cbns3d_cart_mpi

This repository hosts a High-Order 3D Compressible Navier-Stokes Solver. The solver utilizes high-order finite difference schemes to discretize and solve the 3D compressible Navier-Stokes equations on Cartesian grids. Leveraging MPI, the code enables parallel execution across multiple cores or processors, enhancing computational efficiency.

Taylor-Green Vortex Test Case

Features:

• High-Order Accuracy: Employs a 5th-order WENO scheme for inviscid fluxes, coupled with a 4th-order central finite difference scheme for viscous fluxes. Time integration utilizes the robust and efficient 3rd-order TVD Runge-Kutta method.
• Parallel Performance: Leverages MPI with 3D Cartesian topology for efficient parallel execution across multiple cores or processors, enhancing computational performance.
• CMake Build System: Utilizes CMake for streamlined build processes and seamless cross-platform compatibility.
• VTK Output: Simulation results are saved in the VTK XML format, enabling efficient parallel output and facilitating multi-block visualization.

Getting Started

Prerequisites

• C++17 or higher
• CMake 3.15 or higher
• C++ compiler (e.g., g++)
• MPI library (e.g., OpenMPI)
• Python (for running scripts)

Clone the Repository

Clone the repository to your local machine:

git clone https://github.com/chenbei102/cbns3d_cart_mpi.git

Build the Project

1. Create and navigate to the build directory:

   mkdir build
   cd build

2. Generate build files with CMake:

   cmake <path/to/source/code>

3. Compile the project:

   make

Run the Solver

1. Edit the configuration file: Customize the .ini file to define simulation parameters. You can use python/taylor_green.ini as an example.

2. Parse the configuration file: Use the parser_ini.py script to parser the .ini file:

   python <path/to/source/code>/python/parser_ini.py <ini_file_name>

3. Run the solver: Execute the solver using mpirun. For example,

   mpirun -np <numProc> ./cbns3d_cart_mpi <dimX> <dimY> <dimZ>

   Replace <numProc>, <dimX>, <dimY>, and <dimZ> with your specific values.

   Note: Ensure <numProc> equals <dimX> * <dimY> * <dimZ>.

Visualization and Postprocessing

The solver outputs flow field data in the VTK XML format, organized as follows:

• MultiBlock file: output.vtm
• Individual block files: output???.vtr (one for each block)

These files are stored in directories named output???, corresponding to different time series outputs.

To visualize and analyze the results:

1. Open an output.vtm file using ParaView.
2. Perform visualization and postprocessing as required.