Skip to content

Latest commit

 

History

History
252 lines (195 loc) · 8.32 KB

README.md

File metadata and controls

252 lines (195 loc) · 8.32 KB
Raw

Parametric Plasma Source

Python package

ActionsCI

Python package, C++ source and build files for parametric plasma source for use in fusion neutron transport calculations with OpenMC.

The plasma source is based on a paper by C. Fausser et al

Installation

Installing from PyPI

pip install parametric_plasma_source

Installing from source

Installation of the parametric plasma source from source requires cmake to build the underlying C++ code. This can be obtained from your OS's package manager by e.g. sudo apt-get install cmake or from cmake source.

If you intend to develop the code then it is recommended to work in a virtual environment.

The requirements for developing the code can be installed by running:

pip install -r requirements-develop.txt

The package can be built and installed in editable mode by:

pip install -e .

Compiling from source

Compiling of the parametric plasma source from source is another method of using the software.

First clone the openmc repository into your home directory

cd ~
git clone https://github.com/openmc-dev/openmc.git

Then clone the parametric plasma source repository

git clone https://github.com/open-radiation-sources/parametric-plasma-source.git

Then create a build folder and compile from the folder. This assumes that openmc repository was cloned into your home directory /opt/openmc.

cd parametric-plasma-source
mkdir build
cd build
cmake .. -DOPENMC_DIR=/opt/openmc
make

This final make command will create a source_sampling.so file that can then be used.

Usage

The parametric plasma source can be sampled either directly in Python 3, sampled in an OpenMC simulation, or sampled using OpenMC via a standalone executable without simulation.

For a better understanding of the varibles take a look at the C. Fausser et al paper.

Sampling in Python

The parametric plasma source can be imported an used in Python 3 in the following manner:

from parametric_plasma_source import PlasmaSource
from random import random

plasma_params = {
    "elongation": 1.557,
    "ion_density_origin": 1.09e20,
    "ion_density_peaking_factor": 1,
    "ion_density_pedestal": 1.09e20,
    "ion_density_separatrix": 3e19,
    "ion_temperature_origin": 45.9,
    "ion_temperature_peaking_factor": 8.06,
    "ion_temperature_pedestal": 6.09,
    "ion_temperature_separatrix": 0.1,
    "major_radius": 9.06,
    "minor_radius": 2.92258,
    "pedestal_radius": 0.8 * 2.92258,
    "plasma_id": 1,
    "shafranov_shift": 0.44789,
    "triangularity": 0.270,
    "ion_temperature_beta": 6,
}

my_plasma = PlasmaSource(**plasma_params)
sample = my_plasma.sample([random(), random(), random(), random(), random(), random(), random(), random()])
particle_x, particle_y, particle_z = sample[0], sample[1], sample[2]
particle_x_dir, particle_y_dir, particle_z_dir = sample[3], sample[4], sample[5]
particle_energy_mev = sample[6]

Sampling in OpenMC

The parametric plasma source also contains a plugin library for OpenMC to allow the source to be sampled in an OpenMC simulation.

When using the OpenMC sampling the inputs must be provided in meters where applicable (the sampling will convert to cm).

from parametric_plasma_source import PlasmaSource, SOURCE_SAMPLING_PATH
import openmc

plasma_params = {
    "elongation": 1.557,
    "ion_density_origin": 1.09e20,
    "ion_density_peaking_factor": 1,
    "ion_density_pedestal": 1.09e20,
    "ion_density_separatrix": 3e19,
    "ion_temperature_origin": 45.9,
    "ion_temperature_peaking_factor": 8.06,
    "ion_temperature_pedestal": 6.09,
    "ion_temperature_separatrix": 0.1,
    "major_radius": 9.06,
    "minor_radius": 2.92258,
    "pedestal_radius": 0.8 * 2.92258,
    "plasma_id": 1,
    "shafranov_shift": 0.44789,
    "triangularity": 0.270,
    "ion_temperature_beta": 6,
}

my_plasma = PlasmaSource(**plasma_params)
settings = openmc.Settings()
settings.run_mode = "fixed source"
settings.batches = 10
settings.particles = 1000
source = openmc.Source()
source.library = SOURCE_SAMPLING_PATH
source.parameters = str(my_plasma)
settings.source = source
settings.export_to_xml()

Sampling using Executable

It is also possible to generate a source outside of OpenMC by creating the source_generator executable by running cmake -H. -Bbuild and then cmake --build build or cmake --build build --target source_generator. The source_generator can then be run as below:

Usage:
source_generator [OPTIONS]

Options:
 -l,--library      Source library, mandatory
 -n,--particles    Number of particles, default 1000
 -o,--output       Output directory, default {current directory}
 -v,--verbosity    Verbosity, default 5

This will use OpenMC commands to sample the source generated using the specified library with the specified number of particles and output the resulting initial_source.h5 file in the requested output directory. The initial_source.h5 can then be analysed to check the properties of the source being generated.

Running Tests

The tests are run by executing pytest tests from within your virtual environment.

C API for use with Fortran

A C API is provided for linking of the plasma source routine to Fortran. This is particularly useful for compilation with MCNP. To compile a static library and a test program a build script is provided in the parametric-plasma-source/fortran_api folder. This can be run in the following manner:

cd parametric-plasma-source/parametric_plasma_source/fortran_api
./build_lib.sh intel

for use with intel ifort and icpc compilers or

cd parametric-plasma-source/parametric_plasma_source/fortran_api
./build_lib.sh gnu

for use with the gnu gfortran and g++ compilers.

Use with MCNP

In order to use the library with MCNPv6.2 the plasma_source_module.F90 and mcnp_pp.F90 should be placed in the MCNP src folder. The source.F90 provided with MCNP should then be modified to:

subroutine source

  ! .. Use Statements ..
  use mcnp_interfaces_mod, only : expirx
  use mcnp_debug
  use pp_source_mk2_mod

  implicit none

  call parametric_plasma_2

  return
end subroutine source

The MCNP Makefile should also be updated to point to the library during linking. This can be done by adding the first line and modifying the second line in the Makefile:

PPLIB = -lplasmasource -L$(PLASMA_SOURCE)
COMPILE_LINE=$(LD) $(OUT_EXE)$(EXEC) $(F_OBJS) $(C_OBJS) $(ALL_LDFLAGS) $(PLOTLIBS) $(LIB_DMMP) $(EXTRALIBS) \
             $(PPLIB)

When compiling MCNP the PLASMA_SOURCE variable then needs to be set to the folder containing libplasmasource.a. An example of the compliation line would be:

make build CONFIG='intel openmpi omp plot' PLASMA_SOURCE=/plasma/source/dir/

The source parameters are passed using the rdum and idum cards in the mcnp input file

ion_density_pedistal           = rdum(1)
ion_density_seperatrix         = rdum(2)
ion_density_origin             = rdum(3)
ion_temperature_pedistal       = rdum(4)
ion_temperature_seperatrix     = rdum(5)
ion_temperature_origin         = rdum(6)
ion_density_peaking_factor     = rdum(7)
ion_temperature_peaking_factor = rdum(8)
ion_temperature_beta           = rdum(9)
minor_radius                   = rdum(10)
major_radius                   = rdum(11)
pedistal_radius                = rdum(12)
elongation                     = rdum(13)
triangularity                  = rdum(14)
min_toroidal_angle             = rdum(15)
max_toroidal_angle             = rdum(16)
        
number_of_bins                 = idum(2)
plasma_id                      = idum(3)

Note that idum(1) is intentionally left unused. This can be used for source selection if multiple user defined sources are to be compiled in the same executable.

The tests are run by executing pytest tests from within your virtual environment.