GEDM is a Python package developed to estimate the capacity and generation mixes of energy technologies in large-scale electricity systems, with a focus on renewables. It consists of an electricity dispatch model and a capacity expansion model which are formulated as mathematical programming problems. A framework is developed to control the program flow, access an external solver, and convert input and output data. This tool offers flexibility in the structure of the electricity market considered, time-slices, time periods and time horizon configuration, and all assumptions can be modified.

A model instance consists of a collection of CSV files that define the market structure, commodities/fuels, technologies/processes, resource potentials and assumptions and so on. This model framework imports these files, builds optimization problems, solves the problems, constructs result tables and exports to CSV files.

Key features of this framework:

• Flexibilities on structuring the model
• Integration of power system flexibility measures, e.g. storage system operation, interconnection and reserve services.
• Tests on system reliability with extreme supply and demand cases
• Dedicated input/output modules that convert the data into more readable table format
• Integration of VRE potential and hourly generation data based on an existing open access dataset provided by the author
• Freely available under the MIT license

This framework offers flexibility on configuring the following:

• market structure (market -> countries/districts -> zones)
• time-slices
• time period steps
• time horizon and base year
• most assumptions: techno-economic, fuel price, carbon cost, emissions limits, development limits etc.

GEDM has been tested on Windows. Running GEDM requires the following environment:

• the Python3 programming language (version 3.6 or greater)
• Python modules and their dependencies: Numpy, Pyomo
• installing a mathematical problem solver: GEDM has been tested with GAMS and GLPK (It is possible to use other solvers in this framework if Pyomo supports the interface.)
• installing Python API for GAMS when using GAMS language and solver

Using this tool by the following steps:

1. install the above requirements
2. download or clone the GEDM repository
3. check the solver option file "/Input/1_model_config/01_SolverConfig.csv", create a file folder for Pyomo
4. construct a case by editing the input files (you can start with the sample case without changing anything)
5. run the model $ Python scripts/GEDM.py
6. check the output files

Users are expected to have experiences on Python. The main reason is when infeasible solutions returned, the users may want to have control over the program flow and formulations to find the causes, other than changing the settings and assumption. In addition, Pyomo and solvers may occasionally have softeware compatible issues. It would require some Python experience to change the codes and solve it.

The overall framework and program flow is shown below. The electricity dispatch (ED) model (md_main.py, md_ED_*.py) deals with the generation dispatch decisions for all generators, taking into account operation constraints of conventional thermal plants and storage systems, variability of renewable generation, and inter-zone transmission. The ED model solves each daily dispatch problem individually. Essential outputs include CO2 emissions, fuel consumption, electricity flows, and cost etc.

The capacity expansion (CE) model (md_CE_*.py) decides the optimal new installation investment in each time period step. The results will be fed to the ED model for dispatch optimization. A critical feature is the mechanism of identifying extreme cases for testing system reliability. Before the program builds the CE problem, this framework selects the most extreme supply and demand hour in each calendar month, i.e. the hour with the highest residual demand. Therefore, this model optimizes the problem with configured time-slices plus these 12 cases.

An input data processing module (impt_*.py) is developed to convert the inputs according to the market structure, temporal resolution, time zone etc. Base year, time horizon, time periods and time-slices are configurable. The original system demand and renewable generation have to be given in hourly resolution to allow such conversion. Another data processing module (*_output.py) is constructed to extract model solutions and convert the results into readable tables in CSV files.

The ED and CE models are the core components formulated into linear programming problems. Details about the model formulation can be seen in this document.

Most settings and assumptions are separated from the codes. Users can freely design theirs models. The tables below summarize the purpose of the inputs files used to construct the models.

Model configuration (document)

folder path: /Input/1_model_config/

Commodity/fuel assumptions (document)

folder path: /Input/2_commodity/

Process/technology assumptions (document)

folder path: /Input/3_process/

Policy assumptions (document)

folder path: /Input/4_policy/

Existing capacity and calibration (document)

folder path: /Input/5_calibration/

Transmission (document)

folder path: /Input/6_transmission/

Electricity demand profile (document)

folder path: /Input/7_demand/

Renewables (document)

folder path: /Input/8_VRE/

The data processing module extracts solutions returned from the solver, summarizes the results, and exports as readable tables in CSV files. Every model instance has coorresponding output files. There are also files to summarize the results through the time horizon.

(document)

/Output/(Country Code)/

This package is licensed under the MIT License.