A class representing a network of connected reactors. More...

#include <ReactorNet.h>

Inheritance diagram for ReactorNet:

Detailed Description

A class representing a network of connected reactors.

This class is used to integrate the governing equations for a network of reactors that are time dependent (Reactor, ConstPressureReactor) connected by various means, for example Wall, MassFlowController, Valve, or PressureController; or reactors dependent on a single spatial variable (FlowReactor).

Definition at line 29 of file ReactorNet.h.

Public Member Functions
	ReactorNet (shared_ptr< ReactorBase > reactor)
	Create reactor network containing single reactor.

	ReactorNet (span< shared_ptr< ReactorBase > > reactors)
	Create reactor network from multiple reactors.

	ReactorNet (const ReactorNet &)=delete

ReactorNet &	operator= (const ReactorNet &)=delete

void	advance (double t)
	Advance the state of all reactors in the independent variable (time or space).

double	advance (double t, bool applylimit)
	Advance the state of all reactors in the independent variable (time or space).

double	step ()
	Advance the state of all reactors with respect to the independent variable (time or space).

void	solveSteady (int loglevel=0)
	Solve directly for the steady-state solution.

Eigen::SparseMatrix< double >	steadyJacobian (double rdt=0.0)
	Get the Jacobian used by the steady-state solver.

ReactorBase &	reactor (int n)
	Return a reference to the n-th reactor in this network.

bool	verbose () const
	Returns `true` if verbose logging output is enabled.

void	setVerbose (bool v=true)
	Enable or disable verbose logging while setting up and integrating the reactor network.

Integrator &	integrator ()
	Return a reference to the integrator.

void	updateState (span< const double > y)
	Update the state of all the reactors in the network to correspond to the values in the solution vector y.

double	sensitivity (size_t k, size_t p)
	Return the sensitivity of the k-th solution component with respect to the p-th sensitivity parameter.

double	sensitivity (const string &component, size_t p, int reactor=0)
	Return the sensitivity of the component named component with respect to the p-th sensitivity parameter.

void	evalJacobian (double t, span< double > y, span< double > ydot, span< const double > p, Array2D *j)
	Evaluate the Jacobian matrix for the reactor network.

size_t	neq () const override
	Number of equations.

size_t	nReactors () const

void	eval (double t, span< const double > y, span< double > ydot, span< const double > p) override
	Evaluate the right-hand-side ODE function.

void	evalSteady (span< const double > y, span< double > residual)
	Evaluate the governing equations adapted for the steady-state solver.

void	evalDae (double t, span< const double > y, span< const double > ydot, span< const double > p, span< double > residual) override
	eval coupling for IDA / DAEs

void	getState (span< double > y) override
	Fill in the vector y with the current state of the system.

void	getStateDae (span< double > y, span< double > ydot) override
	Fill in the vectors y and ydot with the current state of the system.

virtual void	getDerivative (int k, span< double > dky)
	Return k-th derivative at the current state of the system.

void	getConstraints (span< double > constraints) override
	Given a vector of length neq(), mark which variables should be considered algebraic constraints.

size_t	nparams () const override
	Number of sensitivity parameters.

size_t	globalComponentIndex (const string &component, size_t reactor=0)
	Return the index corresponding to the component named component in the reactor with index reactor in the global state vector for the reactor network.

string	componentName (size_t i) const
	Return the name of the i-th component of the global state vector.

double	upperBound (size_t i) const
	Get the upper bound on the i-th component of the global state vector.

double	lowerBound (size_t i) const
	Get the lower bound on the i-th component of the global state vector.

void	resetBadValues (span< double > y)
	Reset physically or mathematically problematic values, such as negative species concentrations.

size_t	registerSensitivityParameter (const string &name, double value, double scale)
	Used by Reactor and Wall objects to register the addition of sensitivity parameters so that the ReactorNet can keep track of the order in which sensitivity parameters are added.

const string &	sensitivityParameterName (size_t p) const
	The name of the p-th sensitivity parameter added to this ReactorNet.

void	initialize ()
	Initialize the reactor network.

void	reinitialize ()
	Reinitialize the integrator.

void	setNeedsReinit ()
	Called to trigger integrator reinitialization before further integration.

virtual void	setMaxSteps (int nmax)
	Set the maximum number of internal integration steps the integrator will take before reaching the next output point.

void	setAdvanceLimits (span< const double > limits)
	Set absolute step size limits during advance.

bool	hasAdvanceLimits () const
	Check whether ReactorNet object uses advance limits.

bool	getAdvanceLimits (span< double > limits) const
	Retrieve absolute step size limits during advance.

void	preconditionerSetup (double t, span< const double > y, double gamma) override
	Evaluate the setup processes for the Jacobian preconditioner.

void	preconditionerSolve (span< const double > rhs, span< double > output) override
	Evaluate the linear system Ax=b where A is the preconditioner.

AnyMap	solverStats () const
	Get solver stats from integrator.

virtual void	setDerivativeSettings (AnyMap &settings)
	Set derivative settings of all reactors.

size_t	nRootFunctions () const override
	Root finding is enabled only while enforcing advance limits.

void	evalRootFunctions (double t, span< const double > y, span< double > gout) override
	Evaluate the advance-limit root function used to stop integration once a limit is met.

Methods to set up a simulation
void	setLinearSolverType (const string &linSolverType="DENSE")
	Set the type of linear solver used in the integration.

void	setPreconditioner (shared_ptr< SystemJacobian > preconditioner)
	Set preconditioner used by the linear solver.

void	setInitialTime (double time)
	Set the initial value of the independent variable (typically time).

double	getInitialTime () const
	Get the initial value of the independent variable (typically time).

double	maxTimeStep () const
	Get the maximum integrator step.

void	setMaxTimeStep (double maxstep)
	Set the maximum integrator step.

void	setMaxErrTestFails (int nmax)
	Set the maximum number of error test failures permitted by the CVODES integrator in a single step.

void	setRelativeTolerance (double rtol)
	Set the relative tolerance for the integrator.

void	setAbsoluteTolerance (double atol)
	Set the scalar absolute tolerance for the integrator.

void	clearAbsoluteTolerance ()
	Clear the user-specified scalar absolute tolerance.

void	setTolerances (double rtol, double atol)
	Set the relative and scalar absolute tolerances for the integrator.

void	setSensitivityTolerances (double rtol, double atol)
	Set the relative and absolute tolerances for integrating the sensitivity equations.

double	time ()
	Current value of the simulation time [s], for reactor networks that are solved in the time domain.

double	distance ()
	Current position [m] along the length of the reactor network, for reactors that are solved as a function of space.

double	rtol ()
	Relative tolerance.

double	atol ()
	Scalar absolute integration tolerance.

double	rtolSensitivity () const
	Relative sensitivity tolerance.

double	atolSensitivity () const
	Absolute sensitivity tolerance.

string	linearSolverType () const
	Problem type of integrator.

int	maxSteps ()
	Returns the maximum number of internal integration steps the integrator will take before reaching the next output point.

Public Member Functions inherited from FuncEval
virtual void	eval (double t, span< const double > y, span< double > ydot, span< const double > p)
	Evaluate the right-hand-side ODE function.

virtual void	evalDae (double t, span< const double > y, span< const double > ydot, span< const double > p, span< double > residual)
	Evaluate the right-hand-side DAE function.

virtual void	getConstraints (span< double > constraints)
	Given a vector of length neq(), mark which variables should be considered algebraic constraints.

int	evalNoThrow (double t, span< const double > y, span< double > ydot)
	Evaluate the right-hand side using return code to indicate status.

int	evalDaeNoThrow (double t, span< const double > y, span< const double > ydot, span< double > residual)
	Evaluate the right-hand side using return code to indicate status.

virtual void	preconditionerSetup (double t, span< const double > y, double gamma)
	Evaluate the setup processes for the Jacobian preconditioner.

virtual void	preconditionerSolve (span< const double > rhs, span< double > output)
	Evaluate the linear system Ax=b where A is the preconditioner.

virtual void	updatePreconditioner (double gamma)
	Update the preconditioner based on already computed jacobian values.

int	preconditioner_setup_nothrow (double t, span< const double > y, double gamma)
	Preconditioner setup that doesn't throw an error but returns a CVODES flag.

int	preconditioner_solve_nothrow (span< const double > rhs, span< double > output)
	Preconditioner solve that doesn't throw an error but returns a CVODES flag.

virtual size_t	nRootFunctions () const
	Number of event/root functions exposed to the integrator.

virtual void	evalRootFunctions (double t, span< const double > y, span< double > gout)
	Evaluate the event/root functions currently in play.

int	evalRootFunctionsNoThrow (double t, span< const double > y, span< double > gout)
	Wrapper for evalRootFunctions that converts exceptions to return codes.

virtual void	getState (span< double > y)
	Fill in the vector y with the current state of the system.

virtual void	getStateDae (span< double > y, span< double > ydot)
	Fill in the vectors y and ydot with the current state of the system.

virtual size_t	neq () const =0
	Number of equations.

virtual size_t	nparams () const
	Number of sensitivity parameters.

void	suppressErrors (bool suppress)
	Enable or disable suppression of errors when calling eval()

bool	suppressErrors () const
	Get current state of error suppression.

string	getErrors () const
	Return a string containing the text of any suppressed errors.

void	clearErrors ()
	Clear any previously-stored suppressed errors.

Protected Member Functions
virtual void	checkPreconditionerSupported () const
	Check that preconditioning is supported by all reactors in the network.

void	updateTolerances ()
	Update the integrator tolerance vector from the current scalar settings, reactor-specific user tolerances, and reactor-specific default scaling rules.

void	updatePreconditioner (double gamma) override
	Update the preconditioner based on already computed jacobian values.

void	updateNames (ReactorBase &r)
	Create reproducible names for reactors and walls/connectors.

virtual void	getEstimate (double time, int k, span< double > yest)
	Estimate a future state based on current derivatives.

virtual int	lastOrder () const
	Returns the order used for last solution step of the ODE integrator The function is intended for internal use by ReactorNet::advance and deliberately not exposed in external interfaces.

Protected Attributes
vector< shared_ptr< ReactorBase > >	m_reactors

vector< shared_ptr< Reactor > >	m_bulkReactors

vector< shared_ptr< ReactorBase > >	m_surfaces

vector< shared_ptr< ReactorBase > >	m_reservoirs

set< FlowDevice * >	m_flowDevices

set< WallBase * >	m_walls

map< string, int >	m_counts
	Map used for default name generation.

unique_ptr< Integrator >	m_integ

double	m_time = 0.0
	The independent variable in the system.

double	m_initial_time = 0.0
	The initial value of the independent variable in the system.

bool	m_integratorInitialized = false
	True if the integrator has been initialized at least once.

bool	m_needIntegratorInit = true
	True if integrator needs to be (re)initialized.

size_t	m_nv = 0

vector< double >	m_atol

double	m_rtol = 1.0e-9

double	m_rtolsens = 1.0e-4

double	m_atols = 1.0e-15

double	m_atolsens = 1.0e-6

bool	m_atolUserSpecified = false
	True if scalar absolute tolerance was user-specified.

shared_ptr< SystemJacobian >	m_precon

string	m_linearSolverType

double	m_maxstep = 0.0
	Maximum integrator internal timestep. Default of 0.0 means infinity.

bool	m_verbose = false

bool	m_timeIsIndependent = true
	Indicates whether time or space is the independent variable.

vector< string >	m_paramNames
	Names corresponding to each sensitivity parameter.

vector< double >	m_ydot

vector< double >	m_yest

vector< double >	m_advancelimits

vector< double >	m_ybase
	Base state used for evaluating advance limits during a single advance() call when root-finding is enabled.

double	m_ybase_time = 0.0
	Base time corresponding to m_ybase.

bool	m_limit_check_active = false
	Indicates whether the advance-limit root check is active for the current call to `advance(t, applylimit=true)`

vector< double >	m_LHS
	m_LHS is a vector representing the coefficients on the "left hand side" of each governing equation

vector< double >	m_RHS

Protected Attributes inherited from FuncEval
bool	m_suppress_errors = false

vector< string >	m_errors
	Errors occurring during function evaluations.

Additional Inherited Members
Public Attributes inherited from FuncEval
vector< double >	m_sens_params
	Values for the problem parameters for which sensitivities are computed This is the array which is perturbed and passed back as the fourth argument to eval().

vector< double >	m_paramScales
	Scaling factors for each sensitivity parameter.

Constructor & Destructor Documentation

◆ ReactorNet() [1/2]

ReactorNet ( shared_ptr< ReactorBase > reactor )

Create reactor network containing single reactor.

Since: New in Cantera 3.2.

Definition at line 32 of file ReactorNet.cpp.

◆ ReactorNet() [2/2]

ReactorNet ( span< shared_ptr< ReactorBase > > reactors )

Create reactor network from multiple reactors.

Since: New in Cantera 3.2.

Definition at line 37 of file ReactorNet.cpp.

◆ ~ReactorNet()

~ReactorNet ( )

Definition at line 28 of file ReactorNet.cpp.

Member Function Documentation

◆ setLinearSolverType()

void setLinearSolverType ( const string & linSolverType = "DENSE" )

Set the type of linear solver used in the integration.

Parameters

linSolverType type of linear solver. Default type: "DENSE" Other options include: "DIAG", "DENSE", "GMRES", "BAND"

Definition at line 324 of file ReactorNet.cpp.

◆ setPreconditioner()

void setPreconditioner ( shared_ptr< SystemJacobian > preconditioner )

Set preconditioner used by the linear solver.

Parameters

preconditioner preconditioner object used for the linear solver

Definition at line 330 of file ReactorNet.cpp.

◆ setInitialTime()

void setInitialTime ( double time )

Set the initial value of the independent variable (typically time).

Default = 0.0 s. Restarts integration from this value using the current mixture state as the initial condition.

Definition at line 172 of file ReactorNet.cpp.

◆ getInitialTime()

double getInitialTime ( ) const

inline

Get the initial value of the independent variable (typically time).

Since: New in Cantera 3.0.

Definition at line 63 of file ReactorNet.h.

◆ maxTimeStep()

double maxTimeStep ( ) const

inline

Get the maximum integrator step.

Definition at line 68 of file ReactorNet.h.

◆ setMaxTimeStep()

void setMaxTimeStep ( double maxstep )

Set the maximum integrator step.

Definition at line 179 of file ReactorNet.cpp.

◆ setMaxErrTestFails()

void setMaxErrTestFails ( int nmax )

Set the maximum number of error test failures permitted by the CVODES integrator in a single step.

Definition at line 185 of file ReactorNet.cpp.

◆ setRelativeTolerance()

void setRelativeTolerance ( double rtol )

Set the relative tolerance for the integrator.

Parameters

rtol	Relative tolerance for all state variables; must be positive.

Since: New in Cantera 4.0.

Definition at line 190 of file ReactorNet.cpp.

◆ setAbsoluteTolerance()

void setAbsoluteTolerance ( double atol )

Set the scalar absolute tolerance for the integrator.

Parameters

atol	Scalar absolute tolerance for state variables without reactor-specific absolute tolerances; must be positive. Superseded by absolute tolerances set for individual reactors.

See also: ReactorBase::setAbsoluteTolerances

Since: New in Cantera 4.0.

Definition at line 200 of file ReactorNet.cpp.

◆ clearAbsoluteTolerance()

void clearAbsoluteTolerance ( )

Clear the user-specified scalar absolute tolerance.

Restores the default scalar absolute tolerance and enables reactor-specific default scaling rules.

See also: ReactorBase::updateDefaultTolerances

Since: New in Cantera 4.0.

Definition at line 211 of file ReactorNet.cpp.

◆ setTolerances()

void setTolerances	(	double	rtol,
		double	atol
	)

Set the relative and scalar absolute tolerances for the integrator.

Parameters

rtol	If positive, set the relative tolerance for all state variables; ignored if negative.
atol	If positive, set the scalar absolute tolerance for all state variables; ignored if negative. Superseded by absolute tolerances set for individual reactors.

Deprecated:: To be removed after Cantera 4.0. Use setRelativeTolerance() and setAbsoluteTolerance() instead.

Definition at line 218 of file ReactorNet.cpp.

◆ setSensitivityTolerances()

void setSensitivityTolerances	(	double	rtol,
		double	atol
	)

Set the relative and absolute tolerances for integrating the sensitivity equations.

Definition at line 247 of file ReactorNet.cpp.

◆ time()

double time ( )

Current value of the simulation time [s], for reactor networks that are solved in the time domain.

Definition at line 258 of file ReactorNet.cpp.

◆ distance()

double distance ( )

Current position [m] along the length of the reactor network, for reactors that are solved as a function of space.

Definition at line 267 of file ReactorNet.cpp.

◆ rtol()

double rtol ( )

inline

Relative tolerance.

Definition at line 126 of file ReactorNet.h.

◆ atol()

double atol ( )

inline

Scalar absolute integration tolerance.

Definition at line 131 of file ReactorNet.h.

◆ rtolSensitivity()

double rtolSensitivity ( ) const

inline

Relative sensitivity tolerance.

Definition at line 136 of file ReactorNet.h.

◆ atolSensitivity()

double atolSensitivity ( ) const

inline

Absolute sensitivity tolerance.

Definition at line 141 of file ReactorNet.h.

◆ linearSolverType()

string linearSolverType ( ) const

Problem type of integrator.

Definition at line 819 of file ReactorNet.cpp.

◆ maxSteps()

int maxSteps ( )

Returns the maximum number of internal integration steps the integrator will take before reaching the next output point.

Definition at line 344 of file ReactorNet.cpp.

◆ advance() [1/2]

void advance ( double t )

Advance the state of all reactors in the independent variable (time or space).

Take as many internal steps as necessary to reach t.

Parameters

t	Time/distance to advance to (s or m).

Definition at line 349 of file ReactorNet.cpp.

◆ advance() [2/2]

double advance	(	double	t,
		bool	applylimit
	)

Advance the state of all reactors in the independent variable (time or space).

Take as many internal steps as necessary towards t. If applylimit is true, the advance step will be automatically reduced if needed to stay within limits (set by setAdvanceLimit). Returns the time/distance at the end of integration.

Parameters

t	Time/distance to advance to (s or m).
applylimit	Limit advance step (boolean).

Definition at line 357 of file ReactorNet.cpp.

◆ step()

double step ( )

Advance the state of all reactors with respect to the independent variable (time or space).

Returns the new value of the independent variable [s or m].

Definition at line 433 of file ReactorNet.cpp.

◆ solveSteady()

void solveSteady ( int loglevel = 0 )

Solve directly for the steady-state solution.

This approach is generally more efficient than time marching to the steady-state, but imposes a few limitations:

The volume of control volume reactor types (such as Reactor and IdealGasMoleReactor) must be constant; no moving walls can be used.
The mass of constant pressure reactor types (such as ConstPressureReactor and IdealGasConstPressureReactor) must be constant; if flow devices are used, inlet and outlet flows must be balanced.
Only ideal gas reactor types can be used for when the energy equation is disabled (fixed temperature simulations).

Parameters

loglevel

Print information about solver progress to aid in understanding cases where the solver fails to converge. Higher levels are more verbose.

0: No logging.
1: Basic info about each steady-state attempt and round of time stepping.
2: Adds details about each time step and steady-state Newton iteration.
3: Adds details about Newton iterations for each time step.
4: Adds details about state variables that are limiting steady-state Newton step sizes.
5: Adds details about state variables that are limiting time-stepping Newton step sizes.
6: Print current state vector after different solver stages
7: Print current residual vector after different solver stages

See also: SteadyStateSystem, MultiNewton

Since: New in Cantera 3.2. Support for reacting surfaces added in Cantera 4.0.

Definition at line 441 of file ReactorNet.cpp.

◆ steadyJacobian()

Eigen::SparseMatrix< double > steadyJacobian ( double rdt = 0.0 )

Get the Jacobian used by the steady-state solver.

Parameters

rdt	Reciprocal of the pseudo-timestep [1/s]. Default of 0.0 returns the steady-state Jacobian.

Since: New in Cantera 3.2.

Definition at line 453 of file ReactorNet.cpp.

◆ reactor()

ReactorBase & reactor ( int n )

inline

Return a reference to the n-th reactor in this network.

The reactor indices are determined by the order in which the reactors were added to the reactor network.

Definition at line 216 of file ReactorNet.h.

◆ verbose()

bool verbose ( ) const

inline

Returns true if verbose logging output is enabled.

Definition at line 221 of file ReactorNet.h.

◆ setVerbose()

void setVerbose ( bool v = true )

inline

Enable or disable verbose logging while setting up and integrating the reactor network.

Definition at line 227 of file ReactorNet.h.

◆ integrator()

Integrator & integrator ( )

Return a reference to the integrator.

Only valid after adding at least one reactor to the network.

Definition at line 565 of file ReactorNet.cpp.

◆ updateState()

void updateState ( span< const double > y )

Update the state of all the reactors in the network to correspond to the values in the solution vector y.

Definition at line 669 of file ReactorNet.cpp.

◆ sensitivity() [1/2]

double sensitivity	(	size_t	k,
		size_t	p
	)

Return the sensitivity of the k-th solution component with respect to the p-th sensitivity parameter.

The sensitivity coefficient \( S_{ki} \) of solution variable \( y_k \) with respect to sensitivity parameter \( p_i \) is defined as:

\[ S_{ki} = \frac{1}{y_k} \frac{\partial y_k}{\partial p_i} \]

For reaction sensitivities, the parameter is a multiplier on the forward rate constant (and implicitly on the reverse rate constant for reversible reactions) which has a nominal value of 1.0, and the sensitivity is nondimensional.

For species enthalpy sensitivities, the parameter is a perturbation to the molar enthalpy of formation, such that the dimensions of the sensitivity are kmol/J.

Definition at line 632 of file ReactorNet.cpp.

◆ sensitivity() [2/2]

double sensitivity	(	const string &	component,
		size_t	p,
		int	reactor = `0`
	)

inline

Return the sensitivity of the component named component with respect to the p-th sensitivity parameter.

The sensitivity coefficient \( S_{ki} \) of solution variable \( y_k \) with respect to sensitivity parameter \( p_i \) is defined as:

\[ S_{ki} = \frac{1}{y_k} \frac{\partial y_k}{\partial p_i} \]

For reaction sensitivities, the parameter is a multiplier on the forward rate constant (and implicitly on the reverse rate constant for reversible reactions) which has a nominal value of 1.0, and the sensitivity is nondimensional.

For species enthalpy sensitivities, the parameter is a perturbation to the molar enthalpy of formation, such that the dimensions of the sensitivity are kmol/J.

Definition at line 262 of file ReactorNet.h.

◆ evalJacobian()

void evalJacobian	(	double	t,
		span< double >	y,
		span< double >	ydot,
		span< const double >	p,
		Array2D *	j
	)

Evaluate the Jacobian matrix for the reactor network.

Parameters

[in]	t	Time/distance at which to evaluate the Jacobian
[in]	y	Global state vector at t
[out]	ydot	Derivative of the state vector evaluated at t, with respect to t.
[in]	p	sensitivity parameter vector (unused?)
[out]	j	Jacobian matrix, size neq() by neq().

Definition at line 646 of file ReactorNet.cpp.

◆ neq()

size_t neq ( ) const

inlineoverridevirtual

Number of equations.

Implements FuncEval.

Definition at line 280 of file ReactorNet.h.

◆ nReactors()

size_t nReactors ( ) const

inline

Definition at line 284 of file ReactorNet.h.

◆ eval()

void eval	(	double	t,
		span< const double >	y,
		span< double >	ydot,
		span< const double >	p
	)

overridevirtual

Evaluate the right-hand-side ODE function.

Called by the integrator.

Parameters

[in]	t	time.
[in]	y	solution vector, length neq()
[out]	ydot	rate of change of solution vector, length neq()
[in]	p	sensitivity parameter vector, length nparams()

Reimplemented from FuncEval.

Definition at line 573 of file ReactorNet.cpp.

◆ evalSteady()

void evalSteady	(	span< const double >	y,
		span< double >	residual
	)

Evaluate the governing equations adapted for the steady-state solver.

Parameters

[in]	y	Current state vector, length neq()
[out]	residual	For differential variables, this is the time derivative; for algebraic variables, this is the residual of the governing equation.

Since: New in Cantera 4.0.

Definition at line 593 of file ReactorNet.cpp.

◆ evalDae()

void evalDae	(	double	t,
		span< const double >	y,
		span< const double >	ydot,
		span< const double >	p,
		span< double >	residual
	)

overridevirtual

eval coupling for IDA / DAEs

Reimplemented from FuncEval.

Definition at line 609 of file ReactorNet.cpp.

◆ getState()

void getState ( span< double > y )

overridevirtual

Fill in the vector y with the current state of the system.

Used for getting the initial state for ODE systems.

Reimplemented from FuncEval.

Definition at line 710 of file ReactorNet.cpp.

◆ getStateDae()

void getStateDae	(	span< double >	y,
		span< double >	ydot
	)

overridevirtual

Fill in the vectors y and ydot with the current state of the system.

Used for getting the initial state for DAE systems.

Reimplemented from FuncEval.

Definition at line 717 of file ReactorNet.cpp.

◆ getDerivative()

void getDerivative	(	int	k,
		span< double >	dky
	)

virtual

Return k-th derivative at the current state of the system.

Definition at line 677 of file ReactorNet.cpp.

◆ getConstraints()

void getConstraints ( span< double > constraints )

overridevirtual

Given a vector of length neq(), mark which variables should be considered algebraic constraints.

Reimplemented from FuncEval.

Definition at line 625 of file ReactorNet.cpp.

◆ nparams()

size_t nparams ( ) const

inlineoverridevirtual

Number of sensitivity parameters.

Reimplemented from FuncEval.

Definition at line 312 of file ReactorNet.h.

◆ globalComponentIndex()

size_t globalComponentIndex	(	const string &	component,
		size_t	reactor = `0`
	)

Return the index corresponding to the component named component in the reactor with index reactor in the global state vector for the reactor network.

Definition at line 730 of file ReactorNet.cpp.

◆ componentName()

string componentName ( size_t i ) const

Return the name of the i-th component of the global state vector.

The name returned includes both the name of the reactor and the specific component, for example ‘'reactor1: CH4’`.

Definition at line 737 of file ReactorNet.cpp.

◆ upperBound()

double upperBound ( size_t i ) const

Get the upper bound on the i-th component of the global state vector.

Definition at line 752 of file ReactorNet.cpp.

◆ lowerBound()

double lowerBound ( size_t i ) const

Get the lower bound on the i-th component of the global state vector.

Definition at line 767 of file ReactorNet.cpp.

◆ resetBadValues()

void resetBadValues ( span< double > y )

Reset physically or mathematically problematic values, such as negative species concentrations.

This method is used within solveSteady() if certain errors are encountered.

Parameters

[in,out] y current state vector, to be updated; length neq()

Definition at line 782 of file ReactorNet.cpp.

◆ registerSensitivityParameter()

size_t registerSensitivityParameter	(	const string &	name,
		double	value,
		double	scale
	)

Used by Reactor and Wall objects to register the addition of sensitivity parameters so that the ReactorNet can keep track of the order in which sensitivity parameters are added.

Parameters

name	A name describing the parameter, for example the reaction string
value	The nominal value of the parameter
scale	A scaling factor to be applied to the sensitivity coefficient

Returns: the index of this parameter in the vector of sensitivity parameters (global across all reactors)

Definition at line 788 of file ReactorNet.cpp.

◆ sensitivityParameterName()

const string & sensitivityParameterName ( size_t p ) const

inline

The name of the p-th sensitivity parameter added to this ReactorNet.

Definition at line 352 of file ReactorNet.h.

◆ initialize()

void initialize ( )

Initialize the reactor network.

Called automatically the first time advance or step is called.

Definition at line 276 of file ReactorNet.cpp.

◆ reinitialize()

void reinitialize ( )

Reinitialize the integrator.

Used to solve a new problem (different initial conditions) but with the same configuration of the reactor network. Can be called manually, or automatically after calling setInitialTime or modifying a reactor's contents.

Definition at line 306 of file ReactorNet.cpp.

◆ setNeedsReinit()

void setNeedsReinit ( )

inline

Called to trigger integrator reinitialization before further integration.

Definition at line 368 of file ReactorNet.h.

◆ setMaxSteps()

void setMaxSteps ( int nmax )

virtual

Set the maximum number of internal integration steps the integrator will take before reaching the next output point.

Parameters

nmax	The maximum number of steps, setting this value to zero disables this option.

Definition at line 339 of file ReactorNet.cpp.

◆ setAdvanceLimits()

void setAdvanceLimits ( span< const double > limits )

Set absolute step size limits during advance.

Definition at line 684 of file ReactorNet.cpp.

◆ hasAdvanceLimits()

bool hasAdvanceLimits ( ) const

Check whether ReactorNet object uses advance limits.

Definition at line 692 of file ReactorNet.cpp.

◆ getAdvanceLimits()

bool getAdvanceLimits ( span< double > limits ) const

Retrieve absolute step size limits during advance.

Definition at line 701 of file ReactorNet.cpp.

◆ preconditionerSetup()

void preconditionerSetup	(	double	t,
		span< const double >	y,
		double	gamma
	)

overridevirtual

Evaluate the setup processes for the Jacobian preconditioner.

Parameters

[in]	t	time.
[in]	y	solution vector, length neq()
	gamma	the gamma in M=I-gamma*J

Warning: This function is an experimental part of the Cantera API and may be changed or removed without notice.

Reimplemented from FuncEval.

Definition at line 837 of file ReactorNet.cpp.

◆ preconditionerSolve()

void preconditionerSolve	(	span< const double >	rhs,
		span< double >	output
	)

overridevirtual

Evaluate the linear system Ax=b where A is the preconditioner.

Parameters

[in]	rhs	right hand side vector used in linear system
[out]	output	output vector for solution

Warning: This function is an experimental part of the Cantera API and may be changed or removed without notice.

Reimplemented from FuncEval.

Definition at line 828 of file ReactorNet.cpp.

◆ solverStats()

AnyMap solverStats ( ) const

Get solver stats from integrator.

Definition at line 810 of file ReactorNet.cpp.

◆ setDerivativeSettings()

void setDerivativeSettings ( AnyMap & settings )

virtual

Set derivative settings of all reactors.

Parameters

settings the settings map propagated to all reactors and kinetics objects

Definition at line 802 of file ReactorNet.cpp.

◆ nRootFunctions()

size_t nRootFunctions ( ) const

overridevirtual

Root finding is enabled only while enforcing advance limits.

Reimplemented from FuncEval.

Definition at line 494 of file ReactorNet.cpp.

◆ evalRootFunctions()

void evalRootFunctions	(	double	t,
		span< const double >	y,
		span< double >	gout
	)

overridevirtual

Evaluate the advance-limit root function used to stop integration once a limit is met.

When limits are active, this sets gout[0] to 1 - max_i(|y[i]-y_base[i]| / limit[i]) so a zero indicates a component has reached its limit; otherwise gout[0] is positive.

Reimplemented from FuncEval.

Definition at line 499 of file ReactorNet.cpp.

◆ checkPreconditionerSupported()

void checkPreconditionerSupported ( ) const

protectedvirtual

Check that preconditioning is supported by all reactors in the network.

Definition at line 877 of file ReactorNet.cpp.

◆ updateTolerances()

void updateTolerances ( )

protected

Update the integrator tolerance vector from the current scalar settings, reactor-specific user tolerances, and reactor-specific default scaling rules.

Definition at line 233 of file ReactorNet.cpp.

◆ updatePreconditioner()

void updatePreconditioner ( double gamma )

overrideprotectedvirtual

Update the preconditioner based on already computed jacobian values.

Reimplemented from FuncEval.

Definition at line 866 of file ReactorNet.cpp.

◆ updateNames()

void updateNames ( ReactorBase & r )

protected

Create reproducible names for reactors and walls/connectors.

Definition at line 528 of file ReactorNet.cpp.

◆ getEstimate()

void getEstimate	(	double	time,
		int	k,
		span< double >	yest
	)

protectedvirtual

Estimate a future state based on current derivatives.

The function is intended for internal use by ReactorNet::advance and deliberately not exposed in external interfaces.

Definition at line 467 of file ReactorNet.cpp.

◆ lastOrder()

int lastOrder ( ) const

protectedvirtual

Returns the order used for last solution step of the ODE integrator The function is intended for internal use by ReactorNet::advance and deliberately not exposed in external interfaces.

Definition at line 485 of file ReactorNet.cpp.

Member Data Documentation

◆ m_reactors

vector<shared_ptr<ReactorBase> > m_reactors

protected

Definition at line 433 of file ReactorNet.h.

◆ m_bulkReactors

vector<shared_ptr<Reactor> > m_bulkReactors

protected

Definition at line 434 of file ReactorNet.h.

◆ m_surfaces

vector<shared_ptr<ReactorBase> > m_surfaces

protected

Definition at line 435 of file ReactorNet.h.

◆ m_reservoirs

vector<shared_ptr<ReactorBase> > m_reservoirs

protected

Definition at line 436 of file ReactorNet.h.

◆ m_flowDevices

set<FlowDevice*> m_flowDevices

protected

Definition at line 437 of file ReactorNet.h.

◆ m_walls

set<WallBase*> m_walls

protected

Definition at line 438 of file ReactorNet.h.

◆ m_counts

map<string, int> m_counts

protected

Map used for default name generation.

Definition at line 439 of file ReactorNet.h.

◆ m_integ

unique_ptr<Integrator> m_integ

protected

Definition at line 440 of file ReactorNet.h.

◆ m_time

double m_time = 0.0

protected

The independent variable in the system.

May be either time or space depending on the type of reactors in the network.

Definition at line 444 of file ReactorNet.h.

◆ m_initial_time

double m_initial_time = 0.0

protected

The initial value of the independent variable in the system.

Definition at line 447 of file ReactorNet.h.

◆ m_integratorInitialized

bool m_integratorInitialized = false

protected

True if the integrator has been initialized at least once.

Definition at line 449 of file ReactorNet.h.

◆ m_needIntegratorInit

bool m_needIntegratorInit = true

protected

True if integrator needs to be (re)initialized.

Definition at line 450 of file ReactorNet.h.

◆ m_nv

size_t m_nv = 0

protected

Definition at line 451 of file ReactorNet.h.

◆ m_atol

vector<double> m_atol

protected

Definition at line 453 of file ReactorNet.h.

◆ m_rtol

double m_rtol = 1.0e-9

protected

Definition at line 454 of file ReactorNet.h.

◆ m_rtolsens

double m_rtolsens = 1.0e-4

protected

Definition at line 455 of file ReactorNet.h.

◆ m_atols

double m_atols = 1.0e-15

protected

Definition at line 456 of file ReactorNet.h.

◆ m_atolsens

double m_atolsens = 1.0e-6

protected

Definition at line 457 of file ReactorNet.h.

◆ m_atolUserSpecified

bool m_atolUserSpecified = false

protected

True if scalar absolute tolerance was user-specified.

Definition at line 459 of file ReactorNet.h.

◆ m_precon

shared_ptr<SystemJacobian> m_precon

protected

Definition at line 460 of file ReactorNet.h.

◆ m_linearSolverType

string m_linearSolverType

protected

Definition at line 461 of file ReactorNet.h.

◆ m_maxstep

double m_maxstep = 0.0

protected

Maximum integrator internal timestep. Default of 0.0 means infinity.

Definition at line 464 of file ReactorNet.h.

◆ m_verbose

bool m_verbose = false

protected

Definition at line 466 of file ReactorNet.h.

◆ m_timeIsIndependent

bool m_timeIsIndependent = true

protected

Indicates whether time or space is the independent variable.

Definition at line 469 of file ReactorNet.h.

◆ m_paramNames

vector<string> m_paramNames

protected

Names corresponding to each sensitivity parameter.

Definition at line 472 of file ReactorNet.h.

◆ m_ydot

vector<double> m_ydot

protected

Definition at line 474 of file ReactorNet.h.

◆ m_yest

vector<double> m_yest

protected

Definition at line 475 of file ReactorNet.h.

◆ m_advancelimits

vector<double> m_advancelimits

protected

Definition at line 476 of file ReactorNet.h.

◆ m_ybase

vector<double> m_ybase

protected

Base state used for evaluating advance limits during a single advance() call when root-finding is enabled.

Definition at line 479 of file ReactorNet.h.

◆ m_ybase_time

double m_ybase_time = 0.0

protected

Base time corresponding to m_ybase.

Definition at line 481 of file ReactorNet.h.

◆ m_limit_check_active

bool m_limit_check_active = false

protected

Indicates whether the advance-limit root check is active for the current call to advance(t, applylimit=true)

Definition at line 484 of file ReactorNet.h.

◆ m_LHS

vector<double> m_LHS

protected

m_LHS is a vector representing the coefficients on the "left hand side" of each governing equation

Definition at line 487 of file ReactorNet.h.

◆ m_RHS

vector<double> m_RHS

protected

Definition at line 488 of file ReactorNet.h.

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

Detailed Description

Public Member Functions

Protected Member Functions

Protected Attributes

Additional Inherited Members

Constructor & Destructor Documentation

◆ ReactorNet() [1/2]

◆ ReactorNet() [2/2]

◆ ~ReactorNet()

Member Function Documentation

◆ setLinearSolverType()

◆ setPreconditioner()

◆ setInitialTime()

◆ getInitialTime()

◆ maxTimeStep()

◆ setMaxTimeStep()

◆ setMaxErrTestFails()

◆ setRelativeTolerance()

◆ setAbsoluteTolerance()

◆ clearAbsoluteTolerance()

◆ setTolerances()

◆ setSensitivityTolerances()

◆ time()

◆ distance()

◆ rtol()

◆ atol()

◆ rtolSensitivity()

◆ atolSensitivity()

◆ linearSolverType()

◆ maxSteps()

◆ advance() [1/2]

◆ advance() [2/2]

◆ step()

◆ solveSteady()

◆ steadyJacobian()

◆ reactor()

◆ verbose()

◆ setVerbose()

◆ integrator()

◆ updateState()

◆ sensitivity() [1/2]

◆ sensitivity() [2/2]

◆ evalJacobian()

◆ neq()

◆ nReactors()

◆ eval()

◆ evalSteady()

◆ evalDae()

◆ getState()

◆ getStateDae()

◆ getDerivative()

◆ getConstraints()

◆ nparams()

◆ globalComponentIndex()

◆ componentName()

◆ upperBound()

◆ lowerBound()

◆ resetBadValues()

◆ registerSensitivityParameter()

◆ sensitivityParameterName()

◆ initialize()

◆ reinitialize()

◆ setNeedsReinit()

◆ setMaxSteps()

◆ setAdvanceLimits()

◆ hasAdvanceLimits()

◆ getAdvanceLimits()

◆ preconditionerSetup()

◆ preconditionerSolve()

◆ solverStats()

◆ setDerivativeSettings()

◆ nRootFunctions()

◆ evalRootFunctions()

◆ checkPreconditionerSupported()

◆ updateTolerances()

◆ updatePreconditioner()

◆ updateNames()

◆ getEstimate()

◆ lastOrder()

Member Data Documentation