Digital-filter based boundary condition for velocity, i.e. U, to generate synthetic turbulence-alike time-series for LES and DES turbulent flow computations from input turbulence statistics. More...

Detailed Description

Digital-filter based boundary condition for velocity, i.e. U, to generate synthetic turbulence-alike time-series for LES and DES turbulent flow computations from input turbulence statistics.

References:

    Digital-filter method-based generator (DFM) (tag:KSJ):
        Klein, M., Sadiki, A., & Janicka, J. (2003).
        A digital filter based generation of inflow data for spatially
        developing direct numerical or large eddy simulations.
        Journal of computational Physics, 186(2), 652-665.
        DOI:10.1016/S0021-9991(03)00090-1

    Forward-stepwise method-based generator (FSM) (tag:XC)
        Xie, Z. T., & Castro, I. P. (2008).
        Efficient generation of inflow conditions for
        large eddy simulation of street-scale flows.
        Flow, turbulence and combustion, 81(3), 449-470.
        DOI:10.1007/s10494-008-9151-5

    Mass-inflow rate correction (tag:KCX):
        Kim, Y., Castro, I. P., & Xie, Z. T. (2013).
        Divergence-free turbulence inflow conditions for
        large-eddy simulations with incompressible flow solvers.
        Computers & Fluids, 84, 56-68.
        DOI:10.1016/j.compfluid.2013.06.001

In DFM or FSM, a random number set (mostly white noise), and a group of target statistics (mostly mean flow, Reynolds stress tensor profiles and length-scale sets) are merged into a new number set (stochastic time-series, yet consisting of the statistics) by a chain of mathematical operations whose characteristics are designated by the target statistics, so that the realised statistics of the new sets could match the target.

Random number sets ---->-|
                         |
                     DFM or FSM ---> New stochastic time-series consisting
                         |           turbulence statistics
Turbulence statistics ->-|

The main difference between DFM and FSM is that FSM replaces the expensive-to-run streamwise convolution summation in DFM by a simpler and an almost-equivalent-in-effect numerical procedure in order to reduce computational costs. Accordingly, FSM potentially brings computational resource advantages for computations involving relatively large streamwise length-scale sets and small time-steps.

Synthetic turbulence is generated on a virtual rectangular structured-mesh plane, which is parallel to the chosen patch, and is mapped onto this patch by the selected mapping method.

Usage

Example of the boundary condition specification:

<patchName>
{
    // Mandatory entries (unmodifiable)
    type                turbulentDigitalFilterInlet;
    n                   (<nHeight> <nWidth>);
    L                   (<L1> <L2> ... <L9>);
    R                   uniform (<Rxx> <Rxy> <Rxz> <Ryy> <Ryz> <Rzz>);
    UMean               uniform (1 0 0);
    Ubulk               10.0;

    // Optional entries (unmodifiable)
    fsm                 false;
    Gaussian            true;     // always false for FSM
    fixSeed             true;
    continuous          false;
    correctFlowRate     true;
    mapMethod           nearestCell;
    perturb             1e-5;
    C1                  -1.5707;  //-0.5*PI;
    C1FSM               -0.7854   //-0.25*PI;
    C2FSM               -1.5707;  //-0.5*PI;

    // Optional (inherited) entries
    ...
}

where the entries mean:

Property	Description	Type	Req'd	Dflt
`type`	Type name: turbulentDigitalFilterInlet	word	yes	-
`n`	Number of cells on turbulence generation plane	tuple of labels	yes	-
`L`	Integral length-scale set (Lxu Lxv Lxw Lyu Lyv Lyw Lzu Lzv Lzw) [m]	tensor	yes	-
`R`	Reynolds stress tensor set (xx xy xz yy yz zz) [m2/s2]	symmTensorField	yes	-
`UMean`	Mean velocity profile [m/s]	vectorField	yes	-
`Ubulk`	Characteristic patch-normal bulk flow speed [m/s]	scalar	yes	-
`fsm`	Flag to turn on the forward-stepwise method	bool	no	false
`Gaussian`	Autocorrelation function form	bool	no	true
`fixSeed`	Flag to fix random-number generator seed to 1234 or generate a new seed based on clock-time per simulation	bool	no	true
`continuous`	Flag to write random-number sets at output time, and to read them on restart. Otherwise, generate new random-number sets of restart	bool	no false
`correctFlowRate`	Flag to correct mass-inflow rate on turbulence plane in (only) streamwise direction	bool	no	true
`mapMethod`	Interpolation-to-patch method	word	no	nearestCell
`perturb`	Point perturbation for planarInterpolation mapMethod	scalar	no	1e-5
`C1`	Model constant shaping autocorrelation function (KSJ:Eq. 14)	scalar	no	-0.5*PI
`C1FSM`	Model coefficient in FSM (XC:Eq. 14)	scalar	no	-0.25*PI
`C2FSM`	Model coefficient in FSM (XC:Eq. 14)	scalar	no	-0.5*PI

The inherited entries are elaborated in:

fixedValueFvPatchFields.H

Options for the fsm entry:

      false | Method due to (KSJ)
      true  | Method due to (XC)

Options for the Gaussian entry:

      true  | Gaussian function
      false | Exponential function (only option for FSM)

Options for the mapMethod entry:

      nearestCell         | One-to-one direct map, no interpolation
      planarInterpolation | Bilinear interpolation

Patch-profile input is available for two entries:

      R     | Reynolds stress tensor
      UMean | Mean velocity

where the input profiles and profile coordinates are located in:

      Coordinates | $FOAM_CASE/constant/boundaryData/\<patchName\>/points
      R/UMean     | $FOAM_CASE/constant/boundaryData/\<patchName\>/0/\{R/UMean\}

points file contains a list of three-dimensional coordinates, and profile data files provide a value corresponding to each coordinate.

Note

mapMethod=planarInterpolation option needs point coordinates that can form a plane.
adjustTimeStep=true option is currently not fully supported.
In order to obtain Reynolds stress tensor information, experiments, RANS simulations or engineering relations can be used.
continuous=true means deterministic-statistical consistent restart (relatively more expensive), and continuous=false means deterministic discontinuity in synthetic turbulence time-series by keeping statistical consistency (relatively cheaper).
For L, the first three entries should always correspond to the length scales in association with the convective (streamwise) mean flow direction.
Streamwise integral length scales are converted to integral time scales by using Taylor's frozen turbulence hypothesis, and Ubulk.

See also

turbulentDFSEMInletFvPatchVectorField.C

Source files

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

src/finiteVolume/fields/fvPatchFields/derived/turbulentDigitalFilterInlet/turbulentDigitalFilterInletFvPatchVectorField.H