handler)
void refine_and_coarsen_fixed_number(::Triangulation< dim, spacedim > &tria, const ::Vector< Number > &criteria, const double top_fraction_of_cells, const double bottom_fraction_of_cells, const types::global_cell_index max_n_cells=std::numeric_limits< types::global_cell_index >::max())

Interpolate the locally refined solution to a mesh with maximal resolution and transfer velocity and pressure.

  template<int dim>
  void NavierStokesProjection<dim>::interpolate_max_res(const unsigned int level) {
  solution_transfer_velocity(dof_handler_velocity);
  std::vector<const LinearAlgebra::distributed::Vector<double>*> velocities;
  velocities.push_back(&u_n);
  velocities.push_back(&u_n_minus_1);
  solution_transfer_velocity.prepare_for_coarsening_and_refinement(velocities);
 
  solution_transfer_pressure(dof_handler_pressure);
  solution_transfer_pressure.prepare_for_coarsening_and_refinement(pres_n);
 
  for(const auto& cell: triangulation.active_cell_iterators_on_level(level)) {
  if(cell->is_locally_owned())
  cell->set_refine_flag();
  }
  triangulation.execute_coarsening_and_refinement();
 
  setup_dofs();
 
  LinearAlgebra::distributed::Vector<double> transfer_velocity, transfer_velocity_minus_1,
  transfer_pressure;
 
  transfer_velocity.reinit(u_n);
  transfer_velocity.zero_out_ghost_values();
  transfer_velocity_minus_1.reinit(u_n_minus_1);
  transfer_velocity_minus_1.zero_out_ghost_values();
 
  transfer_pressure.reinit(pres_n);
  transfer_pressure.zero_out_ghost_values();
 
  std::vector<LinearAlgebra::distributed::Vector<double>*> transfer_velocities;
 
  transfer_velocities.push_back(&transfer_velocity);
  transfer_velocities.push_back(&transfer_velocity_minus_1);
  solution_transfer_velocity.interpolate(transfer_velocities);
  transfer_velocity.update_ghost_values();
  transfer_velocity_minus_1.update_ghost_values();
 
  solution_transfer_pressure.interpolate(transfer_pressure);
  transfer_pressure.update_ghost_values();
 
  u_n = transfer_velocity;
  u_n_minus_1 = transfer_velocity_minus_1;
  pres_n = transfer_pressure;
  }
 
 

Save maximum resolution to a mesh adapted.

  template<int dim>
  void NavierStokesProjection<dim>::save_max_res() {
  parallel::distributed::Triangulation<dim> triangulation_tmp(MPI_COMM_WORLD);
  GridGenerator::plate_with_a_hole(triangulation_tmp, 0.5, 1.0, 1.0, 1.1, 1.0, 19.0, Point<2>(2.0, 2.0), 0, 1, 1.0, 2, true);
  triangulation_tmp.refine_global(triangulation.n_global_levels() - 1);
 
  DoFHandler<dim> dof_handler_velocity_tmp(triangulation_tmp);
  DoFHandler<dim> dof_handler_pressure_tmp(triangulation_tmp);
  dof_handler_velocity_tmp.distribute_dofs(fe_velocity);
  dof_handler_pressure_tmp.distribute_dofs(fe_pressure);
 
  pres_n_tmp;
  u_n_tmp.reinit(dof_handler_velocity_tmp.n_dofs());
  pres_n_tmp.reinit(dof_handler_pressure_tmp.n_dofs());
 
  DataOut<dim> data_out;
  std::vector<std::string> velocity_names(dim, "v");
  std::vector<DataComponentInterpretation::DataComponentInterpretation>
  component_interpretation_velocity(dim, DataComponentInterpretation::component_is_part_of_vector);
  VectorTools::interpolate_to_different_mesh(dof_handler_velocity, u_n, dof_handler_velocity_tmp, u_n_tmp);
  u_n_tmp.update_ghost_values();
  data_out.add_data_vector(dof_handler_velocity_tmp, u_n_tmp, velocity_names, component_interpretation_velocity);
  VectorTools::interpolate_to_different_mesh(dof_handler_pressure, pres_n, dof_handler_pressure_tmp, pres_n_tmp);
  pres_n_tmp.update_ghost_values();
  data_out.add_data_vector(dof_handler_pressure_tmp, pres_n_tmp, "p", {DataComponentInterpretation::component_is_scalar});
  PostprocessorVorticity<dim> postprocessor;
  data_out.add_data_vector(dof_handler_velocity_tmp, u_n_tmp, postprocessor);
 
  data_out.build_patches(MappingQ1<dim>(), 1, DataOut<dim>::curved_inner_cells);
  const std::string output = "./" + saving_dir + "/solution_max_res_end.vtu";
  data_out.write_vtu_in_parallel(output, MPI_COMM_WORLD);
  }
 
 
void interpolate_to_different_mesh(const DoFHandler< dim, spacedim > &dof_handler_1, const VectorType &u1, const DoFHandler< dim, spacedim > &dof_handler_2, VectorType &u2)

@sect{ NavierStokesProjection::run }

This is the time marching function, which starting at t_0 advances in time using the projection method with time step dt until T.

Its second parameter, verbose indicates whether the function should output information what it is doing at any given moment: we use the ConditionalOStream class to do that for us.

  template<int dim>
  void NavierStokesProjection<dim>::run(const bool verbose, const unsigned int output_interval) {
  ConditionalOStream verbose_cout(std::cout, verbose && Utilities::MPI::this_mpi_process(MPI_COMM_WORLD) == 0);
 
  output_results(1);
  double time = t_0 + dt;
  unsigned int n = 1;
  while(std::abs(T - time) > 1e-10) {
  time += dt;
  n++;
  pcout << "Step = " << n << " Time = " << time << std::endl;
 
  /*--- First stage of TR-BDF2 and we start by setting the proper flag ---*/
  TR_BDF2_stage = 1;
  navier_stokes_matrix.set_TR_BDF2_stage(TR_BDF2_stage);
  for(unsigned int level = 0; level < triangulation.n_global_levels(); ++level)
  mg_matrices[level].set_TR_BDF2_stage(TR_BDF2_stage);
 
  verbose_cout << " Interpolating the velocity stage 1" << std::endl;
  interpolate_velocity();
 
  verbose_cout << " Diffusion Step stage 1 " << std::endl;
  diffusion_step();
 
  verbose_cout << " Projection Step stage 1" << std::endl;
  project_grad(1);
  u_tmp.equ(gamma*dt, u_tmp);
  u_star += u_tmp; /*--- In the rhs of the projection step we need u_star + gamma*dt*grad(pres_n) and we save it into u_star ---*/
  projection_step();
 
  verbose_cout << " Updating the Velocity stage 1" << std::endl;
  u_n_gamma.equ(1.0, u_star);
  project_grad(2);
  grad_pres_int.equ(1.0, u_tmp); /*--- We save grad(pres_int), because we will need it soon ---*/
  u_tmp.equ(-gamma*dt, u_tmp);
  u_n_gamma += u_tmp; /*--- u_n_gamma = u_star - gamma*dt*grad(pres_int) ---*/
  u_n_minus_1 = u_n;
 
  /*--- Second stage of TR-BDF2 ---*/
  TR_BDF2_stage = 2;
  for(unsigned int level = 0; level < triangulation.n_global_levels(); ++level)
  mg_matrices[level].set_TR_BDF2_stage(TR_BDF2_stage);
  navier_stokes_matrix.set_TR_BDF2_stage(TR_BDF2_stage);
 
  verbose_cout << " Interpolating the velocity stage 2" << std::endl;
  interpolate_velocity();
 
  verbose_cout << " Diffusion Step stage 2 " << std::endl;
  diffusion_step();
 
  verbose_cout << " Projection Step stage 2" << std::endl;
  u_tmp.equ((1.0 - gamma)*dt, grad_pres_int);
  u_star += u_tmp; /*--- In the rhs of the projection step we need u_star + (1 - gamma)*dt*grad(pres_int) ---*/
  projection_step();
 
  verbose_cout << " Updating the Velocity stage 2" << std::endl;
  u_n.equ(1.0, u_star);
  project_grad(1);
  u_tmp.equ((gamma - 1.0)*dt, u_tmp);
  u_n += u_tmp; /*--- u_n = u_star - (1 - gamma)*dt*grad(pres_n) ---*/
 
  const double max_vel = get_maximal_velocity();
  pcout<< "Maximal velocity = " << max_vel << std::endl;
  /*--- The Courant number is computed taking into account the polynomial degree for the velocity ---*/
  pcout << "CFL = " << dt*max_vel*(EquationData::degree_p + 1)*
  compute_lift_and_drag();
  if(n % output_interval == 0) {
  verbose_cout << "Plotting Solution final" << std::endl;
  output_results(n);
  }
  /*--- In case dt is not a multiple of T, we reduce dt in order to end up at T ---*/
  if(T - time < dt && T - time > 1e-10) {
  dt = T - time;
  navier_stokes_matrix.set_dt(dt);
  for(unsigned int level = 0; level < triangulation.n_global_levels(); ++level)
  mg_matrices[level].set_dt(dt);
  }
  /*--- Perform the refinement if desired ---*/
  if(refinement_iterations > 0 && n % refinement_iterations == 0) {
  verbose_cout << "Refining mesh" << std::endl;
  refine_mesh();
  }
  }
  if(n % output_interval != 0) {
  verbose_cout << "Plotting Solution final" << std::endl;
  output_results(n);
  }
  if(refinement_iterations > 0) {
  for(unsigned int lev = 0; lev < triangulation.n_global_levels() - 1; ++ lev)
  interpolate_max_res(lev);
  save_max_res();
  }
  }
 
  } // namespace NS_TRBDF2
 
 
double minimal_cell_diameter(const Triangulation< dim, spacedim > &triangulation, const Mapping< dim, spacedim > &mapping=(ReferenceCells::get_hypercube< dim >() .template get_default_linear_mapping< dim, spacedim >()))
::VectorizedArray< Number, width > sqrt(const ::VectorizedArray< Number, width > &)

@sect{ The main function }

The main function looks very much like in all the other tutorial programs. We first initialize MPI, we initialize the class 'NavierStokesProjection' with the dimension as template parameter and then let the method 'run' do the job.

  int main(int argc, char *argv[]) {
  try {
  using namespace NS_TRBDF2;
 
  RunTimeParameters::Data_Storage data;
  data.read_data("parameter-file.prm");
 
  Utilities::MPI::MPI_InitFinalize mpi_init(argc, argv, -1);
 
  const auto& curr_rank = Utilities::MPI::this_mpi_process(MPI_COMM_WORLD);
  deallog.depth_console(data.verbose && curr_rank == 0 ? 2 : 0);
 
  NavierStokesProjection<2> test(data);
  test.run(data.verbose, data.output_interval);
 
  if(curr_rank == 0)
  std::cout << "----------------------------------------------------"
  << std::endl
  << "Apparently everything went fine!" << std::endl
  << "Don't forget to brush your teeth :-)" << std::endl
  << std::endl;
 
  return 0;
  }
  catch(std::exception &exc) {
  std::cerr << std::endl
  << std::endl
  << "----------------------------------------------------"
  << std::endl;
  std::cerr << "Exception on processing: " << std::endl
  << exc.what() << std::endl
  << "Aborting!" << std::endl
  << "----------------------------------------------------"
  << std::endl;
  return 1;
  }
  catch(...) {
  std::cerr << std::endl
  << std::endl
  << "----------------------------------------------------"
  << std::endl;
  std::cerr << "Unknown exception!" << std::endl
  << "Aborting!" << std::endl
  << "----------------------------------------------------"
  << std::endl;
  return 1;
  }
 
  }
unsigned int depth_console(const unsigned int n)
Definition logstream.cc:349
LogStream deallog
Definition logstream.cc:36

Annotated version of runtime_parameters.h

  /* -----------------------------------------------------------------------------
  *
  * SPDX-License-Identifier: LGPL-2.1-or-later
  * Copyright (C) 2022 by Giuseppe Orlando
  *
  * This file is part of the deal.II code gallery.
  *
  * -----------------------------------------------------------------------------
  */
 

We start by including all the necessary deal.II header files

  #include <deal.II/base/parameter_handler.h>
 

@sect{Run time parameters}

Since our method has several parameters that can be fine-tuned we put them into an external file, so that they can be determined at run-time.

  namespace RunTimeParameters {
  using namespace dealii;
 
  class Data_Storage {
  public:
  Data_Storage();
 
  void read_data(const std::string& filename);
 
  double initial_time;
  double final_time;
 
  double Reynolds;
  double dt;
 
  unsigned int n_refines; /*--- Number of refinements ---*/
  unsigned int max_loc_refinements; /*--- Number of maximum local refinements allowed ---*/
  unsigned int min_loc_refinements; /*--- Number of minimum local refinements allowed
  once reached that level ---*/
 
  /*--- Parameters related to the linear solver ---*/
  unsigned int max_iterations;
  double eps;
 
  bool verbose;
  unsigned int output_interval;
 
  std::string dir; /*--- Auxiliary string variable for output storage ---*/
 
  unsigned int refinement_iterations; /*--- Auxiliary variable about how many steps perform remeshing ---*/
 
  protected:
  };
 

In the constructor of this class we declare all the parameters in suitable (but arbitrary) subsections.

  Data_Storage::Data_Storage(): initial_time(0.0),
  final_time(1.0),
  Reynolds(1.0),
  dt(5e-4),
  n_refines(0),
  max_loc_refinements(0),
  min_loc_refinements(0),
  max_iterations(1000),
  eps(1e-12),
  verbose(true),
  output_interval(15),
  refinement_iterations(0) {
  prm.enter_subsection("Physical data");
  {
  prm.declare_entry("initial_time",
  "0.0",
  " The initial time of the simulation. ");
  prm.declare_entry("final_time",
  "1.0",
  " The final time of the simulation. ");
  prm.declare_entry("Reynolds",
  "1.0",
  " The Reynolds number. ");
  }
  prm.leave_subsection();
 
  prm.enter_subsection("Time step data");
  {
  prm.declare_entry("dt",
  "5e-4",
  " The time step size. ");
  }
  prm.leave_subsection();
 
  prm.enter_subsection("Space discretization");
  {
  prm.declare_entry("n_of_refines",
  "100",
  Patterns::Integer(0, 1500),
  " The number of cells we want on each direction of the mesh. ");
  prm.declare_entry("max_loc_refinements",
  "4",
  " The number of maximum local refinements. ");
  prm.declare_entry("min_loc_refinements",
  "2",
  " The number of minimum local refinements. ");
  }
  prm.leave_subsection();
 
  prm.enter_subsection("Data solve");
  {
  prm.declare_entry("max_iterations",
  "1000",
  Patterns::Integer(1, 30000),
  " The maximal number of iterations linear solvers must make. ");
  prm.declare_entry("eps",
  "1e-12",
  " The stopping criterion. ");
  }
  prm.leave_subsection();
 
  prm.declare_entry("refinement_iterations",
  "0",
  " This number indicates how often we need to "
  "refine the mesh");
 
  prm.declare_entry("saving directory", "SimTest");
 
  prm.declare_entry("verbose",
  "true",
  " This indicates whether the output of the solution "
  "process should be verbose. ");
 
  prm.declare_entry("output_interval",
  "1",
  " This indicates between how many time steps we print "
  "the solution. ");
  }
 

We need now a routine to read all declared parameters in the constructor

  void Data_Storage::read_data(const std::string& filename) {
  std::ifstream file(filename);
  AssertThrow(file, ExcFileNotOpen(filename));
 
  prm.parse_input(file);
 
  prm.enter_subsection("Physical data");
  {
  initial_time = prm.get_double("initial_time");
  final_time = prm.get_double("final_time");
  Reynolds = prm.get_double("Reynolds");
  }
  prm.leave_subsection();
 
  prm.enter_subsection("Time step data");
  {
  dt = prm.get_double("dt");
  }
  prm.leave_subsection();
 
  prm.enter_subsection("Space discretization");
  {
  n_refines = prm.get_integer("n_of_refines");
  max_loc_refinements = prm.get_integer("max_loc_refinements");
  min_loc_refinements = prm.get_integer("min_loc_refinements");
  }
  prm.leave_subsection();
 
  prm.enter_subsection("Data solve");
  {
  max_iterations = prm.get_integer("max_iterations");
  eps = prm.get_double("eps");
  }
  prm.leave_subsection();
 
  dir = prm.get("saving directory");
 
  refinement_iterations = prm.get_integer("refinement_iterations");
 
  verbose = prm.get_bool("verbose");
 
  output_interval = prm.get_integer("output_interval");
  }
 
  } // namespace RunTimeParameters