Smart Energy Management

A repository for a Smart Energy Management solution to optimize purchase of energy on the one-day-ahead market based on a price forecast and learnings of historical data of charging events using machine learning methods and optimization tools.

Environment

See requirements.txt

Optimization task

0. setup variables
1. objective
2. start x0
3. bounds
4. constraints
5. MINIMIZE OPTIMIZATION

$$ J = \min_{E_s(t)} \text {cost of energy for the household} $$

6. resulting vector
7. plot results
8. safe results

Figures

1. Plots of the covariance matrix
2. Plots of optimal energy purchase based on price forecast

Results

json and csv for scheduled energy purchase

Utils

1. data_preparation.py for drop of useless variables
2. extract_prices.py to make use of the historical data and one day ahead forecast
3. battery_distribution.py to estimate the battery capacity to be charged at the next day given historical data
4. max_power_distribution.py to estimate the maximal charging power for a given vehicle based on historical data
5. plug_distribution.py to estimate when the vehicle is plugged a necessary condition to optimize charging
6. data_exploration.py for first plots, covariance matrix and filters