/ev-simulation-model

Primary LanguageJupyter Notebook

ev-simulation-model

This package implements a Gaussian Mixture Model to simulate charging session data of electric vehicles. In particular, the session duration [h] and the electricity demand [kWh] can be simulated.

Assume the following variables:

$d$: Charging duration in hours
$e$: Electricity demand in kWh
$z$: Plug-in hour of the day

The model can generate draws from the following distributions:

The joint distribution: $p(d,e)$
The marginal distributions: $p(d)$, $p(e)$, $p(z)$
The conditional distributions: $p(d|z)$, $p(e|z)$, $p(d,e|z)$

The distribution $p(d,e|z)$ is of particular interest because it allows to simulate tuples of charging duration and electricity demand given a certain plug-in hour of the day. The marginal distribution $p(z)$ can be leveraged in a population model where a fixed number of chargers is assumed to determine how many new vehicles plug-in during a certain time interval. For example, to calculate the plug-in probability between 16 and 17 o'clock you can calculate $$\int_{16}^{17} p(z) \thinspace dz \quad .$$ Multiplying the resulting probability with the number of chargers in the population then yields the absolute number of new plug-ins within this time interval.

Developer notes

This project uses the following tools to automate tedious tasks in the development process:

  • poetry: Reproducible dependency management and packaging
  • tox: Test and workflow automation against different Python environments
  • pytest: Unittests
  • black: Code formatting
  • isort: Code imports formatting
  • flake8: Code linting
  • mypy: Static type checking

Where is my setup.py / setup.cfg?

There is no setup.py / setup.cfg configuration file because this package is managed by Poetry's build system. This follows PEP 517 where setuptools is no longer the default build system for Python. Instead it is possible to configure build systems via pyproject.toml. Poetry provides a consistent configuration solely based on the pyproject.toml file. Read more details here: https://setuptools.pypa.io/en/latest/build_meta.html.

Usage

Installation

Install poetry on your machine if not yet installed. Install the project dependencies using:

poetry install

Development process

You can invoke all workflows using the tox CLI:

tox -e format # Format code
tox -e lint # Lint code
tox -e format,lint # Format first and then lint
tox -e python3.9 # Run pytest tests for Python 3.9 environment

# Run all workflows in logical order:
# format -> lint -> pytest against all Python environments
tox

Where to find the Python interpreter path?

It might be useful to find the Python interpreter path for integration with your IDE. Print the path using:

poetry env info --path

For usage from command-line directly run:

poetry run <command> # Will run any command within the Poetry Python virtual environment
poetry shell # Will start a new shell as child process (recommended)
source `poetry env info --path`/bin/activate # Activation in current shell