LevelGodunov Class Reference

#include <LevelGodunov.H>

Detailed Description

Level Godunov - really a hyperbolic level integrator.


Public Member Functions
	LevelGodunov ()
	Default constructor.
	~LevelGodunov ()
	Destructor.
void	define (const DisjointBoxLayout &a_thisDisjointBoxLayout, const DisjointBoxLayout &a_coarserDisjointBoxLayout, const ProblemDomain &a_domain, const int &a_refineCoarse, const Real &a_dx, const GodunovPhysics *const a_godunovFactory, const int &a_normalPredOrder, const bool &a_useFourthOrderSlopes, const bool &a_usePrimLimiting, const bool &a_useCharLimiting, const bool &a_useFlattening, const bool &a_useArtificialViscosity, const Real &a_artificialViscosity, const bool &a_hasCoarser, const bool &a_hasFiner)
	Actual constructor.
Real	step (LevelData< FArrayBox > &a_U, LevelData< FArrayBox > a_flux[CH_SPACEDIM], LevelFluxRegister &a_finerFluxRegister, LevelFluxRegister &a_coarserFluxRegister, const LevelData< FArrayBox > &a_S, const LevelData< FArrayBox > &a_UCoarseOld, const Real &a_TCoarseOld, const LevelData< FArrayBox > &a_UCoarseNew, const Real &a_TCoarseNew, const Real &a_time, const Real &a_dt)
	Take one timestep for this grid level.
void	computeWHalf (LayoutData< FluxBox > &a_WHalf, LevelData< FArrayBox > &a_U, const LevelData< FArrayBox > &a_S, const LevelData< FArrayBox > &a_UCoarseOld, const Real &a_TCoarseOld, const LevelData< FArrayBox > &a_UCoarseNew, const Real &a_TCoarseNew, const Real &a_time, const Real &a_dt)
	Compute the time-centered values of the primitive variables on cell faces.
Real	computeUpdate (LevelData< FArrayBox > &a_dU, LevelFluxRegister &a_finerFluxRegister, LevelFluxRegister &a_coarserFluxRegister, const LevelData< FArrayBox > &a_U, const LayoutData< FluxBox > &a_WHalf, const Real &a_time, const Real &a_dt)
	Compute the increment in the conserved variables from face variables.
Real	getMaxWaveSpeed (const LevelData< FArrayBox > &a_U)
	Return maximum wave speed.
GodunovPhysics *	getGodunovPhysicsPtr ()
	Get the GodunovPhysics pointer from the PatchGodunov member.
const GodunovPhysics *	getGodunovPhysicsPtrConst () const
void	highOrderLimiter (bool a_highOrderLimiter)
	Set whether to use high-order limiter.
Protected Attributes
DisjointBoxLayout	m_grids
LayoutData< PatchGodunov >	m_patchGodunov
PatchGodunov	m_patchGodunovNoBox
	just for global calls
int	m_numGhost
Copier	m_exchangeCopier
LevelData< FArrayBox >	m_U
PiecewiseLinearFillPatch	m_patcher
Real	m_dx
ProblemDomain	m_domain
int	m_refineCoarse
int	m_numCons
int	m_numFluxes
bool	m_hasCoarser
bool	m_hasFiner
int	m_normalPredOrder
bool	m_useFourthOrderSlopes
bool	m_usePrimLimiting
bool	m_useCharLimiting
bool	m_useFlattening
bool	m_useArtificialViscosity
Real	m_artificialViscosity
bool	m_isDefined
Private Member Functions
void	operator= (const LevelGodunov &)
	LevelGodunov (const LevelGodunov &)

Constructor & Destructor Documentation

LevelGodunov::LevelGodunov ( )

Default constructor.

Object requires define() to be called before all other functions.

LevelGodunov::~LevelGodunov ( )

Destructor.

Destroys all objects created by define(). Passed in data references of define() are left alone.

LevelGodunov::LevelGodunov ( const LevelGodunov & ) [private]

Member Function Documentation

void LevelGodunov::define	(	const DisjointBoxLayout &	a_thisDisjointBoxLayout,
		const DisjointBoxLayout &	a_coarserDisjointBoxLayout,
		const ProblemDomain &	a_domain,
		const int &	a_refineCoarse,
		const Real &	a_dx,
		const GodunovPhysics *const	a_godunovFactory,
		const int &	a_normalPredOrder,
		const bool &	a_useFourthOrderSlopes,
		const bool &	a_usePrimLimiting,
		const bool &	a_useCharLimiting,
		const bool &	a_useFlattening,
		const bool &	a_useArtificialViscosity,
		const Real &	a_artificialViscosity,
		const bool &	a_hasCoarser,
		const bool &	a_hasFiner
	)

Actual constructor.

Inside the routine, we cast away const-ness on the data members for the assignment. The arguments passed in are maintained const (coding standards).

Parameters:

a_thisDisjointBoxLayout	box layout at this level
a_coarserDisjointBoxLayout	box layout at next coarser level (or empty if this is coarsest level)
a_domain	problem domain at this level
a_refineCoarse	refinement ratio between this level and next coarser level
a_dx	grid spacing at this level
a_godunovFactory	pointer to GodunovPhysics class
a_normalPredOrder	order of the normal predictor: 1 for PLM, 2 for PPM
a_useFourthOrderSlopes	whether to use 4th-order slopes
a_usePrimLimiting	whether to apply slope limiting to primitive variables
a_useCharLimiting	whether to apply slope limiting to characteristic variables
a_useFlattening	whether to apply slope flattening
a_useArtificialViscosity	whether to apply artificial viscosity
a_artificialViscosity	artificial viscosity coefficient
a_hasCoarser	whether there is a coarser level
a_hasFiner	whether there is a finer level

Real LevelGodunov::step	(	LevelData< FArrayBox > &	a_U,
		LevelData< FArrayBox >	a_flux[CH_SPACEDIM],
		LevelFluxRegister &	a_finerFluxRegister,
		LevelFluxRegister &	a_coarserFluxRegister,
		const LevelData< FArrayBox > &	a_S,
		const LevelData< FArrayBox > &	a_UCoarseOld,
		const Real &	a_TCoarseOld,
		const LevelData< FArrayBox > &	a_UCoarseNew,
		const Real &	a_TCoarseNew,
		const Real &	a_time,
		const Real &	a_dt
	)

Take one timestep for this grid level.

Parameters:

a_U	conserved variables at this level, defined on a_thisDisjointBoxLayout in define()
a_flux	flux, for passing face-centered data in and out of the function
a_finerFluxRegister	flux register with next finer level
a_coarserFluxRegister	flux register with next coarser level
a_S	source term, or if no source term then this is null constructed and not defined
a_UCoarseOld	conserved variables at coarser level at time of last coarser-level update, or empty if no coarser level; may be empty if interpolation not required
a_TCoarseOld	time of last update at coarser level
a_UCoarseNew	conserved variables at coarser level at time of next coarser-level update, or empty if no coarser level; may be empty if interpolation not required
a_TCoarseNew	time of next update at coarser level
a_time	current time
a_dt	time step

void LevelGodunov::computeWHalf	(	LayoutData< FluxBox > &	a_WHalf,
		LevelData< FArrayBox > &	a_U,
		const LevelData< FArrayBox > &	a_S,
		const LevelData< FArrayBox > &	a_UCoarseOld,
		const Real &	a_TCoarseOld,
		const LevelData< FArrayBox > &	a_UCoarseNew,
		const Real &	a_TCoarseNew,
		const Real &	a_time,
		const Real &	a_dt
	)

Compute the time-centered values of the primitive variables on cell faces.

This API is used in cases where some operation over the whole level must be performed on the face-centered variables prior to the final difference update. Examples include incompressible flow and MHD, in which it is necessary to compute the projection of a face-centered vector field on its divergence-free part. To complete the differencing, it is necessary to call the member function computeUpdate().

Parameters:

a_WHalf	primitive variables extrapolated to cell faces and a half time step
a_U	conserved variables on current grids
a_S	source term on current grid
a_UCoarseOld	conserved variables at coarser level at time of last coarser-level update, or empty if no coarser level; may be empty if interpolation not required
a_TCoarseOld	time of last update at coarser level
a_UCoarseNew	conserved variables at coarser level at time of next coarser-level update, or empty if no coarser level; may be empty if interpolation not required
a_TCoarseNew	time of next update at coarser level
a_time	current time
a_dt	time step

Real LevelGodunov::computeUpdate	(	LevelData< FArrayBox > &	a_dU,
		LevelFluxRegister &	a_finerFluxRegister,
		LevelFluxRegister &	a_coarserFluxRegister,
		const LevelData< FArrayBox > &	a_U,
		const LayoutData< FluxBox > &	a_WHalf,
		const Real &	a_time,
		const Real &	a_dt
	)

Compute the increment in the conserved variables from face variables.

Compute dU = dt*dU/dt, the change in the conserved variables over the time step. Function returns the maximum stable time step.

Parameters:

a_dU	change in conserved variables over the time step
a_finerFluxRegister	flux register with next finer level
a_coarserFluxRegister	flux register with next coarser level
a_U	conserved variables at beginning of time step
a_WHalf	primitive variables extrapolated to cell faces and a half time step
a_time	current time
a_dt	time step