This class stores all ContactInteractions for a problem, calls Tribol functions that act on all contact interactions, and agglomerates fields that exist over different ContactInteractions. More...

#include <contact_data.hpp>

Public Member Functions
	ContactData (const mfem::ParMesh &mesh)
	The constructor. More...

	~ContactData ()
	Destructor to finalize Tribol.

void	addContactInteraction (int interaction_id, const std::set< int > &bdry_attr_surf1, const std::set< int > &bdry_attr_surf2, ContactOptions contact_opts)
	Add another contact interaction. More...

void	updateGeometry (int cycle, double time, double &dt, std::optional< std::reference_wrapper< const mfem::Vector >> u_shape=std::nullopt, std::optional< std::reference_wrapper< const mfem::Vector >> u=std::nullopt, bool eval_jacobian=false)
	Updates the contact geometry and gap data. More...

void	updateForcesAndJacobian (int cycle, double time, double &dt, std::optional< std::reference_wrapper< const mfem::Vector >> u_shape=std::nullopt, std::optional< std::reference_wrapper< const mfem::Vector >> u=std::nullopt, std::optional< std::reference_wrapper< const mfem::Vector >> p=std::nullopt)
	Updates the contact forces and Jacobian contributions for the current configuration. More...

void	reset ()
	Resets the contact pressures to zero.

FiniteElementDual	forces () const
	Get the contact constraint residual (i.e. nodal forces) from all contact interactions. More...

mfem::HypreParVector	mergedPressures () const
	Returns pressures from all contact interactions on the contact surface true degrees of freedom. More...

mfem::HypreParVector	mergedGaps (bool zero_inactive=false) const
	Returns nodal gaps from all contact interactions on the contact surface true degrees of freedom. More...

std::unique_ptr< mfem::BlockOperator >	mergedJacobian () const
	Returns a 2x2 block Jacobian on displacement/pressure true degrees of freedom from contact constraints. More...

void	residualFunction (const mfem::Vector &u_shape, const mfem::Vector &u, mfem::Vector &r)
	Computes the residual including contact terms. More...

std::unique_ptr< mfem::BlockOperator >	jacobianFunction (std::unique_ptr< mfem::HypreParMatrix > orig_J) const
	Computes the Jacobian including contact terms, given the non-contact Jacobian terms. More...

std::unique_ptr< mfem::HypreParMatrix >	contactSubspaceTransferOperator ()
	Computes the subspace transfer operator. More...

bool	haveContactInteractions () const
	Have there been contact interactions added? More...

bool	haveLagrangeMultipliers () const
	Are any contact interactions enforced using Lagrange multipliers? More...

int	numPressureDofs () const
	Get the number of Lagrange multiplier true degrees of freedom. More...

Protected Member Functions
void	setPressures (const mfem::Vector &merged_pressures) const
	Set the pressure field. More...

void	setDisplacements (const mfem::Vector &u_shape, const mfem::Vector &u)
	Update the current coordinates based on the new displacement field. More...

Detailed Description

This class stores all ContactInteractions for a problem, calls Tribol functions that act on all contact interactions, and agglomerates fields that exist over different ContactInteractions.

Definition at line 49 of file contact_data.hpp.

Constructor & Destructor Documentation

◆ ContactData()

smith::ContactData::ContactData ( const mfem::ParMesh & mesh )

The constructor.

Parameters

mesh	The volume mesh for the problem

Definition at line 459 of file contact_data.cpp.

Member Function Documentation

◆ addContactInteraction()

void smith::ContactData::addContactInteraction	(	int	interaction_id,
		const std::set< int > &	bdry_attr_surf1,
		const std::set< int > &	bdry_attr_surf2,
		ContactOptions	contact_opts
	)

Add another contact interaction.

Parameters

interaction_id	Unique identifier for the ContactInteraction (used in Tribol)
bdry_attr_surf1	MFEM boundary attributes for the first (mortar) surface
bdry_attr_surf2	MFEM boundary attributes for the second (nonmortar) surface
contact_opts	Defines contact method, enforcement, type, and penalty

Note: Interaction ids must be unique across the entire executable: Tribol uses interaction_id as the unique key for its coupling schemes, so reusing an id (even across different ContactData instances) is an error.

Definition at line 466 of file contact_data.cpp.

◆ contactSubspaceTransferOperator()

std::unique_ptr< mfem::HypreParMatrix > smith::ContactData::contactSubspaceTransferOperator ( )

Computes the subspace transfer operator.

Returns: Contact subspace transfer operator (mapping from contact_dofs to all displacement dofs)

Definition at line 535 of file contact_data.cpp.

◆ forces()

FiniteElementDual smith::ContactData::forces ( ) const

Get the contact constraint residual (i.e. nodal forces) from all contact interactions.

Returns: Nodal contact forces on the true DOFs

Precondition: updateForcesAndJacobian() must be called with the current configuration so the force contributions are up-to-date

Definition at line 489 of file contact_data.cpp.

◆ haveContactInteractions()

bool smith::ContactData::haveContactInteractions ( ) const

inline

Have there been contact interactions added?

Returns: true if contact interactions have been added; false if there are no contact interactions

Definition at line 203 of file contact_data.hpp.

◆ haveLagrangeMultipliers()

bool smith::ContactData::haveLagrangeMultipliers ( ) const

inline

Are any contact interactions enforced using Lagrange multipliers?

Returns: true: at least one contact interaction is using Lagrange multiplier enforcement; false: no contact interactions are using Lagrange multipliers

Definition at line 227 of file contact_data.hpp.

◆ jacobianFunction()

std::unique_ptr< mfem::BlockOperator > smith::ContactData::jacobianFunction ( std::unique_ptr< mfem::HypreParMatrix > orig_J ) const

Computes the Jacobian including contact terms, given the non-contact Jacobian terms.

Parameters

orig_J The non-contact terms of the Jacobian, not including essential boundary conditions

Returns: Jacobian with contact terms, not including essential boundary conditions

Precondition: updateForcesAndJacobian() must be called with the current configuration so the Jacobian contributions are up-to-date

Definition at line 515 of file contact_data.cpp.

◆ mergedGaps()

mfem::HypreParVector smith::ContactData::mergedGaps ( bool zero_inactive = false ) const

Returns nodal gaps from all contact interactions on the contact surface true degrees of freedom.

The type of gap (normal or vector-valued) is set by the ContactType in the ContactOptions struct for the contact interaction. TiedNormal and Frictionless (the two type supported in Tribol) correspond to scalar gap normal. Only linear (order = 1) gap fields are supported.

Parameters

[in] zero_inactive Sets inactive t-dofs to zero gap

Returns: Nodal gap true degrees of freedom on each contact interaction (merged into one mfem::HypreParVector)

Precondition: updateGeometry() or updateForcesAndJacobian() must be called with the current configuration so the gap values are up-to-date

Definition at line 498 of file contact_data.cpp.

◆ mergedJacobian()

std::unique_ptr< mfem::BlockOperator > smith::ContactData::mergedJacobian ( ) const

Returns a 2x2 block Jacobian on displacement/pressure true degrees of freedom from contact constraints.

The element Jacobian contributions are computed upon calling updateForcesAndJacobian(). This method does MPI communication to move Jacobian contributions to the correct rank, then assembles the contributions. The pressure degrees of freedom for all contact interactions are merged into a single block.

Note: The blocks are owned by the BlockOperator

Returns: Pointer to block Jacobian (2x2 BlockOperator of HypreParMatrix)

Definition at line 503 of file contact_data.cpp.

◆ mergedPressures()

mfem::HypreParVector smith::ContactData::mergedPressures ( ) const

Returns pressures from all contact interactions on the contact surface true degrees of freedom.

The type of pressure (normal or vector-valued) is set by the ContactType in the ContactOptions struct for the contact interaction. TiedNormal and Frictionless (the two type supported in Tribol) correspond to scalar normal pressure. Only linear (order = 1) pressure fields are supported.

Returns: Pressure true degrees of freedom on each contact interaction (merged into one mfem::HypreParVector)

Definition at line 496 of file contact_data.cpp.

◆ numPressureDofs()

int smith::ContactData::numPressureDofs ( ) const

inline

Get the number of Lagrange multiplier true degrees of freedom.

Returns: Number of Lagrange multiplier true degrees of freedom

Definition at line 234 of file contact_data.hpp.

◆ residualFunction()

void smith::ContactData::residualFunction	(	const mfem::Vector &	u_shape,
		const mfem::Vector &	u,
		mfem::Vector &	r
	)

Computes the residual including contact terms.

Parameters

[in]	u_shape	Shape displacement vector (size of [displacement] block)
[in]	u	Solution vector ([displacement; pressure] block vector)
[in,out]	r	Residual vector ([force; gap] block vector); takes in initialized residual force vector and adds contact contributions

Precondition: The current coordinates must be up-to-date

Note: This method calls updateForcesAndJacobian() to compute residual and Jacobian contributions based on the current configuration

Definition at line 510 of file contact_data.cpp.

◆ setDisplacements()

void smith::ContactData::setDisplacements	(	const mfem::Vector &	u_shape,
		const mfem::Vector &	u
	)

protected

Update the current coordinates based on the new displacement field.

Parameters

u_shape	Shape displacement vector
u	Current displacement dof values

Definition at line 530 of file contact_data.cpp.

◆ setPressures()

void smith::ContactData::setPressures ( const mfem::Vector & merged_pressures ) const

protected

Set the pressure field.

This sets Tribol's pressure degrees of freedom based on 1) the values in merged_pressure for Lagrange multiplier enforcement 2) the nodal gaps and penalty for penalty enforcement

Note: The nodal gaps must be up-to-date for penalty enforcement

Parameters

merged_pressures Current pressure true dof values in a merged mfem::Vector

Definition at line 528 of file contact_data.cpp.

◆ updateForcesAndJacobian()

void smith::ContactData::updateForcesAndJacobian	(	int	cycle,
		double	time,
		double &	dt,
		std::optional< std::reference_wrapper< const mfem::Vector >>	u_shape = `std::nullopt`,
		std::optional< std::reference_wrapper< const mfem::Vector >>	u = `std::nullopt`,
		std::optional< std::reference_wrapper< const mfem::Vector >>	p = `std::nullopt`
	)

Updates the contact forces and Jacobian contributions for the current configuration.

Parameters

cycle	The current simulation cycle
time	The current time
dt	The timestep size to attempt
u_shape	Optional shape displacement vector; if provided with `u`, the geometry is refreshed first
u	Optional current displacement dof values; if provided with `u_shape`, the geometry is refreshed first
p	Optional current pressure true dof values. Required for Lagrange multiplier enforcement and ignored for penalty enforcement.

Definition at line 481 of file contact_data.cpp.

◆ updateGeometry()

void smith::ContactData::updateGeometry	(	int	cycle,
		double	time,
		double &	dt,
		std::optional< std::reference_wrapper< const mfem::Vector >>	u_shape = `std::nullopt`,
		std::optional< std::reference_wrapper< const mfem::Vector >>	u = `std::nullopt`,
		bool	eval_jacobian = `false`
	)

Updates the contact geometry and gap data.

Parameters

cycle	The current simulation cycle
time	The current time
dt	The timestep size to attempt
u_shape	Optional shape displacement vector
u	Optional current displacement dof values
eval_jacobian	Whether to also evaluate the Jacobian contributions (default false)

Definition at line 474 of file contact_data.cpp.

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

src/smith/physics/contact/contact_data.hpp
src/smith/physics/contact/contact_data.cpp

Public Member Functions

Protected Member Functions

Detailed Description

Constructor & Destructor Documentation

◆ ContactData()

Member Function Documentation

◆ addContactInteraction()

◆ contactSubspaceTransferOperator()

◆ forces()

◆ haveContactInteractions()

◆ haveLagrangeMultipliers()

◆ jacobianFunction()

◆ mergedGaps()

◆ mergedJacobian()

◆ mergedPressures()

◆ numPressureDofs()

◆ residualFunction()

◆ setDisplacements()

◆ setPressures()

◆ updateForcesAndJacobian()

◆ updateGeometry()