Configuration files

Configuration files#

The input parameters are loaded from the three following .json files:

  • settings.json: contains most of the information about the FELiCS run.

  • boundaries.json: contains specific information about the boundary conditions.

  • mixture.json(optional): contains specific information about the physical properties of the fluid considered.

Structure of the settings.json file#

The typical settings.json file is divided into 6 main sections:

{
"BoundaryCondition":{
    "BCsFilePath":      (str)   path to the "boundaries.json" file,
},
"Case":{
    "AnalysisMode":     (str)   type of analysis, one of: ("Modal", "Resolvent", "InputOutput"),
    "CalculateAdjoint": (bool)  if we compute the adjoint spectrum in "Modal" analysis,
    "CoordinateSystem": (str)   coordinate system, one of: ("Cylindrical", "Cartesian"),
    "m":                (float) wavenumber of fluctuations in the third, spectral dimension (for a 2D grid),
    "MeshFilePath":     (str)   path to the mesh file.
    "MixtureFilePath":  (str)   path to the "mixture.json" file.
    "MolVisc":          (float) kinematic viscosity (dynamic viscosity in compressible flows) value when "MolViscModel" is set to "Constant"
    "MolViscModel":     (str)   type of laminar viscosity model used for the viscosity,
    "MolViscPerturbModel": (dict)   type of laminar viscosity model used for fluctuating laminar viscosity, specifying [type, empirical constants], e.g.:
        {
            "type":"Sutherland mean",
            "Constants": {"Ts":170.672}
        },
    "nDim":             (int)   number of dimensions resolved in the mesh.
    "PrandtlNumber":    (int)   spatially global Prandtl number, only relevant for cases where thermal diffusion or conduction is involved
    "Reaction":         (bool)  if using chemistry reactions (DEPRECATED? => DEFINED IN THE MIXTURE FILE?)
    "SetOfEquations":{  (dict)  with fields: [equations, variants, transported variables], all defined as str. e.g.:  
        "Momentum":{
            "Equation":"NSPrimitive",
            "Variable":"u"
            }
        },
    "SpeciesFilePath":  (str)   path to the SpeciesFile (DEPRECATED?),
    "TransVelFluc":     (bool)  if fluctuations are defined in the Fourier dimension (DUPLICATED with "m"?)
    "TurbulenceModel":  (str)   type of turbulence model (currently "None", "Constant", or "File")
},
"Export":{
    "ExportFolder":     (str)   directory (relative to running dir) where outputs are saved,
    "Video":            (bool)  if xdmf files used to generate a video of the mode are exported, 
},
"FlowInput":{
    "AveragingDirection":(str)  dimension in which the mean flow is averaged, set to "None" if unused (DEPRECATED?)
    "MeanFlowFilePath": (str)   mean flow file path and name to load.
},
"IOResolvent":{
    "ForcingBoundaryIndices": (list of int) indices of boundaries where forcing is applied ("InputOutput" analysis),
    "ForcingCoeff":     (list of int)   variables onto which forcing is applied ("InputOutput" analysis),
    "ForcingMode":      (str)   type of forcing "Body" or "Boundary" ("InputOutput" analysis)
    "ForcingNorm":      (str)   norm type for the forcing term ("Resolvent" analysis), one of: ("TKE" - default, "Chu")
    "ResponseNorm":     (str)   norm type for the response term ("Resolvent" analysis), one of ("TKE" - default, "Chu")
    "Omegas":           (list)  angular frequency (see formatting in previous section),
},
"Numerics":{
    "EigenValueGuess":  (list)  eigenvalue guesses for "Modal" analysis (see formatting in previous section),
    "nSolut":           (int)   number of solutions to compute (number of nearest eigenvalues in  "Modal" analysis, number of resolvent modes per frequency in "Resolvent" analysis)
    "PolynomialOrder":  (dict)  sets the polynomial order for each transported variables (list must match "SetOfEquations"). e.g.: {"u": 2,"T": 1,"rho": 1}
}}

The settings.json is directly given to FELiCS when running from the command line via the -file flag:

FELiCS -file settings.json

Structure of the boundaries.json file#

The structure of this file is:

{
"ID1":{            (int as a str)    ID of the boundary. For gmsh *.msh* meshes, this must correspond to the index of an existing *PhysicalNames* entry.
    "name":        (str)             name of boundary type; one of: ("zeroDirichlet", "wall", "custom", "symmetry","none")
    "specifics":{  (dict)            only for boundary types "custom" and "symmetry"
        "variable":(str)             the name of the variable considered; has to be one of the state vector varaibles, defined in the "SetOfEquations" part in the general config.json file
        "type"    :(str)             type of boundary conditions, one of: ("Dirichlet", "Neumann", "None")
        "value"   :(float)           value imposed on the variable (or its gradient)
}}}

Here is an expamle for a boundaries.json file:

{
    "1": {
        "name": "zeroDirichlet"
    },
    "2": {
        "name": "symmetry",
        "specifics": [
            {
                "variable": "ux",
                "type": "Dirichlet",
                "value": 0.0
            },
            {
                "variable": "uy",
                "type": "Neumann",
                "value": 0.0
            },
            {
                "variable": "p",
                "type": "Dirichlet",
                "value": 0.0
            }
        ]
    },
    "3": {
        "name": "wall"
    }
}

Note: The name of the file is not important. The name of this json file should be set in the main settings file via the variable “BCsFilePath”.

Structure of the mixture.json file#

**TODO: ** review (Thomas?)

In most cases that do not involve chemistry modelling or species transport, the mixture.json is not needed.

Below is the structure of the mixture.json file for a case using chemistry:

{
"Species":                        (dict) list of species that are considered, providing how they are calculated and additional species properties, e.g.:
    { 
        "phi":{                   (dict) name of species
            "calc":"transported", (str) specify species that are returned, "all": all species, "transported": only species with separate transport equations, "constraint": only passive species without transport equations,
            "Sc":0.9              (float) Schmidt number of the species
        }
    },
"Reaction_mechanism": (dict) list with type of reaction model and all required settings, e.g.:
{
    "type": "KaiserCnF2023", (str) type of reaction model (currently only "KaiserCnF2023" is implemented)
    "additional_fields":["prefactor"], (str) additional fields to be read in from the input *.fel file 
    "reactions":[{ (dict) list of reactions that provides all involved educts and products and their corresponding stoichiometric coefficients
        "educts": [],
        "stochiometricCoefficientsEducts": [],
        "products": ["progress"],  
        "stochiometricCoefficientsProducts": [1.0]
    }]
}}

Specific formatting for FELiCS settings and json#

  • booleans are defined in lowercase (true, false)

  • EigenValueGuess and Omegas can be either

    • list of strings for complex values: ["1.0-1j", "1.0+1j"]

    • list of floats for real values: [1, 2, 1.2e-1]

    • a combination of both: ["1.0-1j", 1, "1.0+1j", 2]

  • json accepts exponential notation for all float-type inputs

Warning: only the parameters listed by config.getAllSettingsDict() in the src/FELiCS/parameters/config.py file will be considered by FELiCS. Add your new parameters there to be able to use them in the code.

Contents