PEAKTK is a toolkit designed to help researchers, utility companies, and organizations predict the peak day of the month and peak hours of the day of the grid/micro grid. The toolkit is written in Python 3.

Quote from the paper PEAKTK paper explaining the goal of the toolkit:

Our goal is to improve reproducibility of energy systems research by providing a common framework for evaluating and comparing new peak forecasting algorithms. Further, PeakTK provides libraries to enable researchers and practitioners to easily incorporate peak forecasting methods into their research when implementing higher level grid optimizations.

In this toolkit, we provide reference implementations of a range of peak forecasting techniques as follows:

• Peak day of the year prediction algorithms
  • CPP Approach
  • Stopping Approach
  • Probabilistic Approach
• Peak day of the month prediction algorithms
  • Extreme temperature
  • VPeak
  • Probabilistic Approach (monthly)
• Peak hour of the day prediction algorithms
  • LSTM-based hourly probabilistic classification
  • LSTM-based hourly demand prediction
• Peak demand prediction
  • LSTM-based demand forecasting
  • SARIMA
  • SVR-based demand forecasting

PeakTK also includes reference energy datasets for for experimentation and quantitiave comparisons as follows:

• ISO New England dataset
• ESO dataset
• Smart* Apartment dataset

All datasets come with weather dataset

Here is the list that we currently are working on and will finish it:

• Fixing pip install. We are trying to make sure that it work on Apple M1.
• Updating the unified interface. (A few algorithms left)
• Adding Sphinx docs
• Adding example for all type of peak forecasting algorithms.

Here is the list of things that we think it would be great to have:

• Support scikit learn's hyperparameter tuning
• Support walk forward validation
• And, of course, have more algorithms

If you would like to contribute, please feel free to send a pull request and email us!

If you found bug or need help with the toolkit, please feel free to email [us] ([email protected]).