SES-House

This is the source code of a project we developed in the course Smart Energy Systems at the TU Berlin, look here for course description.

Requirements

• Python
• Anaconda
• Gurobi

Data

• ./sample: should contain the sample files for test functions
• ./data: should contain the dataset files for the forecasting and optimization

We use data from PecanStreet in order to do the forecasting and optimization. We use PecanStreet for PV power generation, house loads and price data.

Programming

Run code

In order to run the code you need to have the files mentioned in the configuration, e.g. configs/default.ini.

Create our conda environment and activate it:

conda env create -f conda.yml
conda activate ses-house

Run our default example:

python code/simple_model.py configs/default.ini

Run our Vermont example:

python code/simple_model.py configs/Vermont.ini

Run our Forecasting PV example:

python code/forecast_pv.py

Run our Forecasting using the saved model example:

python code/forecast_pv.py 1

Run our Forecasting Load example:

python code/forecast_load.py

Check the code/forecast_load_conf.py and code/forecast_conf.py to adapt the configuration.

Get started

The following instructions should work Linux environments. To get started, install anaconda. Therefore follow the installation instruction on this website.

make install
conda activate ses-house

To be sure that the code is working do:

make test

When editing the code do this command to format the code:

make black

When something weird happens and for cleanup:

make clean

When you add or remove packages from our ses-house conda environment, update the environment file conda.yml with this command:

make export

Help

If you would like to try this out and have some questions, comments or improvements, feel free to write us or open an issue.