This repository implements a discontinuous Galerkin finite element method (DGFEM) applied to the linearized Euler equations and the acoustic perturbation equations. The solver is based on GMSH library and supports a wide range of features:
- 1D, 2D, 3D problems
- 4-th order Runge-Kutta
- High order elements
- Absorbing and reflecting boundaries
- Support 'json' format configartion file
- Multiple sources support: monopoles, dipoles, quadrupoles, user defined analytical formulation sources and external data (csv and sound 'wave' file supported)
- Complex geometry and unstructured grid (only triangles (2D) and tetrahedrons (3D) elements are supported)
- VTK post-processing (use Paraview)
- User defined obervers position post-processing (text data time variables, Fast Fourier Transform, Pressure Power Spectral Density and sound 'wave' files)
First, make sure the following libraries are installed. If you are running a linux distribution (ubuntu, debian, ...), an installation script is provided.
- Gmsh (v4.13.x)
- Eigen (v3.x)
- Lapack
- Blas
- OpenMP
Libtbb- VTK (v9.x)
- FFTW
git clone https://github.com/skhelladi/DGFEM-CAA.git
cd DGFEM-CAA
sh build.sh
or use
git clone https://github.com/skhelladi/DGFEM-CAA.git
cd DGFEM-CAA
mkdir build && cd build
cmake ../ -DCMAKE_BUILD_TYPE=Release -G "Unix Makefiles" -DGMSH_INCLUDE_DIRS="../gmsh-4.11.0-Linux64-sdk/include" -DGMSH_LIBRARIES="../gmsh-4.11.0-Linux64-sdk/lib/libgmsh.so" -DGMSH_EXECUTABLE="../gmsh-4.11.0-Linux64-sdk/bin/gmsh" -DEIGEN_INCLUDE_DIRS="/usr/include/eigen3"
make -j4
Once the sources sucessfully build, you can start using with the solver. It required two arguments: a mesh file created with Gmsh and a config file containing the solver options. Examples of mesh files and config files are given here.
cd bin
./dgalerkin myconfig.conf
or
cd bin
./dgalerkin myconfig.json
2D propagation of an Gaussian initial condition over a square.
cd build/bin
./dgalerkin ../../doc/config/Square.conf
or
cd build/bin
./dgalerkin ../../doc/config/Square.json
or configure run_caa batch file with the right mesh and configurations files.
sh run_caa
meshFileName = doc/2d/square2.msh
# Initial, final time and time step(t>0)
timeStart=0
timeEnd=0.05
timeStep=0.00001
# Saving rate:
timeRate=0.001
# Element Type:
# ["Lagrange", "IsoParametric", ...]
elementType=Lagrange
# Time integration method:
# ["Euler1", "Euler2", "Runge-Kutta"...]
timeIntMethod=Runge-Kutta
# Boundary condition:
# /!\ The physical group name must match the Gmsh name (case sensitive)
# MyPhysicalName = Absorbing or Reflecting
Reflecting = Reflecting
Absorbing = Absorbing
# Number of thread
numThreads=12
# Mean Flow parameters
v0_x = -30
v0_y = 30
v0_z = 0
rho0 = 1.225
c0 = 100
# Source:
# name = fct,x,y,z,size,intensity,frequency,phase,duration
# - fct supported = [monopole, dipole, quadrupole, formula, file (csv, wav)]
# - if fct = formula => name = fct,"formula expr",x,y,z,size,duration (ex: formulat = 0.1 * sin(2 * pi * 50 * t))
# - if fct = file => name = fct,"filename",x,y,z,size
# suported file formats are : csv, wav
# - (x,y,z) = source position
# - intensity = source intensity
# - frequency = source frequency
# NB: Extended source or Multiple sources are supported.
# (source1 = ..., source2 = ...) indice must change.
source1 = formula, "0.1 * sin(2 * pi * 50 * t)", 0.0,0.0,0.0, 0.1, 0.1
source2 = monopole, 0.0,0.0,0.0, 0.1, 0.1,50,0,0.1
source3 = file,"data/data.csv", 0.0,0.0,0.0, 0.1
source4 = file,"data/data.wav", 0.0,0.0,0.0, 0.1
# source4 = udf, "-0.1 * sin(2 * pi * 50 * t)", -0.5,0.0,0.0, 0.1, 0.1
# Initial condition:
# name = gaussian,x,y,z, size, amplitude
# - fct supported = [gaussian]
# - (x,y,z) = position
# - amplitude = initial amplitude
# NB: Multiple CI are supported and recursively added.
# (initial condition1 = ..., initial condition = ...)
# initialCondtition1 = gaussian, 0,0,0,1,1
# Observers position:
# name = x,y,z, size
# - (x,y,z) = position
# NB: Multiple observers are supported and recursively added.
# (observer1 = ...; observer2 = ...)
observer1 = 2.11792,0.00340081,0.0,0.1
observer2 = -2.11792,0.00340081,0.0,0.1{
"mesh": {
"File": "doc/2d/square2.msh",
"BC": {
"number": 2,
"boundary1": {
"name": "Abs",
"type": "Absorbing"
},
"boundary2": {
"name": "Ref",
"type": "Reflecting"
}
}
},
"solver": {
"time": {
"start": 0.0,
"end": 0.05,
"step": 5e-05,
"rate": 0.001
},
"elementType": "Lagrange",
"timeIntMethod": "Runge-Kutta",
"numThreads": 12
},
"initialization": {
"meanFlow": {
"vx": 30.0,
"vy": 0.0,
"vz": 0.0,
"rho": 1.225,
"c": 100.0
},
"number": 2,
"initialCondition1": {
"type": "gaussian",
"x": 0.2,
"y": 0.0,
"z": 0.0,
"size": 1.0,
"amplitude": 1.0
},
"initialCondition2": {
"type": "gaussian",
"x": -0.2,
"y": 0.0,
"z": 0.0,
"size": 1.0,
"amplitude": 1.0
}
},
"observers": {
"number": 2,
"observer1": {
"x": 2.11792,
"y": 0.00340081,
"z": 0.0,
"size": 0.1
},
"observer2": {
"x": -2.11792,
"y": 0.00340081,
"z": 0.0,
"size": 0.1
}
},
"sources": {
"number": 3,
"source1": {
"type": "formula",
"fct": "0.1 * sin(2 * pi * 50 * t)",
"x": 0.0,
"y": 0.0,
"z": 0.0,
"size": 0.1,
"amplitude": 0.0,
"frequency": 0.0,
"phase": 0.0,
"duration": 0.05
},
"source2": {
"type": "file",
"fct": "data/data2.wav",
"x": 0.0,
"y": 0.0,
"z": 0.0,
"size": 0.1,
"amplitude": 0.0,
"frequency": 0.0,
"phase": 0.0,
"duration": 0.05
},
"source3": {
"type": "monopole",
"fct": "",
"x": 0.0,
"y": 0.0,
"z": 0.0,
"size": 0.1,
"amplitude": 0.1,
"frequency": 50.0,
"phase": 0.0,
"duration": 0.05
}
}
}This project is licensed under the GPL-3 license.
- Sofiane Khelladi
- Pierre-Olivier Vanberg
- Martin Lacroix
- Tom Servais