FEAST v2.0

Introduction:

FEAST is a Fugitive Emissions Abatement Simulation Toolkit for evaluating natural gas leak detection and repair (LDAR) programs. FEAST can be used to estimate natural gas savings and net present value of LDAR programs. Version 2.0 is an update of version 1.0 released in 2016. Version 2.0 was ported from Matlab to Python, a more advanced infrared camera simulation module was added, and several steps were taken to improve the speed and readability of the model's code.

Tutorial:

Clicking link below will open an active Jupyter notebook tutorial using the temporary notebook server system tmpnb. Once the Tutorial loads feel free to edit the notebook and see what FEAST can do (your changes will only be stored for 20 minutes of inactivity).

Tutorial

File Structure:

Python FEAST consists of a directory containing over 30 python module and object files. The file map below illustrates where each file is stored. The map is followed by a short description of the files.

FEAST
|----field_simulation.py
|----Glossary.txt
|----README.rst
|----DetectionModules
        |----__init__.py
        |----abstract_detection_method.py
        |----dd.py
        |----fid.py
        |----helper_functions.py
        |----ir.py
        |----null.py
|----GeneralClassesFunctions
        |----__init__.py
        |----leak_class_functions.py
        |----plotting_functions.py
        |----results_analysis_functions.py
        |----simulation_classes.py
        |----simulation_functions.py
|----InputData
        |----input_data_classes.py
        |----DataObjectInstances
                |----arpae_wind.p
                |----fernandez_leak_reapair_costs_2006.p
                |----fort_worth_leaks.p
                |----fort_worth_wind.p
                |----hitran_database.p
                |----pnnl_methane.p
        |----RawData
                |----ARPAEWind.csv
                |----FernandezRepairCost.csv
                |----FortWorth.csv
                |----FortWorthwindData.csv
                |----HITRAN_database.csv
                |----pnnl_methane.csv
        |----RawDataProcessingScripts
                |----arpae_wind_reader.py
                |----fernandez_repair_cost_reader.py
                |----fort_worth_data_prep.py
                |----fort_worth_wind_reader.py
                |----HITRAN_reader.py
                |----pnnl_reader.py
                |----repair_cost_data_reader.py
                |----wind_data_reader.py

File descriptions

field_simulation.py

contains one function of the same name (field_simulation). One call to field_simulation() creates one realization of a FEAST scenario. field_simulation() accepts several optional input arguments to change parameters from their default settings.

DetectionModules:

DetectionModules is the directory containing all of the LDAR program files:

abstract_detection_method.py

defines a parent class with the attributes and methods that all LDAR programs have.

helper_functions.py

contains a few short functions that are used by multiple LDAR programs.

null.py

defines the null detection method.

dd.py

defines a "Distributed Detector" LDAR program

ir.py

defines LDAR programs that use an infrared camera. An a manual subclass and an airborne, automated subclass are included.

fid.py

defines an LDAR program based on a flame ionization detector

GeneralClassesFunctions:

GeneralClassesFunctions contains files that define classes and functions that are not directly specified by LDAR programs or input data. Each module in the directory is described below:

leak_class_functions.py

defines the Leak class used to store all the data required to define a set of leaks. The module also contains function definitions used to create and manipulate leak objects.

plotting_functions.py

defines functions for plotting simulation results.

results_analysis_functions.py

defines functions that compile results from numerous realizations of a scenario to calculate mean net present value, detected leak size distributions and other statistics. plotting_functions.py calls results_analysis_functions.py to produce plots.

simulation_classes.py

defines classes that are necessary for a simulation. These classes are GasField, FinanceSettings, Atmosphere, Time and Results.

simulation_functions.py

defines functions that are necessary for a simulation but are neither part of a LDAR program nor methods of a class. The functions are listed below:

-sample_wr	Generates a list of random samples with replacement from a set.
-new_leak_count
	Calculates the number of new leaks to generate at a time step
-save_results	Generates a Results object at the end of a simulation and saves it.
-set_kwargs_attrs
	Allows any attribute specified in a class to be set using key word arguments
-gauss_leak_model
	Calculates the concentration of gas due to a leak at specified location and conditions.

InputData:

InputData is a directory containing raw data files, scripts for processing those raw data files and python object files created from the raw data. PyFEAST only uses the python object files, but the raw files and processing files are included for transparency and to allow for alternative processing files to be added in the future. The following list describes the subdirectories and class file in InputData.

input_data_classes.py

Defines all of the input data classes used by PyFEAST.

DataObjectInstances

Contains python data object files used by PyFEAST

RawData

Contains raw csv files for wind speed, leak data sets and other inputs to PyFEAST.

RawDataProcessingScripts

Contains the scripts used to produce the objects in DataObjectInstaces from the csv files in RawData.

Author:

Chandler Kemp https://github.com/ChandlerKemp

Acknowledgments:

JP Addison reviewed all code developed for the Python implementation of FEAST.