Main Page   Namespace List   Class Hierarchy   Alphabetical List   Compound List   File List   Compound Members   File Members  

PatchGodunov Class Reference

#include <PatchGodunov.H>

Collaboration diagram for PatchGodunov:

Collaboration graph
[legend]
List of all members.

Public Methods

 PatchGodunov ()
 Constructor.

virtual ~PatchGodunov ()
 Destructor.

virtual void define (ProblemDomain &a_domain, const Real &a_dx)
 Define the object.

virtual void setPhysIBC (PhysIBC *a_bc)
 Set the initial and boundary condition object.

virtual PhysIBC * getPhysIBC () const
 Get the initial and boundary condition object.

virtual void setSlopeParameters (bool a_fourthOrderSlopes, bool a_flattening, bool a_limitSlopes=true)
 Set parameters for slope computations.

virtual bool useFourthOrderSlopes ()
 Fourth-order slope query.

virtual bool useFlattening ()
 Slope flattening query.

virtual bool limitSlopes ()
 Slope limiter query.

virtual void setArtificialViscosity (bool a_useArtificialViscosity, Real a_artificialViscosity)
 Set parameters for artificial viscosity.

virtual bool useArtificialViscosity ()
 Artificial viscosity query.

virtual Real artificialViscosityCoefficient ()
 Artificial viscosity coefficient.

virtual PatchGodunov * new_patchGodunov () const=0
 Factory method - this object is its own factory.

virtual void setCurrentTime (const Real &a_currentTime)
 Set the current time before calling updateState().

virtual void setCurrentBox (const Box &a_currentBox)
 Set the current box before calling updateState().

virtual void updateState (FArrayBox &a_U, FArrayBox a_F[CH_SPACEDIM], Real &a_maxWaveSpeed, const FArrayBox &a_S, const Real &a_dt, const Box &a_box)
 Update the conserved variables and return the final fluxes used for this.

virtual Real getMaxWaveSpeed (const FArrayBox &a_U, const Box &a_box)=0
 Compute the maximum wave speed.

virtual int numConserved ()=0
 Number of conserved variables.

virtual Vector< string > stateNames ()
 Names of the conserved variables.

virtual int numFluxes ()=0
 Number of flux variables.


Protected Methods

virtual bool isDefined () const
 Is the object completely defined.

virtual int numPrimitives ()=0
 Number of primitive variables.

virtual int numSlopes ()=0
 Number of primitive variables for which slopes are computed.

virtual void consToPrim (FArrayBox &a_W, const FArrayBox &a_U, const Box &a_box)=0
 Compute the primitive variables from the conserved variables within a_box.

virtual void computeFlattening (FArrayBox &a_flattening, const FArrayBox &a_W, const Box &a_box)
 Compute the slope flattening coefficients.

virtual void slope (FArrayBox &a_dW, const FArrayBox &a_W, const FArrayBox &a_flattening, const int &a_dir, const Box &a_box)
 Compute the 2nd or 4th order slopes of the primitive variables.

virtual void normalPred (FArrayBox &a_WMinus, FArrayBox &a_WPlus, const FArrayBox &a_W, const FArrayBox &a_dW, const Real &a_scale, const int &a_dir, const Box &a_box)=0
 Extrapolate the primitive variables to the cell faces.

virtual void incrementWithSource (FArrayBox &a_W, const FArrayBox &a_S, const Real &a_scale, const Box &a_box)
 Increment the primitive variables by a source term.

virtual void riemann (FArrayBox &a_F, const FArrayBox &a_WLeft, const FArrayBox &a_WRight, const int &a_dir, const Box &a_box)=0
 Compute a Riemann problem and generate fluxes at the faces.

virtual void updatePrim (FArrayBox &a_WMinus, FArrayBox &a_WPlus, const FArrayBox &a_F, const Real &a_scale, const int &a_dir, const Box &a_box)=0
 Update the primitive variables using fluxes (of the conserved variables).

virtual void artificialViscosity (FArrayBox &a_F, const FArrayBox &a_U, const FArrayBox &a_divVel, const int &a_dir, const Box &a_box)
 Apply artificial viscosity to the fluxes.

virtual void updateCons (FArrayBox &a_U, const FArrayBox &a_F, const Real &a_scale, const int &a_dir, const Box &a_box)=0
 Update the conserved variable using fluxes and a scaling factor.

virtual void postUpdateCons (FArrayBox &a_U, const FArrayBox &a_Uold, const Real &a_dt, const Real &a_dx, const Box &a_box)
virtual Interval velocityInterval ()=0
 Interval within the primitive variables corresponding to the velocities.

virtual int pressureIndex ()=0
 Component index within the primitive variables of the pressure.

virtual int bulkModulusIndex ()=0
 Component index within the primitive variables of the bulk modulus.

virtual void applyLimiter (FArrayBox &a_dW, const FArrayBox &a_dWLeft, const FArrayBox &a_dWRight, const int &a_dir, const Box &a_box)
 Apply a slope limiter to computed slopes.

virtual void divVel (FArrayBox &a_divVel, const FArrayBox &a_W, const int a_dir, const Box &a_box)
 Compute the face centered divergence of the velocity.


Protected Attributes

bool m_isDefined
ProblemDomain m_domain
Real m_dx
bool m_useFourthOrderSlopes
bool m_useFlattening
bool m_limitSlopes
bool m_isSlopeSet
bool m_useArtificialViscosity
Real m_artificialViscosity
bool m_isArtViscSet
PhysIBC * m_bc
bool m_isBCSet
Real m_currentTime
bool m_isCurrentTimeSet
Box m_currentBox
bool m_isCurrentBoxSet

Detailed Description

The base class PatchGodunov provides an implementation of a second-order, unsplit Godunov method acting on a single grid/patch. PatchGodunov provides an interface to the level integrator, LevelGodunov, which manages the entire level and flux corrections (via flux registers) to the coarser and finer levels. In addition, the physics dependent code is not provided in PatchGodunov but is supplied by the user by subclassing PatchGodunov and implementing the pure virtual functions. Some parameters can also be adjusted to modify the algorithm. All functions are virtual so any of them can be reimplemented by the user.

There are three types of grid variables that appear in the unsplit Godunov method: conserved variables, primitive variables, and fluxes, denoted below by U, W, F, respectively. It is often convenient to have the number of primitive variables and fluxes exceed the number of conserved variables. In the case of primitive variables, redundant quantities are carried that parameterize the equation of state in order to avoid multiple calls to that the equation of state function. In the case of fluxes, it is often convenient to split the flux for some variables into multiple components, e.g., dividing the momentum flux into advective and pressure terms. The API given here provides the flexibility to support these various possibilities.

Constructor & Destructor Documentation

PatchGodunov::PatchGodunov (    ) 
 

Constructor.

virtual PatchGodunov::~PatchGodunov (    )  [virtual]
 

Destructor.

Member Function Documentation

virtual void PatchGodunov::applyLimiter (  FArrayBox &    a_dW,
const FArrayBox &    a_dWLeft,
const FArrayBox &    a_dWRight,
const int &    a_dir,
const Box &    a_box
)  [protected, virtual]
 

Apply a slope limiter to computed slopes.

Given the center difference, a_dW, and the left and right differences, a_dWLeft and a_dWRight, in direction, a_dir, apply a slope limiter within a_box to generate final slopes, a_dW. This is called by slope().

A default implementation is provided which implements a van Leer limiter, see the documentation for details.

virtual void PatchGodunov::artificialViscosity (  FArrayBox &    a_F,
const FArrayBox &    a_U,
const FArrayBox &    a_divVel,
const int &    a_dir,
const Box &    a_box
)  [protected, virtual]
 

Apply artificial viscosity to the fluxes.

Return fluxes, a_F, updated using artificial viscosity based on the conserved variables, a_U, and the face centered divergence of the velocity, a_divVel, in a given direction, a_dir, within a_box.

virtual Real PatchGodunov::artificialViscosityCoefficient (    )  [virtual]
 

Artificial viscosity coefficient.

virtual int PatchGodunov::bulkModulusIndex (    )  [protected, pure virtual]
 

Component index within the primitive variables of the bulk modulus.

Return the component index withn the primitive variables for the bulk modulus. Used for slope flattening (slope computation) used as a normalization to measure shock strength.

virtual void PatchGodunov::computeFlattening (  FArrayBox &    a_flattening,
const FArrayBox &    a_W,
const Box &    a_box
)  [protected, virtual]
 

Compute the slope flattening coefficients.

Compute the slope flattening coefficients, a_flattening, using the primitive variables, a_W, within a_box.

virtual void PatchGodunov::consToPrim (  FArrayBox &    a_W,
const FArrayBox &    a_U,
const Box &    a_box
)  [protected, pure virtual]
 

Compute the primitive variables from the conserved variables within a_box.

virtual void PatchGodunov::define (  ProblemDomain &    a_domain,
const Real &    a_dx
)  [virtual]
 

Define the object.

virtual void PatchGodunov::divVel (  FArrayBox &    a_divVel,
const FArrayBox &    a_W,
const int    a_dir,
const Box &    a_box
)  [protected, virtual]
 

Compute the face centered divergence of the velocity.

Compute the face centered divergence of the velocity, a_divVel, from the primitive variables, a_W, in a direction, a_dir, within a_box.

virtual Real PatchGodunov::getMaxWaveSpeed (  const FArrayBox &    a_U,
const Box &    a_box
)  [pure virtual]
 

Compute the maximum wave speed.

virtual PhysIBC* PatchGodunov::getPhysIBC (    )  const [virtual]
 

Get the initial and boundary condition object.

virtual void PatchGodunov::incrementWithSource (  FArrayBox &    a_W,
const FArrayBox &    a_S,
const Real &    a_scale,
const Box &    a_box
)  [protected, virtual]
 

Increment the primitive variables by a source term.

The default implementation does nothing to the primitive variables.

virtual bool PatchGodunov::isDefined (    )  const [protected, virtual]
 

Is the object completely defined.

Return true if the object is completely defined.

virtual bool PatchGodunov::limitSlopes (    )  [virtual]
 

Slope limiter query.

Return true if the application is using slope limiting

virtual PatchGodunov* PatchGodunov::new_patchGodunov (    )  const [pure virtual]
 

Factory method - this object is its own factory.

Return a pointer to new PatchGodunov object with its initial and boundary condtions, slope parameters, and artificial viscosity information defined.

virtual void PatchGodunov::normalPred (  FArrayBox &    a_WMinus,
FArrayBox &    a_WPlus,
const FArrayBox &    a_W,
const FArrayBox &    a_dW,
const Real &    a_scale,
const int &    a_dir,
const Box &    a_box
)  [protected, pure virtual]
 

Extrapolate the primitive variables to the cell faces.

Extrapolate the primitive variables, a_W, in the minus and plus direction, a_dir, using the slopes a_dW within a_box. See document for details.

virtual int PatchGodunov::numConserved (    )  [pure virtual]
 

Number of conserved variables.

Return the number of conserved variables.

virtual int PatchGodunov::numFluxes (    )  [pure virtual]
 

Number of flux variables.

Return the number of flux variables. This can be greater than the number of conserved variables if addition fluxes/face-centered quantities are computed.

virtual int PatchGodunov::numPrimitives (    )  [protected, pure virtual]
 

Number of primitive variables.

Return the number of primitive variables. This may be greater than the number of conserved variables if derived/redundant quantities are also stored for convenience.

virtual int PatchGodunov::numSlopes (    )  [protected, pure virtual]
 

Number of primitive variables for which slopes are computed.

Return the number of primitive variables for which slopes are computed. Only slopes corresponding to primitive variables in the interval 0 to numSlopes() - 1 are computed and only primitive variables in that interval are updated using the slopes.

virtual void PatchGodunov::postUpdateCons (  FArrayBox &    a_U,
const FArrayBox &    a_Uold,
const Real &    a_dt,
const Real &    a_dx,
const Box &    a_box
)  [protected, virtual]
 

virtual int PatchGodunov::pressureIndex (    )  [protected, pure virtual]
 

Component index within the primitive variables of the pressure.

Return the component index withn the primitive variables for the pressure. Used for slope flattening (slope computation).

virtual void PatchGodunov::riemann (  FArrayBox &    a_F,
const FArrayBox &    a_WLeft,
const FArrayBox &    a_WRight,
const int &    a_dir,
const Box &    a_box
)  [protected, pure virtual]
 

Compute a Riemann problem and generate fluxes at the faces.

Given input left and right states in a direction, a_dir, compute a Riemann problem and generate fluxes at the faces within a_box.

virtual void PatchGodunov::setArtificialViscosity (  bool    a_useArtificialViscosity,
Real    a_artificialViscosity
)  [virtual]
 

Set parameters for artificial viscosity.

virtual void PatchGodunov::setCurrentBox (  const Box &    a_currentBox )  [virtual]
 

Set the current box before calling updateState().

virtual void PatchGodunov::setCurrentTime (  const Real &    a_currentTime )  [virtual]
 

Set the current time before calling updateState().

virtual void PatchGodunov::setPhysIBC (  PhysIBC *    a_bc )  [virtual]
 

Set the initial and boundary condition object.

virtual void PatchGodunov::setSlopeParameters (  bool    a_fourthOrderSlopes,
bool    a_flattening,
bool    a_limitSlopes = true
)  [virtual]
 

Set parameters for slope computations.

virtual void PatchGodunov::slope (  FArrayBox &    a_dW,
const FArrayBox &    a_W,
const FArrayBox &    a_flattening,
const int &    a_dir,
const Box &    a_box
)  [protected, virtual]
 

Compute the 2nd or 4th order slopes of the primitive variables.

Compute the limited slope, a_dW, of the primitive variables, a_W, over the range of indices, 0 to numSlopes()-1. This only used a_flattening if 4th order slopes are being computed and slope flattening is turned on. This also calls applyLimiter() to do slope limiting.

virtual Vector<string> PatchGodunov::stateNames (    )  [virtual]
 

Names of the conserved variables.

Return the names of the conserved variables. A default implementation is provided that puts in generic names (i.e., "variable#" which "#" ranges for 0 to numConserved()-1.

virtual void PatchGodunov::updateCons (  FArrayBox &    a_U,
const FArrayBox &    a_F,
const Real &    a_scale,
const int &    a_dir,
const Box &    a_box
)  [protected, pure virtual]
 

Update the conserved variable using fluxes and a scaling factor.

Given the fluxes, a_F, in a direction, a_dir, and a scaling factor, a_scale, update the conserved variables, a_U, within a_box:

a_U = a_U - a_scale * (a_F(a_dir,HiSide) - a_F(a_dir,LoSide))

virtual void PatchGodunov::updatePrim (  FArrayBox &    a_WMinus,
FArrayBox &    a_WPlus,
const FArrayBox &    a_F,
const Real &    a_scale,
const int &    a_dir,
const Box &    a_box
)  [protected, pure virtual]
 

Update the primitive variables using fluxes (of the conserved variables).

Given the fluxes, a_F, in a direction, a_dir, and a scaling factor, a_scale, update the primitive variables, a_WMinus and a_WPlus, within a_box. Note: The fluxes are fluxes of conserved variables.

virtual void PatchGodunov::updateState (  FArrayBox &    a_U,
FArrayBox    a_F[CH_SPACEDIM],
Real &    a_maxWaveSpeed,
const FArrayBox &    a_S,
const Real &    a_dt,
const Box &    a_box
)  [virtual]
 

Update the conserved variables and return the final fluxes used for this.

Update the conserved variables and return the final fluxes that were used for this. Compute the fluxes using a second-order, unsplit Godunov method based on the input conserved variables, a_U, and source terms, a_S. If there are no source terms then a_S should be null constructed. Also return the maximum wave speed.

virtual bool PatchGodunov::useArtificialViscosity (    )  [virtual]
 

Artificial viscosity query.

Return true if the application is using artificial viscosity.

virtual bool PatchGodunov::useFlattening (    )  [virtual]
 

Slope flattening query.

Return true if the application is using slope flattening.

virtual bool PatchGodunov::useFourthOrderSlopes (    )  [virtual]
 

Fourth-order slope query.

Return true if you are using fourth-order slopes and return false if you are using second-order slopes.

virtual Interval PatchGodunov::velocityInterval (    )  [protected, pure virtual]
 

Interval within the primitive variables corresponding to the velocities.

Return the interval of component indices within the primitive variable of the velocities. Used for slope flattening (slope computation) and computing the divergence of the velocity (artificial viscosity).

Member Data Documentation

Real PatchGodunov::m_artificialViscosity [protected]
 

PhysIBC* PatchGodunov::m_bc [protected]
 

Box PatchGodunov::m_currentBox [protected]
 

Real PatchGodunov::m_currentTime [protected]
 

ProblemDomain PatchGodunov::m_domain [protected]
 

Real PatchGodunov::m_dx [protected]
 

bool PatchGodunov::m_isArtViscSet [protected]
 

bool PatchGodunov::m_isBCSet [protected]
 

bool PatchGodunov::m_isCurrentBoxSet [protected]
 

bool PatchGodunov::m_isCurrentTimeSet [protected]
 

bool PatchGodunov::m_isDefined [protected]
 

bool PatchGodunov::m_isSlopeSet [protected]
 

bool PatchGodunov::m_limitSlopes [protected]
 

bool PatchGodunov::m_useArtificialViscosity [protected]
 

bool PatchGodunov::m_useFlattening [protected]
 

bool PatchGodunov::m_useFourthOrderSlopes [protected]
 

The documentation for this class was generated from the following file:
Generated on Wed Apr 16 14:27:44 2003 for Chombo by doxygen1.2.16