Options

This page explains the arguments of each class that can be defined in prob.h. Most of these options are passed as an argument in set-up through a light structure. All arguments are defined as constexpr so they are known and evaluated at compile time. This helps optimize the code (at compile time expense). A quick tip on checking the available options is to look in defaultparm_t in NumParam.h where all options are defined to avoid name conflicts.

closures_dt

transport_const_t

Example use is

typedef closures_dt<indicies_t, transport_const_t<viscparm_t>, ...  > ProbClosures;

where specific values of conductivity and viscosity are selected

struct viscparm_t {

  public:

  static constexpr Real conductivity = 0.001;      // conductivity (for constant value)
  static constexpr Real viscosity    = 1e-5;       // viscosity    (for constant value)
};

Smagorinsky_t / WALE_t

Both Smagorinsky and WALE are sub-classes of LES and have the same parameters.

Example use is

typedef closures_dt< ... , Smagorinsky_t<LESparm_t,indicies_t> > ProbClosures;

with LESparm_t containing all parameters required for LES models. Some may not be used if specific models are used, but ALL have to be defined if one od these templates are used.

struct LESparm_t {

  public:

  static constexpr int  order = 2;         
  static constexpr Real Pr_o_Prsgs = 0.8;  
  static constexpr Real Scsgs = 0.7;       
  static constexpr Real Cs = 0.18;          
  static constexpr Real CI = 0.08;        
  static constexpr bool fixDelta = false;  
  static constexpr Real Delta = 0.02;      
};

Variable Name	Type	Value	Description
order	int	2	Order of the numerical scheme used for gradient estimation.
Pr_o_Prsgs	Real	0.8	Ratio of molecular Prandtl number (Pr) to subgrid-scale Prandtl number (Prsgs).
Scsgs	Real	0.7	Subgrid-scale Schmidt number (Scsgs).
Cs	Real	0.18	Smagorinsky constant; used in sgs models.
CI	Real	0.08	Yoshizawa constant; associated with isotropic sgs contibution.
fixDelta	bool	false	Flag to indicate whether the filter width is fixed.
Delta	Real	0.02	Filter width (only used if fixDelta is true)

rhs_dt

skew_t

Example use is

typedef rhs_dt<skew_t<methodparm_t, ProbClosures>,  .... > ProbRHS;

with

struct methodparm_t {

  public:

  static constexpr bool dissipation = true;         
  static constexpr int  order = 4;                  
  static constexpr Real C2skew=0.1,C4skew=0.016;   

};

Variable Name	Type	Value	Description
dissipation	bool	true	Flag indicating whether dissipation is applied (true = yes).
order	int	4	Order of the numerical scheme used 2/4/6.
C2skew	Real	0.1	Coefficient for 2nd-order skew-symmetric dissipation.
C4skew	Real	0.016	Coefficient for 4th-order skew-symmetric dissipation.

viscous_t

Example use is

typedef rhs_dt<skew_t<... viscous_t<methodparm_t, ProbClosures> ..  > ProbRHS;

struct methodparm_t {

  public:

  static constexpr int  order = 2;                  // order numerical scheme
  static constexpr bool use_LES= false;             // LES model 
};

Variable Name	Type	Value	Description
order	int	2	Order of the numerical scheme used 2/4/6.
use_LES	bool	false	Flag indicating sgs model used (true =yes).

If use_LES = true, sub-grid viscosty/conductivity/ect.. will be added to the viscosity

IBM

Example use is

typedef eib_t<TypeWall,ibmparm_t,ProbClosures> ProbIB;

struct ibmparm_t {

  public:

  static constexpr int  interp_order = 1;
  static constexpr int  extrap_order = 1;
  static constexpr Real alpha= 0.6;      
};

This controls IBM options, and it is needed to define stencils for extrapolation. Order is interp_order+1, in the above example the interpolation order is 2

Variable Name	Type	Value	Description
interp_order	int	1	Order of the interpolation scheme -1 used in IBM.
extrap_order	int	1	Order of the extrapolation scheme -1 used in IBM.
alpha	Real	0.6	Ratio between first IP and IB points (in mesh units)

Default values

The default options are (any struct can be replace by defaultparm_t)

// default values 
struct defaultparm_t {

public:


static constexpr int order = 2;                   // order of scheme (used in many options)

static constexpr bool dissipation = false;        // use in some schemes to activate dissipation

#ifdef AMREX_USE_GPIBM
static constexpr bool ibm = true;
#else
static constexpr bool ibm = false;
#endif

//  Skew 
static constexpr Real C2skew=0.1,C4skew=0.0016;   // Skew symmetric default

// Transport properties
static constexpr Real conductivity = 0.0262;      // conductivity (for constant value)
static constexpr Real viscosity   = 1.85e-5;      // viscosity    (for constant value)

// ebm/walls options
static constexpr Real Twall = 300;              // wall temperature (for isothermal wall type)
static constexpr bool solve_diffwall = false;   // do not solve viscous fluxes at teh wall (for ebm)

// viscous options 
static constexpr bool use_LES = false;

// LES options
static constexpr Real Pr_o_Prsgs = 1.0;          // Pr/Prsgs
static constexpr Real Scsgs = 0.7;              // sgs Schmidt number
static constexpr Real Cs = 0.1;                 // Smagorinsky constant
static constexpr Real CI = 0.1;                 // Yoshizawa constant
static constexpr bool fixDelta = false;         // (it will use local mesh size as filter width)
static constexpr Real Delta = 0.01;             // Filter width  

};