#include <nonlinear_implicit_assembly.h>

Inheritance diagram for MAST::NonlinearImplicitAssembly:

Collaboration diagram for MAST::NonlinearImplicitAssembly:

[legend]

Classes
class	PostAssemblyOperation
	user-provided object to perform actions after assembly and before returning to the solver. More...

Public Member Functions
	NonlinearImplicitAssembly ()
	constructor associates this assembly object with the system More...

virtual	~NonlinearImplicitAssembly ()
	destructor resets the association of this assembly object with the system More...

Real	res_l2_norm () const
	L2 norm of the last-assembled residual. More...

Real	first_iter_res_l2_norm () const

void	reset_residual_norm_history ()
	reset L2 norm of the last-assembled residual More...

void	set_post_assembly_operation (MAST::NonlinearImplicitAssembly::PostAssemblyOperation &post)
	sets the PostAssemblyOperation object for use after assembly. More...

virtual void	residual_and_jacobian (const libMesh::NumericVector< Real > &X, libMesh::NumericVector< Real > R, libMesh::SparseMatrix< Real > J, libMesh::NonlinearImplicitSystem &S)
	function that assembles the matrices and vectors quantities for nonlinear solution More...

virtual void	linearized_jacobian_solution_product (const libMesh::NumericVector< Real > &X, const libMesh::NumericVector< Real > &dX, libMesh::NumericVector< Real > &JdX, libMesh::NonlinearImplicitSystem &S)
	calculates the product of the Jacobian and a perturbation in solution vector $[J] \{\Delta X\}$ . More...

virtual void	second_derivative_dot_solution_assembly (const libMesh::NumericVector< Real > &X, const libMesh::NumericVector< Real > &dX, libMesh::SparseMatrix< Real > &d_JdX_dX, libMesh::NonlinearImplicitSystem &S)
	calculates $d ([J] \{\Delta X\})/ dX$ . More...

virtual bool	sensitivity_assemble (const MAST::FunctionBase &f, libMesh::NumericVector< Real > &sensitivity_rhs)
	Assembly function. More...

Public Member Functions inherited from MAST::AssemblyBase
	AssemblyBase ()
	constructor takes a reference to the discipline that provides the boundary conditions, volume loads, properties, etc. More...

virtual	~AssemblyBase ()
	virtual destructor More...

const MAST::PhysicsDisciplineBase &	discipline () const

MAST::PhysicsDisciplineBase &	discipline ()

MAST::AssemblyElemOperations &	get_elem_ops ()

virtual void	set_discipline_and_system (MAST::PhysicsDisciplineBase &discipline, MAST::SystemInitialization &system)
	attaches a system to this discipline More...

void	attach_elem_parameter_dependence_object (MAST::AssemblyBase::ElemParameterDependence &dep)
	This object, if provided by user, will be used to reduce unnecessary computations in sensitivity analysis assembly operations. More...

void	clear_elem_parameter_dependence_object ()

virtual void	clear_discipline_and_system ()
	clears association with a system to this discipline More...

virtual void	set_elem_operation_object (MAST::AssemblyElemOperations &elem_ops)
	attaches a element operation to this object, and associated `this` with the element operation object. More...

virtual void	clear_elem_operation_object ()
	clears the association of `this` object with the assembly element operation object. More...

const MAST::NonlinearSystem &	system () const

MAST::NonlinearSystem &	system ()

MAST::SystemInitialization &	system_init ()

void	set_solver_monitor (MAST::AssemblyBase::SolverMonitor &monitor)
	attaches the solver monitor, which is a user provided routine that is called each time More...

MAST::AssemblyBase::SolverMonitor *	get_solver_monitor ()

void	clear_solver_monitor ()
	clears the monitor object More...

void	attach_solution_function (MAST::MeshFieldFunction &f)
	tells the assembly object that this function is will need to be initialized before each residual evaluation More...

void	detach_solution_function ()
	removes the attachment of the solution function More...

virtual void	calculate_output (const libMesh::NumericVector< Real > &X, MAST::OutputAssemblyElemOperations &output)
	calculates the value of quantity . More...

virtual void	calculate_output_derivative (const libMesh::NumericVector< Real > &X, MAST::OutputAssemblyElemOperations &output, libMesh::NumericVector< Real > &dq_dX)
	calculates $\frac{\partial q(X, p)}{\partial X}$ More...

virtual void	calculate_output_direct_sensitivity (const libMesh::NumericVector< Real > &X, const libMesh::NumericVector< Real > *dXdp, const MAST::FunctionBase &p, MAST::OutputAssemblyElemOperations &output)
	evaluates the sensitivity of the outputs in the attached discipline with respect to the parametrs in `params`. More...

virtual Real	calculate_output_adjoint_sensitivity (const libMesh::NumericVector< Real > &X, const libMesh::NumericVector< Real > &dq_dX, const MAST::FunctionBase &p, MAST::AssemblyElemOperations &elem_ops, MAST::OutputAssemblyElemOperations &output, const bool include_partial_sens=true)
	Evaluates the total sensitivity of `output` wrt `p` using the adjoint solution provided in `dq_dX` for a linearization about solution `X`. More...

std::unique_ptr< libMesh::NumericVector< Real > >	build_localized_vector (const libMesh::System &sys, const libMesh::NumericVector< Real > &global) const
	localizes the parallel vector so that the local copy stores all values necessary for calculation of the element quantities More...

Protected Attributes
MAST::NonlinearImplicitAssembly::PostAssemblyOperation *	_post_assembly
	this object, if non-NULL is user-provided to perform actions after assembly and before returning to the solver More...

Real	_res_l2_norm
	L2 norm of the last-assembled residual. More...

Real	_first_iter_res_l2_norm

Protected Attributes inherited from MAST::AssemblyBase
MAST::AssemblyElemOperations *	_elem_ops
	provides assembly elem operations for use by this class More...

MAST::PhysicsDisciplineBase *	_discipline
	PhysicsDisciplineBase object for which this class is assembling. More...

MAST::SystemInitialization *	_system
	System for which this assembly is performed. More...

MAST::MeshFieldFunction *	_sol_function
	system solution that will be initialized before each solution More...

MAST::AssemblyBase::SolverMonitor *	_solver_monitor
	User provided solver monitor is attached to the linear nonlinear solvers, if provided. More...

MAST::AssemblyBase::ElemParameterDependence *	_param_dependence
	If provided by user, this object is used by sensitiivty analysis to check for whether or the current design parameter influences an element. More...

Additional Inherited Members
Public Attributes inherited from MAST::AssemblyBase
bool	close_matrix
	flag to control the closing fo the Jacobian after assembly More...

Detailed Description

Definition at line 36 of file nonlinear_implicit_assembly.h.

Constructor & Destructor Documentation

MAST::NonlinearImplicitAssembly::NonlinearImplicitAssembly ( )

constructor associates this assembly object with the system

Definition at line 40 of file nonlinear_implicit_assembly.cpp.

Here is the caller graph for this function:

MAST::NonlinearImplicitAssembly::~NonlinearImplicitAssembly ( )

virtual

destructor resets the association of this assembly object with the system

Definition at line 49 of file nonlinear_implicit_assembly.cpp.

Here is the call graph for this function:

Here is the caller graph for this function:

Member Function Documentation

Real MAST::NonlinearImplicitAssembly::first_iter_res_l2_norm ( ) const

inline

Definition at line 76 of file nonlinear_implicit_assembly.h.

void MAST::NonlinearImplicitAssembly::linearized_jacobian_solution_product	(	const libMesh::NumericVector< Real > &	X,
		const libMesh::NumericVector< Real > &	dX,
		libMesh::NumericVector< Real > &	JdX,
		libMesh::NonlinearImplicitSystem &	S
	)

virtual

calculates the product of the Jacobian and a perturbation in solution vector $[J] \{\Delta X\}$ .

For a single discipline system the solution vector and linearized solution provided here are used. For a multiphysics system, the user must ensure that all relevant multidisciplinary data-structures are initialized before calling this method.

Definition at line 256 of file nonlinear_implicit_assembly.cpp.

Here is the call graph for this function:

Here is the caller graph for this function:

Real MAST::NonlinearImplicitAssembly::res_l2_norm ( ) const

inline

L2 norm of the last-assembled residual.

Definition at line 74 of file nonlinear_implicit_assembly.h.

void MAST::NonlinearImplicitAssembly::reset_residual_norm_history ( )

inline

reset L2 norm of the last-assembled residual

Definition at line 81 of file nonlinear_implicit_assembly.h.

Here is the call graph for this function:

void MAST::NonlinearImplicitAssembly::residual_and_jacobian	(	const libMesh::NumericVector< Real > &	X,
		libMesh::NumericVector< Real > *	R,
		libMesh::SparseMatrix< Real > *	J,
		libMesh::NonlinearImplicitSystem &	S
	)

virtual

function that assembles the matrices and vectors quantities for nonlinear solution

Reimplemented from MAST::AssemblyBase.

Reimplemented in MAST::LevelSetNonlinearImplicitAssembly.

Definition at line 66 of file nonlinear_implicit_assembly.cpp.

Here is the call graph for this function:

Here is the caller graph for this function:

void MAST::NonlinearImplicitAssembly::second_derivative_dot_solution_assembly	(	const libMesh::NumericVector< Real > &	X,
		const libMesh::NumericVector< Real > &	dX,
		libMesh::SparseMatrix< Real > &	d_JdX_dX,
		libMesh::NonlinearImplicitSystem &	S
	)

virtual

calculates $d ([J] \{\Delta X\})/ dX$ .

Definition at line 369 of file nonlinear_implicit_assembly.cpp.

Here is the call graph for this function:

Here is the caller graph for this function:

bool MAST::NonlinearImplicitAssembly::sensitivity_assemble	(	const MAST::FunctionBase &	f,
		libMesh::NumericVector< Real > &	sensitivity_rhs
	)

virtual

Assembly function.

This function will be called to assemble the RHS of the sensitivity equations (which is -1 times sensitivity of system residual) prior to a solve and must be provided by the user in a derived class. The method provides dR/dp for f parameter.

If the routine is not able to provide sensitivity for this parameter, then it should return false, and the system will attempt to use finite differencing.

Reimplemented from MAST::AssemblyBase.

Reimplemented in MAST::LevelSetNonlinearImplicitAssembly.

Definition at line 480 of file nonlinear_implicit_assembly.cpp.

Here is the call graph for this function:

Here is the caller graph for this function:

void MAST::NonlinearImplicitAssembly::set_post_assembly_operation ( MAST::NonlinearImplicitAssembly::PostAssemblyOperation & post )

sets the PostAssemblyOperation object for use after assembly.

Note that calling clear_discipline_and_system() will clear this pointer and the user will have to call this function again.

Definition at line 57 of file nonlinear_implicit_assembly.cpp.

Here is the call graph for this function:

Here is the caller graph for this function:

Member Data Documentation

Real MAST::NonlinearImplicitAssembly::_first_iter_res_l2_norm

protected

Definition at line 159 of file nonlinear_implicit_assembly.h.

MAST::NonlinearImplicitAssembly::PostAssemblyOperation* MAST::NonlinearImplicitAssembly::_post_assembly

protected

this object, if non-NULL is user-provided to perform actions after assembly and before returning to the solver

Definition at line 154 of file nonlinear_implicit_assembly.h.

Real MAST::NonlinearImplicitAssembly::_res_l2_norm

protected

L2 norm of the last-assembled residual.

Definition at line 159 of file nonlinear_implicit_assembly.h.

The documentation for this class was generated from the following files:

/home/travis/build/MASTmultiphysics/mast-multiphysics/src/base/nonlinear_implicit_assembly.h
/home/travis/build/MASTmultiphysics/mast-multiphysics/src/base/nonlinear_implicit_assembly.cpp

Classes

Public Member Functions

Protected Attributes

Additional Inherited Members

Detailed Description

Constructor & Destructor Documentation

Member Function Documentation

Member Data Documentation