MFEM v4.7.0
Finite element discretization library
|
Class for standard nodal finite elements. More...
#include <fe_base.hpp>
Public Member Functions | |
NodalFiniteElement (int D, Geometry::Type G, int Do, int O, int F=FunctionSpace::Pk) | |
Construct NodalFiniteElement with given. | |
const DofToQuad & | GetDofToQuad (const IntegrationRule &ir, DofToQuad::Mode mode) const override |
Return a DofToQuad structure corresponding to the given IntegrationRule using the given DofToQuad::Mode. | |
void | GetLocalInterpolation (ElementTransformation &Trans, DenseMatrix &I) const override |
Return the local interpolation matrix I (Dof x Dof) where the fine element is the image of the base geometry under the given transformation. | |
void | GetLocalRestriction (ElementTransformation &Trans, DenseMatrix &R) const override |
Return a local restriction matrix R (Dof x Dof) mapping fine dofs to coarse dofs. | |
void | GetTransferMatrix (const FiniteElement &fe, ElementTransformation &Trans, DenseMatrix &I) const override |
Return interpolation matrix, I, which maps dofs from a coarse element, fe, to the fine dofs on this finite element. | |
void | Project (Coefficient &coeff, ElementTransformation &Trans, Vector &dofs) const override |
Given a coefficient and a transformation, compute its projection (approximation) in the local finite dimensional space in terms of the degrees of freedom. | |
void | Project (VectorCoefficient &vc, ElementTransformation &Trans, Vector &dofs) const override |
Given a vector coefficient and a transformation, compute its projection (approximation) in the local finite dimensional space in terms of the degrees of freedom. (VectorFiniteElements) | |
void | ProjectMatrixCoefficient (MatrixCoefficient &mc, ElementTransformation &T, Vector &dofs) const override |
Given a matrix coefficient and a transformation, compute an approximation ("projection") in the local finite dimensional space in terms of the degrees of freedom. For VectorFiniteElements, the rows of the coefficient are projected in the vector space. | |
void | Project (const FiniteElement &fe, ElementTransformation &Trans, DenseMatrix &I) const override |
Compute the embedding/projection matrix from the given FiniteElement onto 'this' FiniteElement. The ElementTransformation is included to support cases when the projection depends on it. | |
void | ProjectGrad (const FiniteElement &fe, ElementTransformation &Trans, DenseMatrix &grad) const override |
Compute the discrete gradient matrix from the given FiniteElement onto 'this' FiniteElement. The ElementTransformation is included to support cases when the matrix depends on it. | |
void | ProjectDiv (const FiniteElement &fe, ElementTransformation &Trans, DenseMatrix &div) const override |
Compute the discrete divergence matrix from the given FiniteElement onto 'this' FiniteElement. The ElementTransformation is included to support cases when the matrix depends on it. | |
const Array< int > & | GetLexicographicOrdering () const |
Get an Array<int> that maps lexicographically ordered indices to the indices of the respective nodes/dofs/basis functions. | |
Public Member Functions inherited from mfem::ScalarFiniteElement | |
ScalarFiniteElement (int D, Geometry::Type G, int Do, int O, int F=FunctionSpace::Pk) | |
Construct ScalarFiniteElement with given. | |
virtual void | SetMapType (int M) |
Set the FiniteElement::MapType of the element to either VALUE or INTEGRAL. Also sets the FiniteElement::DerivType to GRAD if the FiniteElement::MapType is VALUE. | |
void | NodalLocalInterpolation (ElementTransformation &Trans, DenseMatrix &I, const ScalarFiniteElement &fine_fe) const |
Get the matrix I that defines nodal interpolation between this element and the refined element fine_fe. | |
void | ScalarLocalInterpolation (ElementTransformation &Trans, DenseMatrix &I, const ScalarFiniteElement &fine_fe) const |
Get matrix I "Interpolation" defined through local L2-projection in the space defined by the fine_fe. | |
void | ScalarLocalL2Restriction (ElementTransformation &Trans, DenseMatrix &R, const ScalarFiniteElement &coarse_fe) const |
Get restriction matrix R defined through local L2-projection in the space defined by the coarse_fe. | |
Public Member Functions inherited from mfem::FiniteElement | |
FiniteElement (int D, Geometry::Type G, int Do, int O, int F=FunctionSpace::Pk) | |
Construct FiniteElement with given. | |
int | GetDim () const |
Returns the reference space dimension for the finite element. | |
int | GetRangeDim () const |
Returns the vector dimension for vector-valued finite elements, which is also the dimension of the interpolation operation. | |
int | GetCurlDim () const |
Returns the dimension of the curl for vector-valued finite elements. | |
Geometry::Type | GetGeomType () const |
Returns the Geometry::Type of the reference element. | |
int | GetDof () const |
Returns the number of degrees of freedom in the finite element. | |
int | GetOrder () const |
Returns the order of the finite element. In the case of anisotropic orders, returns the maximum order. | |
bool | HasAnisotropicOrders () const |
Returns true if the FiniteElement basis may be using different orders/degrees in different spatial directions. | |
const int * | GetAnisotropicOrders () const |
Returns an array containing the anisotropic orders/degrees. | |
int | Space () const |
Returns the type of FunctionSpace on the element. | |
int | GetRangeType () const |
Returns the FiniteElement::RangeType of the element, one of {SCALAR, VECTOR}. | |
int | GetDerivRangeType () const |
Returns the FiniteElement::RangeType of the element derivative, either SCALAR or VECTOR. | |
int | GetMapType () const |
Returns the FiniteElement::MapType of the element describing how reference functions are mapped to physical space, one of {VALUE, INTEGRAL H_DIV, H_CURL}. | |
int | GetDerivType () const |
Returns the FiniteElement::DerivType of the element describing the spatial derivative method implemented, one of {NONE, GRAD, DIV, CURL}. | |
int | GetDerivMapType () const |
Returns the FiniteElement::DerivType of the element describing how reference function derivatives are mapped to physical space, one of {VALUE, INTEGRAL, H_DIV, H_CURL}. | |
virtual void | CalcShape (const IntegrationPoint &ip, Vector &shape) const =0 |
Evaluate the values of all shape functions of a scalar finite element in reference space at the given point ip. | |
void | CalcPhysShape (ElementTransformation &Trans, Vector &shape) const |
Evaluate the values of all shape functions of a scalar finite element in physical space at the point described by Trans. | |
virtual void | CalcDShape (const IntegrationPoint &ip, DenseMatrix &dshape) const =0 |
Evaluate the gradients of all shape functions of a scalar finite element in reference space at the given point ip. | |
void | CalcPhysDShape (ElementTransformation &Trans, DenseMatrix &dshape) const |
Evaluate the gradients of all shape functions of a scalar finite element in physical space at the point described by Trans. | |
const IntegrationRule & | GetNodes () const |
Get a const reference to the nodes of the element. | |
virtual void | CalcVShape (const IntegrationPoint &ip, DenseMatrix &shape) const |
Evaluate the values of all shape functions of a vector finite element in reference space at the given point ip. | |
virtual void | CalcVShape (ElementTransformation &Trans, DenseMatrix &shape) const |
Evaluate the values of all shape functions of a vector finite element in physical space at the point described by Trans. | |
void | CalcPhysVShape (ElementTransformation &Trans, DenseMatrix &shape) const |
Equivalent to the CalcVShape() method with the same arguments. | |
virtual void | CalcDivShape (const IntegrationPoint &ip, Vector &divshape) const |
Evaluate the divergence of all shape functions of a vector finite element in reference space at the given point ip. | |
void | CalcPhysDivShape (ElementTransformation &Trans, Vector &divshape) const |
Evaluate the divergence of all shape functions of a vector finite element in physical space at the point described by Trans. | |
virtual void | CalcCurlShape (const IntegrationPoint &ip, DenseMatrix &curl_shape) const |
Evaluate the curl of all shape functions of a vector finite element in reference space at the given point ip. | |
virtual void | CalcPhysCurlShape (ElementTransformation &Trans, DenseMatrix &curl_shape) const |
Evaluate the curl of all shape functions of a vector finite element in physical space at the point described by Trans. | |
virtual void | GetFaceDofs (int face, int **dofs, int *ndofs) const |
Get the dofs associated with the given face. *dofs is set to an internal array of the local dofc on the face, while *ndofs is set to the number of dofs on that face. | |
virtual void | CalcHessian (const IntegrationPoint &ip, DenseMatrix &Hessian) const |
Evaluate the Hessians of all shape functions of a scalar finite element in reference space at the given point ip. | |
virtual void | CalcPhysHessian (ElementTransformation &Trans, DenseMatrix &Hessian) const |
Evaluate the Hessian of all shape functions of a scalar finite element in reference space at the given point ip. | |
virtual void | CalcPhysLaplacian (ElementTransformation &Trans, Vector &Laplacian) const |
Evaluate the Laplacian of all shape functions of a scalar finite element in reference space at the given point ip. | |
virtual void | CalcPhysLinLaplacian (ElementTransformation &Trans, Vector &Laplacian) const |
virtual void | ProjectFromNodes (Vector &vc, ElementTransformation &Trans, Vector &dofs) const |
Given a vector of values at the finite element nodes and a transformation, compute its projection (approximation) in the local finite dimensional space in terms of the degrees of freedom. Valid for VectorFiniteElements. | |
virtual void | ProjectDelta (int vertex, Vector &dofs) const |
Project a delta function centered on the given vertex in the local finite dimensional space represented by the dofs. | |
virtual void | ProjectCurl (const FiniteElement &fe, ElementTransformation &Trans, DenseMatrix &curl) const |
Compute the discrete curl matrix from the given FiniteElement onto 'this' FiniteElement. The ElementTransformation is included to support cases when the matrix depends on it. | |
virtual void | GetFaceMap (const int face_id, Array< int > &face_map) const |
Return the mapping from lexicographic face DOFs to lexicographic element DOFs for the given local face face_id. | |
virtual const StatelessDofTransformation * | GetDofTransformation () const |
Return a DoF transformation object for this particular type of basis. | |
virtual | ~FiniteElement () |
Deconstruct the FiniteElement. | |
Protected Member Functions | |
void | ProjectCurl_2D (const FiniteElement &fe, ElementTransformation &Trans, DenseMatrix &curl) const |
Protected Attributes | |
Array< int > | lex_ordering |
Protected Attributes inherited from mfem::FiniteElement | |
int | dim |
Dimension of reference space. | |
int | vdim |
Vector dimension of vector-valued basis functions. | |
int | cdim |
Dimension of curl for vector-valued basis functions. | |
Geometry::Type | geom_type |
Geometry::Type of the reference element. | |
int | func_space |
int | range_type |
int | map_type |
int | deriv_type |
int | deriv_range_type |
int | deriv_map_type |
int | dof |
Number of degrees of freedom. | |
int | order |
Order/degree of the shape functions. | |
int | orders [Geometry::MaxDim] |
Anisotropic orders. | |
IntegrationRule | Nodes |
DenseMatrix | vshape |
Array< DofToQuad * > | dof2quad_array |
Container for all DofToQuad objects created by the FiniteElement. | |
Additional Inherited Members | |
Public Types inherited from mfem::FiniteElement | |
enum | RangeType { UNKNOWN_RANGE_TYPE = -1 , SCALAR , VECTOR } |
Enumeration for range_type and deriv_range_type. More... | |
enum | MapType { UNKNOWN_MAP_TYPE = -1 , VALUE , INTEGRAL , H_DIV , H_CURL } |
Enumeration for MapType: defines how reference functions are mapped to physical space. More... | |
enum | DerivType { NONE , GRAD , DIV , CURL } |
Enumeration for DerivType: defines which derivative method is implemented. More... | |
Static Public Member Functions inherited from mfem::FiniteElement | |
static bool | IsClosedType (int b_type) |
Return true if the BasisType of b_type is closed (has Quadrature1D points on the boundary). | |
static bool | IsOpenType (int b_type) |
Return true if the BasisType of b_type is open (doesn't have Quadrature1D points on the boundary). | |
static int | VerifyClosed (int b_type) |
Ensure that the BasisType of b_type is closed (has Quadrature1D points on the boundary). | |
static int | VerifyOpen (int b_type) |
Ensure that the BasisType of b_type is open (doesn't have Quadrature1D points on the boundary). | |
static int | VerifyNodal (int b_type) |
Ensure that the BasisType of b_type nodal (satisfies the interpolation property). | |
Static Protected Member Functions inherited from mfem::ScalarFiniteElement | |
static const ScalarFiniteElement & | CheckScalarFE (const FiniteElement &fe) |
Class for standard nodal finite elements.
Definition at line 714 of file fe_base.hpp.
|
inline |
Construct NodalFiniteElement with given.
D | Reference space dimension |
G | Geometry type (of type Geometry::Type) |
Do | Number of degrees of freedom in the FiniteElement |
O | Order/degree of the FiniteElement |
F | FunctionSpace type of the FiniteElement |
Definition at line 733 of file fe_base.hpp.
|
overridevirtual |
Return a DofToQuad structure corresponding to the given IntegrationRule using the given DofToQuad::Mode.
See the documentation for DofToQuad for more details.
Reimplemented from mfem::FiniteElement.
Reimplemented in mfem::NodalTensorFiniteElement.
Definition at line 703 of file fe_base.cpp.
|
inline |
Get an Array<int> that maps lexicographically ordered indices to the indices of the respective nodes/dofs/basis functions.
Lexicographic ordering of nodes is defined in terms of reference-space coordinates (x,y,z). Lexicographically ordered nodes are listed first in order of increasing x-coordinate, and then in order of increasing y-coordinate, and finally in order of increasing z-coordinate.
For example, the six nodes of a quadratic triangle are lexicographically ordered as follows:
5 |\ 3 4 | \ 0-1-2
The resulting array may be empty if the DOFs are already ordered lexicographically, or if the finite element does not support creating this permutation. The array returned is the same as the array given by TensorBasisElement::GetDofMap, but it is also available for non-tensor elements.
Definition at line 795 of file fe_base.hpp.
|
inlineoverridevirtual |
Return the local interpolation matrix I (Dof x Dof) where the fine element is the image of the base geometry under the given transformation.
Reimplemented from mfem::FiniteElement.
Definition at line 740 of file fe_base.hpp.
|
overridevirtual |
Return a local restriction matrix R (Dof x Dof) mapping fine dofs to coarse dofs.
The fine element is the image of the base geometry under the given transformation, Trans.
The assumption in this method is that a subset of the coarse dofs can be expressed only in terms of the dofs of the given fine element.
Rows in R corresponding to coarse dofs that cannot be expressed in terms of the fine dofs will be marked as invalid by setting the first entry (column 0) in the row to infinity().
This method assumes that the dimensions of R are set before it is called.
Reimplemented from mfem::FiniteElement.
Definition at line 764 of file fe_base.cpp.
|
inlineoverridevirtual |
Return interpolation matrix, I, which maps dofs from a coarse element, fe, to the fine dofs on this finite element.
Trans represents the mapping from the reference element of this element into a subset of the reference space of the element fe, thus allowing the "coarse" FiniteElement to be different from the "fine" FiniteElement as when h-refinement is combined with p-refinement or p-derefinement. It is assumed that both finite elements use the same FiniteElement::MapType.
Reimplemented from mfem::FiniteElement.
Reimplemented in mfem::NodalTensorFiniteElement.
Definition at line 747 of file fe_base.hpp.
|
overridevirtual |
Given a coefficient and a transformation, compute its projection (approximation) in the local finite dimensional space in terms of the degrees of freedom.
The approximation used to project is usually local interpolation of degrees of freedom. The derived class could use other methods not implemented yet, e.g. local L2 projection.
Reimplemented from mfem::FiniteElement.
Definition at line 796 of file fe_base.cpp.
|
overridevirtual |
Compute the embedding/projection matrix from the given FiniteElement onto 'this' FiniteElement. The ElementTransformation is included to support cases when the projection depends on it.
Reimplemented from mfem::FiniteElement.
Definition at line 857 of file fe_base.cpp.
|
overridevirtual |
Given a vector coefficient and a transformation, compute its projection (approximation) in the local finite dimensional space in terms of the degrees of freedom. (VectorFiniteElements)
The approximation used to project is usually local interpolation of degrees of freedom. The derived class could use other methods not implemented yet, e.g. local L2 projection.
Reimplemented from mfem::FiniteElement.
Definition at line 813 of file fe_base.cpp.
|
protected |
Definition at line 724 of file fe_base.cpp.
|
overridevirtual |
Compute the discrete divergence matrix from the given FiniteElement onto 'this' FiniteElement. The ElementTransformation is included to support cases when the matrix depends on it.
Reimplemented from mfem::FiniteElement.
Definition at line 945 of file fe_base.cpp.
|
overridevirtual |
Compute the discrete gradient matrix from the given FiniteElement onto 'this' FiniteElement. The ElementTransformation is included to support cases when the matrix depends on it.
Reimplemented from mfem::FiniteElement.
Definition at line 916 of file fe_base.cpp.
|
overridevirtual |
Given a matrix coefficient and a transformation, compute an approximation ("projection") in the local finite dimensional space in terms of the degrees of freedom. For VectorFiniteElements, the rows of the coefficient are projected in the vector space.
Reimplemented from mfem::FiniteElement.
Definition at line 835 of file fe_base.cpp.
|
protected |
Definition at line 720 of file fe_base.hpp.