gt-frc / neutpy

NeutPy - A code for the calculation of neutral particle transport in plasmas based on the transmission and escape probability method

NeutPy is the Python 2.x port of GTNEUT, which was written by John Mandrekas.

GTNEUT is a two-dimensional code for the calculation of the transport of neutral particles in fusion plasmas. It is based on the Transmission and Escape Probabilities (TEP) method and can be considered a computationally efficient alternative to traditional Monte Carlo methods. The code has been benchmarked extensively against Monte Carlo and has been used to model the distribution of neutrals in fusion experiments.

The original physics background can be found at

Mandrekas, John. (2004). GTNEUT: A code for the calculation of neutral particle transport in plasmas based on the Transmission and Escape Probability method. Computer Physics Communications. 161. 36-64. 10.1016/j.cpc.2004.04.009.

The original FORTRAN 95 GTNEUT code can be found at The Fusion Research Center GitHub at https://github.com/gt-frc/GTNEUT

Installation

Triangle Installation

The Triangle 2D mesh generator is required for NeutPy mesh generation. This guide can be used to install Triangle locally. If you imagine using triangle otherwise, consider following the steps below but with consideration for global installation (e.g., cloning to /opt/ instead of your home directory). Ensure that you have a C compiler installed. Download the Triangle zip file from https://www.cs.cmu.edu/~quake/triangle.html or

$ cd ~

$ git clone https://github.com/libigl/triangle.git

$ cd triangle

Make your bin directory

$ mkdir bin

Read the README file for instructions on how to compile. It's pretty basic. We recommend simply compiling triangle alone with (using GCC) since we do not use showme.

$ gcc -O -o bin/triangle triangle.c -lm

If you want to fully compile triangle and showme, edit the makefile, noting any special options from the README.

Keep SRC = ./ and set BIN = ./bin/

Make triangle

$ make

After triangle is compiled, set executable

$ cd bin

$ sudo chmod +x triangle

Set link (this allows triangle to be called on command line as triangle) to /usr/local/bin or some other directory on your PATH.

$ sudo ln -s /full/path/to/triangle /usr/local/bin/triangle

Install NeutPy

Using pip

$ pip install neutpy

From GitHub

If you'd like to work on the actual neutpy code, you can clone from GitHub:

$ cd /your/future/neutpy/home/

• Master branch

Clone master branch from github

$ git clone https://github.com/gt-frc/neutpy.git

• Other branches

You can clone another branch from github as follows:

$ git clone -b <branch> https://github.com/gt-frc/neutpy.git

Enter NeutPy

$ cd neutpy

Setup your virtual environment (install virtualenv using apt, yum, etc.)

$ virtualenv --python=/usr/bin/python2.7 venv

Activate it

$ source venv/bin/activate

Install packages

$ pip install -r requirements.txt

Usage

NeutPy requires 6 input files and 2 configuration files.

Configuration files:

toneutpy.conf is the main input file for a shot. This file will change with each shot.

neutpy.conf is the main NeutPy configuration file and includes parameters that probably wouldn't change between shots and other variables. This file needs to be in the current working directory and cannot be renamed. Grab an example copy from the FRC GitHub (https://github.com/gt-frc/neutpy/)

Data files:

The names of the data files included follow the GT3 gt3_shotid_timeid_profile.dat convention but can be defined differently in your toneutpy.conf file.

Ion/Electron density and temperature data are X/Y (normalized rho/value) two-column data. Temperatures are in keV. Densities should be given in #/m^3. Psi data are non-normalized 3-column R/Z/value data, with R/Z in meters.

Example File Structure

gt3_shotid_timeid_ne.dat (Electron density profile)

gt3_shotid_timeid_ni.dat (Ion density profile)

gt3_shotid_timeid_Te.dat (Electron temperature profile)

gt3_shotid_timeid_Ti.dat (Ion temperature profile)

gt3_shotid_timeid_psirz.dat (Non-normalized psi grid)

gt3_diid_wall.dat (Machine wall coordinates (R/Z))

Example Usage

Import the neutrals class.

>>> from neutpy import neutrals

There are three main entry points into NeutPy: from_file, from_mesh, and from_gt3.

from_file()

>>> neuts = neutrals()
INITIALIZING NEUTPY
>>> neuts.from_file('/relative/path/to/your/toneutpy.conf')

where the filename is relative to the CWD.

from_gt3()

>>> neuts = neutrals()
INITIALIZING NEUTPY
>>> neuts.from_gt3(<GT3 object>)

The from_mesh method is in development.

neutpy's Issues

Need to verify if Triangle is installed

Need some way to check to see if Triangle is installed long before it is called so that we don't waste time setting up the meshing if Triangle can't be found.

Neutpy still fails on negative triangularity unit test

As the title says.

Density/temperature gradients may take too long

The density and temperature gradients take a good deal of time compared to the rest of the pre-work for neutpy in _core_nT(). Can it be made faster?

Need more in-depth unit testing

Currently, the unittests simply run the whole of neutpy via from_file. This probably should be done in a more granular manner.

Need robust way to handle minimum wall temp/densities

Currently, the minimum wall densities and temperatures are hardcoded at 1E15 and .02KeV prior to transport into the SOL. These should be handled in a more robust manner.

X-point search needs to try to look for multiple x-points

Currently, neutpy works for single-null reactors and only searches for 1 x-point in _get_sep_lines(). Can/should be made more robust in the future.

Wall R0 should be read in or determined automatically

Currently, the R0 of a wall segment is hardcoded at 1.4m. This should be determined correctly.

R_max needs to be determined

In the SOL particle diffusion, r_max is chosen arbitrarily but affects the neutral density calculation. The change isn't substantial but is worth investigating.

Python 3 port

This shouldn't be too hard to port to Python 3. Print statements and exception handling are the main issues. I believe shapely cut() can be omitted/replaced now?