The Grid Event Signature Library (GESL) initiative at DOE’s Oak Ridge National Laboratory (ORNL) and Lawrence Livermore National Laboratory (LLNL) is focused on the development of the well-defined, curated, and free-to-access power grid data repository with the goals of advancing the field of machine learning and artificial intelligence (ML/AI) for the grid and facilitating swift response against malfunctions of grid infrastructure.

Accelerating ML/AI-enabled capabilities of the electric grid is one of the core missions of the DOE. The goal of the GESL is to develop a community resource of curated signature data for use in power system AI algorithm development.

• The GESL houses accurately and precisely labeled power system event signatures.
• Signature data is easily accessible, free to use, and publicly available.
• The GESL will leverage existing mature technologies where possible and practical, such as open-source databases and visualization tools.

• In the power systems industry, there are scarce few databases that have proper event labeling along with public accessibility. Other existing public data sets of grid signatures often lack critical metadata or only contain a single or very few examples of an event type, and data formats vary widely across these data sets. Furthermore, because of the dynamic nature of the electric grid structure, data sets should evolve over time to update themselves with new signal waveforms and characteristics, which is a major shortcoming of current data sets.
• ORNL and LLNL have partnered with universities and private sector industry partners to develop a framework, the GESL, that attempts to address these challenges. The GESL is an expandable database architecture for power system event signatures from devices that monitor different assets on the power system. The GESL can be used for the development and testing of new ML algorithms for interactive signature identification, matching, and predictive analytics.

The GESL currently consists of both Phasor Measurement Unit (PMU) data and Point-on-Wave (PoW) data captured from 10 distinct, anonymized providers across the United States. These waveforms span a variety of time durations, sampling rates, voltage levels, and measurement parameters such as voltage and current oscillography, frequency, and rate-of-change-of-frequency (ROCOF).

Users can use the standard GESL GUI within the Dashboard to download waveform data as well as the provided API commands. The API and associated instructions can be found at Applications/API

No. Users can use free tools like Python and Octave along with commercial software like MATLAB to post process and analyze the downloaded waveforms.

Please send any and all inquiries to geslsupport@ornl.gov.

Publications include not only papers, but also presentations for conferences/meetings or educational purposes. All documents and papers that report on research that uses the GESL Data Set may cite the following:

ORNL, LLNL, September 14, 2023, "Grid Event Signature Library", GESL, https://gesl.ornl.gov.
ORNL, LLNL. (2023). Grid Event Signature Library. GESL. https://gesl.ornl.gov
ORNL, LLNL. (2023). "Grid Event Signature Library." https://gesl.ornl.gov.
ORNL, LLNL. "Grid Event Signature Library." https://gesl.ornl.gov

@misc{GESL2023Dataset_ORNL_LLNL,
title = { {Grid Event Signature Library} },
author = { {Oak Ridge National Laboratory and Lawrence Livermore National Laboratory} },
howpublished = {Available: \url{https://gesl.ornl.gov/} }
,}

Plot(s) on the dashboard are compressed and randomly omit data points to provide an approximate visual representation due to data size and/or resource/application limitations. When analyzing GESL data, dashboard plots should be taken as approximations only, therefore data should be downloaded for complete analysis.