Public Member Functions | List of all members
DRGEP< CompType, ThermoType > Class Template Reference

The DRGEP algorithm [1] is based on. More...

Inheritance diagram for DRGEP< CompType, ThermoType >:
Inheritance graph
[legend]
Collaboration diagram for DRGEP< CompType, ThermoType >:
Collaboration graph
[legend]

Public Member Functions

 TypeName ("DRGEP")
 
 DRGEP (const IOdictionary &dict, TDACChemistryModel< CompType, ThermoType > &chemistry)
 
virtual ~DRGEP ()
 
virtual void reduceMechanism (const scalarField &c, const scalar T, const scalar p)
 
- Public Member Functions inherited from chemistryReductionMethod< CompType, ThermoType >
 TypeName ("chemistryReductionMethod")
 
 declareRunTimeSelectionTable (autoPtr, chemistryReductionMethod, dictionary,(const IOdictionary &dict, TDACChemistryModel< CompType, ThermoType > &chemistry),(dict, chemistry))
 
 chemistryReductionMethod (const IOdictionary &dict, TDACChemistryModel< CompType, ThermoType > &chemistry)
 
virtual ~chemistryReductionMethod ()
 
bool active () const
 
bool log () const
 
const List< bool > & activeSpecies () const
 
label NsSimp ()
 
label nSpecie ()
 
scalar tolerance () const
 

Additional Inherited Members

- Static Public Member Functions inherited from chemistryReductionMethod< CompType, ThermoType >
static autoPtr< chemistryReductionMethod< CompType, ThermoType > > New (const IOdictionary &dict, TDACChemistryModel< CompType, ThermoType > &chemistry)
 
- Protected Attributes inherited from chemistryReductionMethod< CompType, ThermoType >
const IOdictionarydict_
 
const dictionary coeffsDict_
 
Switch active_
 
Switch log_
 
TDACChemistryModel< CompType, ThermoType > & chemistry_
 
List< boolactiveSpecies_
 
label NsSimp_
 
const label nSpecie_
 
scalar tolerance_
 

Detailed Description

template<class CompType, class ThermoType>
class Foam::chemistryReductionMethods::DRGEP< CompType, ThermoType >

The DRGEP algorithm [1] is based on.

|sum_i=1->Nr vAi wi dBi| rAB = ————————— , max(PA, CA)

PA = sum_i=1->Nr (max (0, vAi wi)) -> production of species A

CA = sum_i=1->Nr (max (0, -vAi wi)) -> consumption of species A

where i is the reaction index, Nr the number of reactions, vAi is the net stoechiometric coefficient of species A in the ith reaction (vAi = v''-v') , wi is the progress variable of reaction i and dBi equals 1 if reaction i involves B and O otherwise. rAB show the error introduced to the production rates of A when B and all the reactions including it are removed. It is computed as in [2] so that the algorithm is O(Nr).

DAC uses a initial set of species that represents the major parts of the combustion mechanism, i.e. H2/O2, fuel decomposition and CO2 production. Usually, it includes the fuel, HO2 and CO. Then it computes the dependence of these set to the other species. This is done by introducing R-value defined by

R_V0 (V) = max_SP(product(rij)) ,

where SP is the set of all possible paths leading from V0 to V and product(rij) is the chain product of the weights of the edges along the given path. The R-value for the initial set species is 1.

When the R-value of a species is larger than a user-defined tolerance then the species is included in the simplified mechanism. Otherwise, the species is removed along with all the reactions including it.

During this process, instead of looking over all species like described in [1], the algorithm implemented here creates dynamic list to retain the initialized edges only (see [2]).

To avoid using the target species when they are not contributing yet or anymore to the system, a coefficient based on the exchange of element is introduced:

NTa |PT - CT| alphaTa = —————- Pa

Pa = sum_speciesS NSa max(0, PS-CS)

where 'a' refers to different elements present in the system (namely C, H, O and N for conventionail hydrocarbon combustion), NTa is the number of element a in species T. When this coefficient alpha is below the specified threshold, the species is removed from the search initiating set. In the original paper from Pepiot-Desjardins et al.[2], this coefficient is further transformed to compute a global normalized scaling coefficient but here as it is dynamically computed, alpha is not introduced in the calculation of R.

References:

    [1] Pepiot-Desjardins, P., & Pitsch, H. (2008).
    An efficient error-propagation-based reduction method for large
    chemical kinetic mechanisms.
    Combustion and Flame, 154(1), 67-81.

    [2] Lu, T., & Law, C. K. (2006).
    Linear time reduction of large kinetic mechanisms with directed
    relation graph: n-Heptane and iso-octane.
    Combustion and Flame, 144(1), 24-36.
Source files

Definition at line 117 of file DRGEP.H.

Constructor & Destructor Documentation

◆ DRGEP()

DRGEP ( const IOdictionary dict,
TDACChemistryModel< CompType, ThermoType > &  chemistry 
)

Definition at line 28 of file DRGEP.C.

References chemistry, Foam::endl(), Foam::exit(), Foam::FatalError, FatalErrorInFunction, dictionary::found(), Foam::Info, and dictionary::subDict().

Here is the call graph for this function:

◆ ~DRGEP()

~DRGEP
virtual

Definition at line 99 of file DRGEP.C.

Member Function Documentation

◆ TypeName()

TypeName ( "DRGEP< CompType, ThermoType >"  )

◆ reduceMechanism()

void reduceMechanism ( const scalarField c,
const scalar  T,
const scalar  p 
)
virtual

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