Line data    Source code

       1             : // The libMesh Finite Element Library.
       2             : // Copyright (C) 2002-2025 Benjamin S. Kirk, John W. Peterson, Roy H. Stogner
       3             : 
       4             : // This library is free software; you can redistribute it and/or
       5             : // modify it under the terms of the GNU Lesser General Public
       6             : // License as published by the Free Software Foundation; either
       7             : // version 2.1 of the License, or (at your option) any later version.
       8             : 
       9             : // This library is distributed in the hope that it will be useful,
      10             : // but WITHOUT ANY WARRANTY; without even the implied warranty of
      11             : // MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
      12             : // Lesser General Public License for more details.
      13             : 
      14             : // You should have received a copy of the GNU Lesser General Public
      15             : // License along with this library; if not, write to the Free Software
      16             : // Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
      17             : 
      18             : 
      19             : 
      20             : #ifndef LIBMESH_ERROR_ESTIMATOR_H
      21             : #define LIBMESH_ERROR_ESTIMATOR_H
      22             : 
      23             : // Local Includes
      24             : #include "libmesh/libmesh_common.h"
      25             : #include "libmesh/system_norm.h"
      26             : 
      27             : // C++ includes
      28             : #include <cstddef>
      29             : #include <map>
      30             : #include <string>
      31             : #include <vector>
      32             : #include <memory>
      33             : 
      34             : namespace libMesh
      35             : {
      36             : 
      37             : // Forward Declarations
      38             : class ErrorVector;
      39             : class EquationSystems;
      40             : class System;
      41             : template <typename T> class NumericVector;
      42             : 
      43             : namespace Parallel {
      44             :   class Communicator;
      45             : }
      46             : 
      47             : enum ErrorEstimatorType : int;
      48             : 
      49             : /**
      50             :  * This class holds functions that will estimate the error
      51             :  * in a finite element solution on a given mesh.  These error
      52             :  * estimates can be useful in their own right, or may be used
      53             :  * to guide adaptive mesh refinement.
      54             :  *
      55             :  * \note The computed errors are stored as floats rather
      56             :  * than doubles since the required precision is low.
      57             :  *
      58             :  * \author Benjamin S. Kirk
      59             :  * \date 2003
      60             :  */
      61             : class ErrorEstimator
      62             : {
      63             : public:
      64             : 
      65             :   /**
      66             :    * Constructor.  Empty.  Derived classes should reset error_norm as
      67             :    * appropriate.
      68             :    */
      69       35034 :   ErrorEstimator() = default;
      70             : 
      71             :   /**
      72             :    * Copy/move ctor, copy/move assignment operator, and destructor are
      73             :    * all explicitly defaulted for this simple class.
      74             :    */
      75             :   ErrorEstimator (const ErrorEstimator &) = default;
      76             :   ErrorEstimator (ErrorEstimator &&) = default;
      77             :   ErrorEstimator & operator= (const ErrorEstimator &) = default;
      78             :   ErrorEstimator & operator= (ErrorEstimator &&) = default;
      79        5456 :   virtual ~ErrorEstimator() = default;
      80             : 
      81             : 
      82             :   /**
      83             :    * This pure virtual function must be redefined
      84             :    * in derived classes to compute the error for each
      85             :    * active element and place it in the "error_per_cell" vector.
      86             :    *
      87             :    * If solution_vector is not nullptr, the estimator will
      88             :    * (if able) attempt to estimate an error in that field
      89             :    * instead of in system.solution.
      90             :    *
      91             :    * If estimate_parent_error is not false, the estimator will (if
      92             :    * able) attempt to give a consistent estimate of errors in parent
      93             :    * elements that would be generated by coarsening.
      94             :    */
      95             :   virtual void estimate_error (const System & system,
      96             :                                ErrorVector & error_per_cell,
      97             :                                const NumericVector<Number> * solution_vector = nullptr,
      98             :                                bool estimate_parent_error = false) = 0;
      99             : 
     100             :   /**
     101             :    * This virtual function can be redefined
     102             :    * in derived classes, but by default computes the sum of
     103             :    * the error_per_cell for each system in the equation_systems.
     104             :    *
     105             :    * Currently this function ignores the error_norm member variable,
     106             :    * and uses the function argument error_norms instead.
     107             :    *
     108             :    * This function is named estimate_errors instead of estimate_error
     109             :    * because otherwise C++ can get confused.
     110             :    */
     111             :   virtual void estimate_errors (const EquationSystems & equation_systems,
     112             :                                 ErrorVector & error_per_cell,
     113             :                                 const std::map<const System *, SystemNorm> & error_norms,
     114             :                                 const std::map<const System *, const NumericVector<Number> *> * solution_vectors = nullptr,
     115             :                                 bool estimate_parent_error = false);
     116             : 
     117             :   /**
     118             :    * When calculating many error vectors at once, we need a data structure to
     119             :    * hold them all
     120             :    */
     121             :   typedef std::map<std::pair<const System *, unsigned int>, std::unique_ptr<ErrorVector>> ErrorMap;
     122             : 
     123             :   /**
     124             :    * This virtual function can be redefined
     125             :    * in derived classes, but by default it calls estimate_error
     126             :    * repeatedly to calculate the requested error vectors.
     127             :    *
     128             :    * Currently this function ignores the error_norm.weight() values
     129             :    * because it calculates each variable's error individually, unscaled.
     130             :    *
     131             :    * The user selects which errors get computed by filling a map with error
     132             :    * vectors: If errors_per_cell[&system][v] exists, it will be filled with the
     133             :    * error values in variable \p v of \p system
     134             :    */
     135             :   virtual void estimate_errors (const EquationSystems & equation_systems,
     136             :                                 ErrorMap & errors_per_cell,
     137             :                                 const std::map<const System *, const NumericVector<Number> *> * solution_vectors = nullptr,
     138             :                                 bool estimate_parent_error = false);
     139             : 
     140             :   /**
     141             :    * \returns The type for the ErrorEstimator subclass.
     142             :    */
     143             :   virtual ErrorEstimatorType type() const = 0;
     144             : 
     145             :   /**
     146             :    * When estimating the error in a single system, the \p error_norm
     147             :    * is used to control the scaling and norm choice for each variable.
     148             :    * Not all estimators will support all norm choices.  The default
     149             :    * scaling is for all variables to be weighted equally.  The default
     150             :    * norm choice depends on the error estimator.
     151             :    *
     152             :    * Part of this functionality was supported via component_scale and
     153             :    * sobolev_order in older libMesh versions, and a small part was
     154             :    * supported via component_mask in even older versions.  Hopefully
     155             :    * the encapsulation here will allow us to avoid changing this API
     156             :    * again.
     157             :    */
     158             :   SystemNorm error_norm;
     159             : 
     160             : protected:
     161             : 
     162             :   /**
     163             :    * This method takes the local error contributions in
     164             :    * \p error_per_cell from each processor and combines
     165             :    * them to get the global error vector.
     166             :    */
     167             :   void reduce_error (std::vector<ErrorVectorReal> & error_per_cell,
     168             :                      const Parallel::Communicator & comm) const;
     169             : };
     170             : 
     171             : 
     172             : } // namespace libMesh
     173             : 
     174             : #endif // LIBMESH_ERROR_ESTIMATOR_H