Detailed Description

template<class T>
class gismo::gsElasticityAssembler< T >

Assembles the stiffness matrix and the right-hand side vector for linear and nonlinear elasticity for 2D plain stress and 3D continua. The matrix and vector have a block structure associated with components of the displacement vector, each block corresponding to one component. Supports mixed displacement-pressure formulation.

Inheritance diagram for gsElasticityAssembler< T >:

Collaboration diagram for gsElasticityAssembler< T >:

Public Member Functions
const std::vector< gsMatrix< T > > &	allFixedDofs () const
	Returns all the Dirichlet values (if applicable)

virtual bool	assemble (const gsMatrix< T > &solutionVector, const std::vector< gsMatrix< T > > &fixedDoFs)

bool	check ()
	checks for consistency and legal values of the stored members.

virtual gsAssembler *	clone () const
	Clone this Assembler, making a deep copy.

void	computeDirichletDofs (short_t unk=0)
	Triggers computation of the Dirichlet dofs.

virtual void	constructCauchyStresses (const gsMultiPatch< T > &displacement, const gsMultiPatch< T > &pressure, gsPiecewiseFunction< T > &result, stress_components::components component=stress_components::von_mises) const
	Construct Cauchy stresses for evaluation or visualization.

virtual void	constructCauchyStresses (const gsMultiPatch< T > &displacement, gsPiecewiseFunction< T > &result, stress_components::components component=stress_components::von_mises) const
	Construct Cauchy stresses for evaluation or visualization.

virtual void	constructPressure (const gsMatrix< T > &solVector, const std::vector< gsMatrix< T > > &fixedDoFs, gsMultiPatch< T > &pressure) const
	@ brief Construct pressure from computed solution vector

virtual void	constructSolution (const gsMatrix< T > &, gsMultiPatch< T > &, const gsVector< index_t > &) const
	Construct solution from computed solution vector for a set of unknowns. The result is a vectorfield, where each component is given the corresponding entry of.

virtual void	constructSolution (const gsMatrix< T > &, gsMultiPatch< T > &, short_t=0) const
	Construct solution from computed solution vector for a single unknows.

virtual void	constructSolution (const gsMatrix< T > &solVector, const std::vector< gsMatrix< T > > &fixedDDofs, gsMultiPatch< T > &result, const gsVector< index_t > &unknowns) const
	Constructs solution as a gsMultiPatch object from the solution vector and fixed DoFs.

virtual void	constructSolution (const gsMatrix< T > &solVector, const std::vector< gsMatrix< T > > &fixedDoFs, gsMultiPatch< T > &displacement) const
	Construct displacement from computed solution vector and fixed degrees of freedom.

virtual void	constructSolution (const gsMatrix< T > &solVector, const std::vector< gsMatrix< T > > &fixedDoFs, gsMultiPatch< T > &displacement, gsMultiPatch< T > &pressure) const
	Construct displacement and pressure from computed solution vector and fixed degrees of freedom.

virtual gsAssembler *	create () const
	Create an empty Assembler of the derived type and return a pointer to it. Call the initialize functions to set the members.

virtual void	eliminateFixedDofs ()
	Eliminates new Dirichelt degrees of fredom.

void	finalize ()
	finishes the assembling of the system matrix, i.e. calls its .makeCompressed() method.

const gsMatrix< T > &	fixedDofs (short_t unk=0) const
	Returns the Dirichlet values for a unknown (if applicable)

virtual void	getFixedDofs (index_t patch, boxSide side, gsMatrix< T > &ddofs) const

	gsElasticityAssembler (const gsMultiPatch< T > &patches, const gsMultiBasis< T > &basis, const gsBoundaryConditions< T > &bconditions, const gsFunction< T > &body_force)
	Constructor for displacement formulation.

	gsElasticityAssembler (const gsMultiPatch< T > &patches, const gsMultiBasis< T > &basisDisp, const gsMultiBasis< T > &basisPres, const gsBoundaryConditions< T > &bconditions, const gsFunction< T > &body_force)
	Constructor of mixed formulation (displacement + pressure)

void	homogenizeFixedDofs (short_t unk=0)
	Sets any Dirichlet values to homogeneous (if applicable)

void	initialize (const gsPde< T > &pde, const gsBasisRefs< T > &basis, const gsOptionList &opt=defaultOptions())
	Intitialize function for, sets data fields using the pde, a vector of bases and assembler options.

void	initialize (const gsPde< T > &pde, const gsMultiBasis< T > &bases, const gsOptionList &opt=defaultOptions())
	Intitialize function for, sets data fields using the pde, a multi-basis and assembler options.

void	initialize (const gsPde< T > &pde, const gsStdVectorRef< gsMultiBasis< T > > &bases, const gsOptionList &opt=defaultOptions())
	Intitialize function for, sets data fields using the pde, a vector of multi-basis and assembler options.

void	initialize (typename gsPde< T >::Ptr pde, const gsStdVectorRef< gsMultiBasis< T > > &bases, const gsOptionList &opt=defaultOptions())
	Intitialize function for, sets data fields using the pde, a vector of multi-basis and assembler options.

const gsSparseMatrix< T > &	matrix () const
	Returns the left-hand global matrix.

gsMultiBasis< T > &	multiBasis (index_t k=0)
	Return the multi-basis. Note: if the basis is altered, then refresh() should be called.

const gsMultiBasis< T > &	multiBasis (index_t k=0) const
	Return the multi-basis.

index_t	numColNz () const
	Provides an estimation of the number of non-zero matrix entries per column. This value can be used for sparse matrix memory allocation.

virtual int	numDofs () const
	Returns number of free degrees of freedom.

virtual index_t	numFixedDofs () const
	get the size of the Dirichlet vector for elimination

size_t	numMultiBasis () const
	Returns the number of multi-bases.

const gsMultiPatch< T > &	patches () const
	Return the multipatch.

const gsPde< T > &	pde () const
	Return the Pde.

void	penalizeDirichletDofs (short_t unk=0)

T	penalty (index_t k) const
	Penalty constant for patch k, used for Nitsche and / Discontinuous Galerkin methods.

template<class ElementVisitor >
void	push ()
	Iterates over all elements of the domain and applies the ElementVisitor.

template<class BElementVisitor >
void	push (const bcContainer &BCs)
	Iterates over all elements of the boundaries BCs and applies the BElementVisitor.

template<class BElementVisitor >
void	push (const BElementVisitor &visitor, const boundary_condition< T > &BC)
	Applies the BElementVisitor to the boundary condition BC.

template<class ElementVisitor >
void	push (const ElementVisitor &visitor)
	Iterates over all elements of the domain and applies the ElementVisitor.

template<class InterfaceVisitor >
void	pushInterface ()
	Iterates over all elements of interfaces and applies the InterfaceVisitor.

virtual void	refresh ()
	Refresh routine to set dof-mappers.

const gsMatrix< T > &	rhs () const
	Returns the left-hand side vector(s) ( multiple right hand sides possible )

void	setFixedDofs (const gsMatrix< T > &coefMatrix, short_t unk=0, size_t patch=0)
	the user can manually set the dirichlet Dofs for a given patch and unknown, based on the Basis coefficients

virtual void	setFixedDofs (const std::vector< gsMatrix< T > > &ddofs)
	set all fixed degrees of freedom

virtual void	setFixedDofs (index_t patch, boxSide side, const gsMatrix< T > &ddofs, bool oneUnk=false)
	Set Dirichet degrees of freedom on a given side of a given patch from a given matrix.

void	setFixedDofVector (gsMatrix< T > vals, short_t unk=0)
	the user can manually set the dirichlet Dofs for a given patch and unknown.

void	setSparseSystem (gsSparseSystem< T > &sys)
	Swaps the actual sparse system with the given one.

const gsSparseSystem< T > &	system () const
	Returns the left-hand global matrix.

virtual void	updateSolution (const gsMatrix< T > &solVector, gsMultiPatch< T > &result, T theta=(T)(1)) const
	Update solution by adding the computed solution vector to the current solution specified by.

Static Public Member Functions
static gsOptionList	defaultOptions ()
	Returns the list of default options for assembly.

Protected Member Functions
template<class ElementVisitor >
void	apply (ElementVisitor &visitor, size_t patchIndex=0, boxSide side=boundary::none)
	Generic assembly routine for volume or boundary integrals.

template<class InterfaceVisitor >
void	apply (InterfaceVisitor &visitor, const boundaryInterface &bi)
	Generic assembly routine for patch-interface integrals.

virtual void	assemble ()
	Main assemble routine, to be implemented in derived classes.

virtual void	assemble (bool)
	assembly procedure for linear problems

virtual bool	assemble (const gsMatrix< T > &solutionVector, const std::vector< gsMatrix< T > > &fixedDDoFs)=0

virtual void	assemble (const gsMultiPatch< T > &)
	Main non-linear assemble routine with input from current solution.

virtual void	assemble (const gsMultiPatch< T > &displacement)

virtual void	assemble (const gsMultiPatch< T > &displacement, const gsMultiPatch< T > &pressure)

void	computeDirichletDofsIntpl (const gsDofMapper &mapper, const gsMultiBasis< T > &mbasis, const short_t unk_=0)
	calculates the values of the eliminated dofs based on Interpolation.

void	computeDirichletDofsL2Proj (const gsDofMapper &mapper, const gsMultiBasis< T > &mbasis, const short_t unk_=0)
	calculates the values of the eliminated dofs based on L2 Projection.

virtual void	reserve ()
	a custom reserve function to allocate memory for the sparse matrix

void	scalarProblemGalerkinRefresh ()
	A prototype of the refresh function for a "standard" scalar problem. Creats one mapper based on the set options and initializes the sparse system (without allocating memory.

Protected Attributes
std::vector< gsMultiBasis< T > >	m_bases

std::vector< gsMatrix< T > >	m_ddof

short_t	m_dim

gsOptionList	m_options
	Options.

memory::shared_ptr< gsPde< T > >	m_pde_ptr

gsSparseSystem< T >	m_system
	Global sparse linear system.

Member Function Documentation

◆ apply()

template<class T >

template<class ElementVisitor >

void apply	(	ElementVisitor &	visitor,
		size_t	patchIndex = `0`,
		boxSide	side = `boundary::none`
	)

protectedinherited

Generic assembly routine for volume or boundary integrals.

Parameters

[in]	visitor	The visitor for the boundary or volume integral
[in]	patchIndex	The considered patch
[in]	side	The considered boundary side, only necessary for boundary integrals.

◆ assemble() [1/4]

template<class T >

virtual bool assemble	(	const gsMatrix< T > &	solutionVector,
		const std::vector< gsMatrix< T > > &	fixedDDoFs
	)

protectedvirtual

Assembles the tangential linear system for Newton's method given the current solution in the form of free and fixed/Dirichelt degrees of freedom. Checks if the current solution is valid (Newton's solver can exit safely if invalid).

Implements gsBaseAssembler< T >.

Reimplemented in gsMuscleAssembler< T >, gsThermoAssembler< T >, gsMuscleAssembler< T >, gsMuscleAssembler< T >, and gsThermoAssembler< T >.

◆ assemble() [2/4]

template<class T >

bool assemble	(	const gsMatrix< T > &	solutionVector,
		const std::vector< gsMatrix< T > > &	fixedDoFs
	)

virtual

Assembles the tangential linear system for Newton's method given the current solution in the form of free and fixed/Dirichelt degrees of freedom. Checks if the current solution is valid (Newton's solver can exit safely if invalid).

Implements gsBaseAssembler< T >.

Reimplemented in gsMuscleAssembler< T >, gsThermoAssembler< T >, gsMuscleAssembler< T >, gsMuscleAssembler< T >, and gsThermoAssembler< T >.

◆ assemble() [3/4]

template<class T >

void assemble ( const gsMultiPatch< T > & displacement )

protectedvirtual

@ brief Assembles the tangential matrix and the residual for a iteration of Newton's method for displacement formulation; set assembleMatrix to false to only assemble the residual; ATTENTION: rhs() returns a negative residual (-r) !!!

Reimplemented from gsBaseAssembler< T >.

Reimplemented in gsMuscleAssembler< T >, gsThermoAssembler< T >, gsMuscleAssembler< T >, and gsThermoAssembler< T >.

◆ assemble() [4/4]

template<class T >

void assemble	(	const gsMultiPatch< T > &	displacement,
		const gsMultiPatch< T > &	pressure
	)

protectedvirtual

@ brief Assembles the tangential matrix and the residual for a iteration of Newton's method for mixed formulation; set assembleMatrix to false to only assemble the residual; ATTENTION: rhs() returns a negative residual (-r) !!!

Reimplemented in gsMuscleAssembler< T >, and gsThermoAssembler< T >.

◆ computeDirichletDofs()

template<class T >

void computeDirichletDofs ( short_t unk = 0 )

inherited

Triggers computation of the Dirichlet dofs.

Parameters

[in] unk the considered unknown

◆ computeDirichletDofsIntpl()

template<class T >

void computeDirichletDofsIntpl	(	const gsDofMapper &	mapper,
		const gsMultiBasis< T > &	mbasis,
		const short_t	unk_ = `0`
	)

protectedinherited

calculates the values of the eliminated dofs based on Interpolation.

Parameters

[in]	mapper	the dofMapper for the considered unknown
[in]	mbasis	the multipabasis for the considered unknown
[in]	unk_	the considered unknown

◆ computeDirichletDofsL2Proj()

template<class T >

void computeDirichletDofsL2Proj	(	const gsDofMapper &	mapper,
		const gsMultiBasis< T > &	mbasis,
		const short_t	unk_ = `0`
	)

protectedinherited

calculates the values of the eliminated dofs based on L2 Projection.

Parameters

[in]	mapper	the dofMapper for the considered unknown
[in]	mbasis	the multipabasis for the considered unknown
[in]	unk_	the considered unknown

◆ constructSolution() [1/2]

template<class T >

virtual void constructSolution	(	const gsMatrix< T > &	solVector,
		gsMultiPatch< T > &	result,
		const gsVector< index_t > &	unknowns
	)		const

inlinevirtual

Construct solution from computed solution vector for a set of unknowns. The result is a vectorfield, where each component is given the corresponding entry of.

unknowns. This method assumes that the specified unknowns have the: same basis.

Parameters

[in]	solVector	the solution vector obtained from the linear system
[out]	result	the solution seen as vectorfield in form of a gsMultiBasis
[in]	unknowns	the considered vector of unknowns

Reimplemented from gsBaseAssembler< T >.

◆ constructSolution() [2/2]

template<class T >

virtual void constructSolution	(	const gsMatrix< T > &	solVector,
		gsMultiPatch< T > &	result,
		short_t	unk = `0`
	)		const

inlinevirtual

Construct solution from computed solution vector for a single unknows.

Parameters

[in]	solVector	the solution vector obtained from the linear system
[out]	result	the solution in form of a gsMultiBasis
[in]	unk	the considered unknown

Reimplemented from gsBaseAssembler< T >.

◆ fixedDofs()

template<class T >

const gsMatrix< T > & fixedDofs ( short_t unk = 0 ) const

inlineinherited

Returns the Dirichlet values for a unknown (if applicable)

Parameters

[in] unk the considered unknown

◆ getFixedDofs()

template<class T >

void getFixedDofs	(	index_t	patch,
		boxSide	side,
		gsMatrix< T > &	ddofs
	)		const

virtualinherited

get fixed degrees of freedom corresponding to a given part of the bdry. each column of the resulting matrix correspond to one variable/component of the vector-valued vairable

◆ homogenizeFixedDofs()

template<class T >

void homogenizeFixedDofs ( short_t unk = 0 )

inlineinherited

Sets any Dirichlet values to homogeneous (if applicable)

Parameters

[in] unk the considered unknown

◆ penalizeDirichletDofs()

template<class T >

void penalizeDirichletDofs ( short_t unk = 0 )

inherited

Enforce Dirichlet boundary conditions by diagonal penalization

Parameters

[in] unk the considered unknown

◆ setFixedDofs() [1/2]

template<class T >

void setFixedDofs	(	const gsMatrix< T > &	coefMatrix,
		short_t	unk = `0`,
		size_t	patch = `0`
	)

inherited

the user can manually set the dirichlet Dofs for a given patch and unknown, based on the Basis coefficients

Parameters

[in]	coefMatrix	the coefficients of the function
[in]	unk	the consideren unknown
[in]	patch	the patch index

◆ setFixedDofs() [2/2]

template<class T >

void setFixedDofs	(	index_t	patch,
		boxSide	side,
		const gsMatrix< T > &	ddofs,
		bool	oneUnk = `false`
	)

virtualinherited

Set Dirichet degrees of freedom on a given side of a given patch from a given matrix.

A usual source of degrees of freedom is another geometry where it is known that the corresponding bases match. The function is not safe in that it assumes a correct numbering of DoFs in a given matrix. To allocate space for these DoFs in the assembler, add an empty/zero Dirichlet boundary condition to gsBoundaryCondtions container that is passed to the assembler constructor.

Reimplemented in gsNsTimeIntegrator< T >.

◆ setFixedDofVector()

template<class T >

void setFixedDofVector	(	gsMatrix< T >	vals,
		short_t	unk = `0`
	)

inherited

the user can manually set the dirichlet Dofs for a given patch and unknown.

Parameters

[in]	vals	the values of the eliminated dofs.
[in]	unk	the considered unknown

◆ updateSolution()

template<class T >

void updateSolution	(	const gsMatrix< T > &	solVector,
		gsMultiPatch< T > &	result,
		T	theta = `(T)(1)`
	)		const

virtualinherited

Update solution by adding the computed solution vector to the current solution specified by.

result. This method assumes that all: unknowns have the same basis.

Parameters

[in]	solVector	the solution vector obtained from the linear system
[out]	result	the solution in form of a gsMultiBasis,

solVector is added to the: coefficients of result.

Parameters

[in] theta damping factor for update, theta = 1 corresponds to a full step.

Member Data Documentation

◆ m_bases

template<class T >

std::vector< gsMultiBasis<T> > m_bases

protectedinherited

The discretization bases corresponding to the patches and to the number of solution fields that are to be computed. One entry of the vector corresponds to one or more unknown (in case of a system of PDEs)

◆ m_ddof

template<class T >

std::vector<gsMatrix<T> > m_ddof

protectedinherited

Fixed DoF values (if applicable, for instance eliminated Dirichlet DoFs). One entry of the vector corresponds to exactly one unknown, i.e. the size of m_ddof must fit m_system.colBlocks().

◆ m_dim

template<class T >

short_t m_dim

protected

Dimension of the problem parametric dim = physical dim = deformation dim

◆ m_pde_ptr

template<class T >

memory::shared_ptr<gsPde<T> > m_pde_ptr

protectedinherited

The PDE: contains multi-patch domain, boundary conditions and coeffcient functions

Detailed Description

Public Member Functions

Static Public Member Functions

Protected Member Functions

Protected Attributes

Member Function Documentation

◆ apply()

◆ assemble() [1/4]

◆ assemble() [2/4]

◆ assemble() [3/4]

◆ assemble() [4/4]

◆ computeDirichletDofs()

◆ computeDirichletDofsIntpl()

◆ computeDirichletDofsL2Proj()

◆ constructSolution() [1/2]

◆ constructSolution() [2/2]

◆ fixedDofs()

◆ getFixedDofs()

◆ homogenizeFixedDofs()

◆ penalizeDirichletDofs()

◆ setFixedDofs() [1/2]

◆ setFixedDofs() [2/2]

◆ setFixedDofVector()

◆ updateSolution()

Member Data Documentation

◆ m_bases

◆ m_ddof

◆ m_dim

◆ m_pde_ptr