DFSR

Divergence-free Spectral Representation (DFSR) method is an inflow/inlet turbulence generation method developed for large-eddy simulation of the atmospheric boundary layer(ABL) flows. The technique uses the spectral representation method to generate ABL-like turbulence with prescribed two-point flow statistics. The code is developed as a utility application using OpenFOAM's framework. DFSR offers the following key features:

Capable of running on unstructured meshes
Uses LAPACK subroutine for efficient matrix factorization
Employes the FFT technique for faster time-series computation
Parallel implementation using OpenMPI

In this version of the code, the divergence-free condition is imposed on the generated turbulence by adding a gradient of a velocity potential, following the procedure developed by Shirani et al.(1981)(similar to approach used in the paper). The final velocity field is written for each time step in the corresponding case directory. Mathematical details of the implemented procedure can be found in Melaku and Bitsuamlak(2021). If you find this work useful in your research and use the DFSR method or parts of it in your work, please do not forget to cite the paper.

Installing DFSR

Requirements

Before compiling the source code, you need to have the following packages already installed.

OpenFOAM, preferably v8
Intel's oneAPI MKL package or LAPACK library

Download the installer here: https://www.intel.com/content/www/us/en/developer/tools/oneapi/base-toolkit-download.html?operatingsystem=linux&distributions=webdownload&options=online

Follow the Command Line Download and Installation instructions (sudo is not required, but we only selected the MKL package to install - 7.3 GB)
And make sure to follow these instructions: https://www.intel.com/content/www/us/en/develop/documentation/get-started-with-dpcpp-compiler/top.html
i.e. source the environment settings script (in our case that worked with $ source $HOME/intel/oneapi/setvars.sh intel64 but might be dependent on where you installed it - also, you might want to add this line to the end of your ~/.bashrc file, as it needs to be sourced every time you open a new terminal window)
Fastest Fourier Transform in the West(FFTW) package

Download the source code from here: wget https://www.fftw.org/fftw-3.3.10.tar.gz and in the un-zipped folder run the following:

$ ./configure
$ make
$ sudo make install

Build from source

Follow the following procedure to compile the source code.

Clone the code to your computer:

$ git clone https://github.com/abiyfantaye/DFSR.git DFSR
Copy the DFSR/src/DFSRTurb directory to $WM_PROJECT_USER_DIR/applications/utilities/preProcessing/:

$ mkdir -p $WM_PROJECT_USER_DIR/applications/utilities/preProcessing
$ cp -r DFSR/src/DFSRTurb $WM_PROJECT_USER_DIR/applications/utilities/preProcessing/
Compile the code in $WM_PROJECT_USER_DIR/applications/utilities/preProcessing/DFSRTurb directory:

$ cd $WM_PROJECT_USER_DIR/applications/utilities/preProcessing/DFSRTurb
$ chmod +x Allwclean
$ chmod +x Allwmake
$ sh ./Allwclean
$ sh ./Allwmake

The turbulence generator is now fully embedded in OpenFOAM and can be launched as DFSRTurb in serial or parallel. If you have trouble installing this application, please drop an email to afmelaku@berkeley.edu cc·'ing to Prof. Girma Bitsuamlak at gbitsuam@uwo.ca.

Simulation Example

Directory tree of example/openTerrain:

├── 0
│   ├── nut
│   ├── p
│   └── U
├── constant
│   ├── boundaryData
│   │   └── windProfile
│   │       ├── profile
│   │       └── sampledData
│   │           ├── samplingPoints
│   │           └── vtkSlices
│   ├── DFSRTurbDict
│   ├── transportProperties
│   └── turbulenceProperties
├── setUp
└── system
    ├── blockMeshDict
    ├── controlDict
    ├── decomposeParDict
    ├── fvSchemes
    ├── fvSolution
    ├── probes1
    └── probes2

Input parameters

The directory example/openTerrain contains an empty domain LES case configured for open terrain conditions. The basic parameters for the DFSR method are defined in constant/DFSRTurbDict. The method reads the ABL profile from constant/boundaryData/windProfile/profile file. This file specifies the variation of the mean velocity, turbulence intensities, and length scales over the height in the following tabular format:

z[m]	Uav[m/s]	Iu	Iv	Iw	Lu [m]	Lv[m]	Lw[m]
0.02	14.56	0.160	0.120	0.120	0.750	0.500	0.300
.....	.......	.....	.....	.....	.....	.....	.....
.....	.......	.....	.....	.....	.....	.....	.....

The parameters that need to be defined in constant/DFSRTurbDict dictionary are:

Parameter	Description	Optional	Default
`patchName`	Name of the patch to apply the inflow	no	--
`startTime`	Starting time of the generated turbulence	yes	0.0
`endTime`	Ending time of the generated turbulence	no	--
`timeStep`	Time step of the inflow, can be different from the solver time step. It should be `1/(2*fMax)` to take advantage of the FFT algorithm.	no	--
`fMax`	The cut-off frequency	no	--
`nFreq`	The number of frequency steps/segments	no	1024
`nInterp`	Number of spectral interpolation frequencies	no	50
`seed`	Seed of the random number generator used to sample random phase angles.	yes	--
`C`	Coherency decay coefficients.	no	--
`cohUav`	Mean velocity used to define the coherency function	yes	--
`scaleI`	A factor to scale turbulence intensity profiles in each direction to account for the decay of turbulence downstream in the CFD simulation.	yes	(1 1 1)
`scaleL`	A factor to scale integral length scale profiles in each direction.	yes	(1 1 1)
`correctDivergence`	Apply div-free corrections or not	yes	off
`writeSamples`	Write data for selected sampling points or not	yes	on
`writeInflow`	Write the final inflow data or not	yes	on

Output

Finally, if the controle writeInflow is turned on, the whole inflow data is saved in constant/boundaryData/<patchName>/ directory. This inflow data is later read by the solver with timeVaryingMappedFixedValue velocity boundary condition at the inlet. Please check the OpenFOAM test case in example/openTerrain directory for the details.

If writeSamples is turned on, the code reads sampling points from constant/boundaryData/windProfile/sampledData/samplingPoints. Then, the time-series of the generated velocity components for each point in samplingPoints are written into Ux,Uy, and Uz files in the same directory. In addition, the velocity field over the inflow patch is written to VTK slices for selected time instances in the same directory.

Excuting the test case

Follow the following procedure to run the test case in example/openTerrain directory:

Generate block mesh:

$ blockMesh
Decompose the case:

$ decomposePar
Run the DFSR inflow generator in parallel:

$ mpirun -np 4 DFSRTurb -parallel
Check the generated samples on paraview:

$ paraview constant/boundaryData/windProfile/sampledData/vtkSlices/Ut_*
Finally run the LES case:

$ mpirun -np 4 pisoFoam -parallel

Reference

[1] Melaku, A.F. and Bitsuamlak, G.T., 2021. A divergence-free inflow turbulence generator using spectral representation method for large-eddy simulation of ABL flows. Journal of Wind Engineering and Industrial Aerodynamics, 212, p.104580.

[2] Shirani, E., Ferziger, J.H. and Reynolds, W.C., 1981. Mixing of a passive scalar in isotropic and sheared homogeneous turbulence. Mosciences Division, Department of Mechanical Engineering, Stanford University.

abiyfantaye / DFSR