Energy System Time Series Suite (ESTSS) - Data Archive

Energy System Time Series Suite - Data Archive

This archive contains variously sized sets of declustered time series within the context of energy systems. These series demonstrate low discrepancy and high heterogeneity in feature space, resulting in a roughly uniform distribution within this space.

For detailed information, please refer to the corresponding GitHub project:
https://github.com/s-guenther/estss/

For associated research, see
https://doi.org/10.1186/s42162-024-00304-8

Data is provided in .csv format. The GitHub project includes a Python function to load this data as a dictionary of pandas data frames.

Should you utilize this data, kindly also cite the associated research paper. For any queries, please feel free to reach out to us through GitHub or the contact details provided at the end of this readme file.

Folder Content

• `ts_*.csv`: Contains declustered load profile time series in tabular format.
  • Size: `(n+1) x (m+1)`, with `n` representing time steps (1000 per series) and `m` the number of series.
  • Includes a header row and index column. Headers indicate series id, and the index column numbers each time step, starting from `0`.
  • The first half of the series `(m/2)` consistently display a constant sign (negative). They are sequentially numbered from 0.
  • The second half `(m/2)` display varying signs. Numbering starts from `1,000,000`.
• `features_*.csv`: Tabulates features corresponding to the time series.
  • Size: `(m+1) x (f+1)`, where `m` is the number of time series and `f` is the number of features
  • Includes a header row and index column. Indexes represent time series id (matching `ts_*.csv` headers), and headers name the features.
• `norm_space_*.csv`: Shows feature vectors in normalized feature space where time series are declustered. Provided for completeness; typically not needed by users.
  • Size: `(m+1) x (g+1)`, where `m` is the number of timer series and `g` is the number of selected features space features. (a subset of `f` from `features_*.csv`).
  • Format matches `features_*.csv`.
• `info_*.csv`: Maps declustered datasets to the manifolded dataset. Provided for completeness; typically not needed by users.
  • Size: `(m+1) x 2`, with `m` as series count. Columns contain manifolded set time series ids.
  • Includes an index column and a header. The index holds the remapped id of declustered series. Header `0` is non-significant.

Each `ts_*.csv`, `features_*.csv`, `norm_space_*.csv`, and `info_*.csv` file comes in four versions to accommodate various set sizes:

• `*_4096.csv`
• `*_1024.csv`
• `*_256.csv`
• `*_64.csv`

These represent sets with 4096, 1024, 256, and 64 time series, respectively,offering different densities in feature space population. The objective is to balance computational load and resolution for individual research needs.

Contact

ESTSS - Energy System Time Series Suite
Copyright (C) 2023
Sebastian Günther
sebastian.guenther@ifes.uni-hannover.de

Leibniz Universität Hannover
Institut für Elektrische Energiesysteme
Fachgebiet für Elektrische Energiespeichersysteme

Leibniz University Hannover
Institute of Electric Power Systems
Electric Energy Storage Systems Section

https://www.ifes.uni-hannover.de/ees.html