The driver for a transient initial-value problem. More...

Detailed Description

The driver for a transient initial-value problem.

#include <TransientDriver.h>

Inheritance diagram for TransientDriver:

[legend]

Public Member Functions
	TransientDriver (const OptionSet &options)

void	setup () override
	Setup this object.

void	diagnose () const override
	Check for common problems.

bool	run () override
	Let the driver run, return `true` upon successful completion, and return `false` otherwise.

Public Member Functions inherited from ModelDriver
	ModelDriver (const OptionSet &options)

const Model &	model () const

const Model &	postprocessor () const

Public Member Functions inherited from Driver
	Driver (const OptionSet &options)
	Construct a new Driver object.

Public Member Functions inherited from NEML2Object
	NEML2Object ()=delete

	NEML2Object (NEML2Object &&)=delete

	NEML2Object (const NEML2Object &)=delete

NEML2Object &	operator= (NEML2Object &&)=delete

NEML2Object &	operator= (const NEML2Object &)=delete

virtual	~NEML2Object ()=default

	NEML2Object (const OptionSet &options)
	Construct a new NEML2Object object.

const OptionSet &	input_options () const

const std::string &	name () const
	A readonly reference to the object's name.

const std::string &	type () const
	A readonly reference to the object's type.

const std::string &	path () const
	A readonly reference to the object's path.

const std::string &	doc () const
	A readonly reference to the object's docstring.

Factory *	factory () const
	Get the factory that created this object.

const Settings &	settings () const
	Settings.

template<typename T = NEML2Object>
const T *	host () const
	Get a readonly pointer to the host.

template<typename T = NEML2Object>
T *	host ()
	Get a writable pointer to the host.

template<typename T >
const T &	resolve_tensor (const std::string &name)
	Resolve a TensorName to a Tensor.

template<class T >
std::shared_ptr< T >	get_object (const std::string &section, const std::string &name)

template<class T = Solver>
std::shared_ptr< T >	get_solver (const std::string &name)
	Get a solver from the factory.

template<class T = Data>
std::shared_ptr< T >	get_data (const std::string &name)
	Get a data from the factory.

template<class T = Model>
std::shared_ptr< T >	get_model (const std::string &name)
	Get a model from the factory.

template<class T = Driver>
std::shared_ptr< T >	get_driver (const std::string &name)
	Get a driver from the factory.

template<class T = WorkScheduler>
std::shared_ptr< T >	get_scheduler (const std::string &name)
	Get a scheduler from the factory.

Public Member Functions inherited from DiagnosticsInterface
	DiagnosticsInterface ()=delete

	DiagnosticsInterface (NEML2Object *object)

	DiagnosticsInterface (DiagnosticsInterface &&)=delete

	DiagnosticsInterface (const DiagnosticsInterface &)=delete

DiagnosticsInterface &	operator= (const DiagnosticsInterface &)=delete

DiagnosticsInterface &	operator= (DiagnosticsInterface &&)=delete

virtual	~DiagnosticsInterface ()=default

const NEML2Object &	object () const
	Get the object.

Static Public Member Functions
static OptionSet	expected_options ()

Static Public Member Functions inherited from ModelDriver
static OptionSet	expected_options ()

Static Public Member Functions inherited from Driver
static OptionSet	expected_options ()

Static Public Member Functions inherited from NEML2Object
static OptionSet	expected_options ()

Protected Member Functions
virtual bool	solve ()
	Solve the initial value problem.

virtual void	advance_step ()
	Advance in time: the state becomes old state, and forces become old forces.

virtual void	update_forces ()
	Update the driving forces for the current time step.

virtual void	apply_ic ()
	Apply the initial conditions.

virtual void	apply_predictor ()
	Apply the predictor to calculate the initial guess for the current time step.

virtual void	solve_step ()
	Perform the constitutive update for the current time step.

virtual void	postprocess ()
	Postprocess the output of the current time step.

virtual void	store_input ()
	Save the input of the current time step.

Protected Attributes
const VariableName	_time_name
	VariableName for the time.

Scalar	_time
	The current time.

Size	_step_count = 0
	The current step count.

const Size	_nsteps
	Total number of steps.

ValueMap	_in
	The input to the constitutive model.

const EnumSelection	_predictor
	The predictor used to set the initial guess.

std::vector< ValueMap >	_result_in
	Inputs from all time steps.

std::vector< ValueMap >	_result_out
	Outputs from all time steps.

std::vector< VariableName >	_driving_force_names
	Driving forces (other than time)

std::vector< Tensor >	_driving_forces

Protected Attributes inherited from ModelDriver
const std::shared_ptr< Model >	_model
	The model which the driver uses to perform constitutive updates.

const std::vector< VariableName >	_intmd_vars
	Variable names for which to tag intermediate shapes.

const std::vector< TensorShape >	_intmd_shapes
	Corresponding tensor shapes for the intermediate variables.

const std::shared_ptr< Model >	_postprocessor
	Postprocessor model.

const Device	_device
	The device on which to evaluate the model.

const bool	_show_model_info
	Set to true to list all the model parameters at the beginning.

Protected Attributes inherited from Driver
bool	_verbose
	Whether to print out additional (debugging) information during the execution.

std::string	_save_as
	The destination file name or file path.

virtual torch::nn::ModuleDict	result () const
	The results (input and output) from all time steps.

virtual std::string	save_as_path () const
	The destination file/path to save the results.

virtual void	output () const
	Save the results into the destination file/path.

Constructor & Destructor Documentation

◆ TransientDriver()

TransientDriver ( const OptionSet & options )

Member Function Documentation

◆ advance_step()

void advance_step ( )

protectedvirtual

Advance in time: the state becomes old state, and forces become old forces.

◆ apply_ic()

void apply_ic ( )

protectedvirtual

Apply the initial conditions.

◆ apply_predictor()

void apply_predictor ( )

protectedvirtual

Apply the predictor to calculate the initial guess for the current time step.

Reimplemented in LDISolidMechanicsDriver.

◆ diagnose()

void diagnose ( ) const

overridevirtual

Check for common problems.

This method serves as the entry point for diagnosing common problems in object setup. The idea behind this method is that while some errors could be detected at construction time, i.e., when the object's constructor is called, it doesn't hinder other objects' creation. We therefore would like to defer the detection of errors until after all objects have been created, collect all errors at once, and present the user with a complete understanding of all errors encountered.

Note, however, if an error could interfere with other objects' creation, it should be raised right away inside the constructor, instead of inside this method.

Reimplemented from ModelDriver.

◆ expected_options()

OptionSet expected_options ( )

static

◆ output()

void output ( ) const

protectedvirtual

Save the results into the destination file/path.

◆ postprocess()

void postprocess ( )

protectedvirtual

Postprocess the output of the current time step.

◆ result()

torch::nn::ModuleDict result ( ) const

virtual

The results (input and output) from all time steps.

Outputting

Returns: torch::nn::ModuleDict The results (input and output) from all time steps. The keys of the dict are "input" and "output". The values are torch::nn::ModuleList with length equal to the number of time steps. Each ModuleList contains the input/output variables at each time step in the form of torch::nn::ModuleDict whose keys are variable names and values are variable values.

◆ run()

bool run ( )

overridevirtual

Let the driver run, return true upon successful completion, and return false otherwise.

Implements Driver.

◆ save_as_path()

std::string save_as_path ( ) const

virtual

The destination file/path to save the results.

◆ setup()

void setup ( )

overridevirtual

Setup this object.

This method is called automatically if you use the Factory method get_object or get_object, right after construction. This serves as the entry point for things that are not convenient/possible to do at construction time, but are necessary before this object can be used (by others).

Reimplemented from ModelDriver.

◆ solve()

bool solve ( )

protectedvirtual

Solve the initial value problem.

◆ solve_step()

void solve_step ( )

protectedvirtual

Perform the constitutive update for the current time step.

◆ store_input()

void store_input ( )

protectedvirtual

Save the input of the current time step.

◆ update_forces()

void update_forces ( )

protectedvirtual

Update the driving forces for the current time step.

Reimplemented in LDISolidMechanicsDriver, and SolidMechanicsDriver.

Member Data Documentation

◆ _driving_force_names

std::vector<VariableName> _driving_force_names

protected

Driving forces (other than time)

◆ _driving_forces

std::vector<Tensor> _driving_forces

protected

◆ _in

ValueMap _in

protected

The input to the constitutive model.

◆ _nsteps

const Size _nsteps

protected

Total number of steps.

◆ _predictor

const EnumSelection _predictor

protected

The predictor used to set the initial guess.

◆ _result_in

std::vector<ValueMap> _result_in

protected

Inputs from all time steps.

◆ _result_out

std::vector<ValueMap> _result_out

protected

Outputs from all time steps.

◆ _save_as

std::string _save_as

protected

The destination file name or file path.

◆ _step_count

Size _step_count = 0

protected

The current step count.

◆ _time

Scalar _time

protected

The current time.

◆ _time_name

const VariableName _time_name

protected

VariableName for the time.

Detailed Description

Public Member Functions

Static Public Member Functions

Protected Member Functions

Protected Attributes

Constructor & Destructor Documentation

◆ TransientDriver()

Member Function Documentation

◆ advance_step()

◆ apply_ic()

◆ apply_predictor()

◆ diagnose()

◆ expected_options()

◆ output()

◆ postprocess()

◆ result()

◆ run()

◆ save_as_path()

◆ setup()

◆ solve()

◆ solve_step()

◆ store_input()

◆ update_forces()

Member Data Documentation

◆ _driving_force_names

◆ _driving_forces

◆ _in

◆ _nsteps

◆ _predictor

◆ _result_in

◆ _result_out

◆ _save_as

◆ _step_count

◆ _time

◆ _time_name