Monte Carlo Solution to Coulomb Collisions using IPPL

The repository to my semester project in computational physics in FS24. The complete report can be found here.

RUN THE CODE

After installing IPPL, paste the dsmc-code folder inside the downloaded ippl folder. In that folder add the code to the CMakeLists.txt:

..............
    FetchContent_MakeAvailable(googletest)

    add_subdirectory (unit_tests)
endif ()

# add these to lines relatively at the end
add_subdirectory (05.03.dsmc/dsmc-code-simple)
message (STATUS "Added 05.03.dsmc source.")

configure_file (${CMAKE_CURRENT_SOURCE_DIR}/cmake/${PROJECT_NAME}Config.cmake.in
    ${CMAKE_CURRENT_BINARY_DIR}/${PROJECT_NAME}Config_install.cmake )

install (
..............

Go to the terminal and run the following commands

module load gcc/11.4.0 cmake/3.26.3 cuda/12.1.1 openmpi/4.1.4 
./ippl-build-scripts/999-build-everything -t serial -k -f -i -u
./ippl-build-scripts/999-build-everything -t openmp -k -f -i -u
./ippl-build-scripts/999-build-everything -t cuda -k -f -i -u

Choose how you want to compile the code. Then there are three possibilities to run the code (inside the build_... folder):

1. ./Nanbu convergence 5 5.0 Nanbu 200 0.49049 false --info 10: Choose the Trubnikov test (with "convergence"), 5 is the number of realizations, 5.0 is the final time of the simulation, Nanbu is the collision model (also possible: TakAbe, Nanbu2 and Bird), 200 is the number of particles, 0.49049 is $\nu_0\Delta t$ and false inidicates no debug output.
2. ./Nanbu delta 64 5 5.0 Nanbu 200 0.001 200 false false false --info 10: Choose delta "function" as initial distribution, a 64^3 grid, 5 realizations, 5.0 as the final time, Nanbu as the collision algorithm, 200 as number of particles, 0.001 as $d_x$ (explained in the report), and three bools (use adaptive mesh grid, use collisions and output debug).
3. ./Nanbu sphere 32 156055 1000 true true false true Nanbu 327.59496 0.0 1.0 10 --info 10: Choose the "Cold Sphere Heating" testcase, a 32^3 grid, 156055 particles, 1000 timesteps, four bools (use self consisten electrical field, compute collisions, output debug, use adaptive mesh), Nanbu is the collision algorithm, 327.59496 is the timestepsize, 0.0 is a initial velocity scaling factor, 1.0 is a multipliert to the confinement force and 10 is the number of realizations.

Make sure to create a data folder in the same folder as the Nanbu executable.

Reproduce Report Plots

All plots created for the report are produced using Python 3 and the Jupyter Notebooks listed here. They are separated by test case and contain all code necessary. The same folder also contains all data files (mostly .csv) generated using the command described above with the parameters mentioned in the report at the respective section.

Final Project Presentation

The final results were presented at PSI on August 22nd, 2024. LaTeX code and PDF for the final presentation can be found here.