OPFData (GridOpt) Release 1: 300k Solved AC Optimal Power Flow Problems per PGLib Grid with Load Perturbations

Provider
Google DeepMind (Lovett et al.)
License
CC-BY-4.0 (source)
Stored at
Google Cloud Storage, gs://gridopt-dataset/dataset_release_1 (us)
Access
Free and public. No account with the data provider is needed.
Category
Energy
Source
https://arxiv.org/abs/2406.07234

About this dataset

DeepMind GridOpt (OPFData), a large dataset of solved AC optimal power flow (AC-OPF) problems for training graph neural network and other machine learning OPF solvers. Base grids come from the PGLib-OPF benchmark library and range from small IEEE test systems to large GOC and PEGASE networks; the folder listing shows grids such as pglib_opf_case118_ieee, pglib_opf_case10000_goc and pglib_opf_case13659_pegase. Problems were solved with PowerModels.jl and the IPOPT interior-point solver, and each grid holds 300,000 solved examples per perturbation type, split into 20 tar.gz shards (about 317 MB each for case118, 19 GB for case10000, 30 GB for case13659). This folder is the load-perturbation release: each example perturbs the bus loads of the base grid and stores the network as a heterogeneous graph (buses, generators, loads, shunts, AC lines, transformers) with the optimal generator dispatch, voltages and branch flows as targets in JSON. PyTorch Geometric reads it through torch_geometric.datasets.OPFDataset. The topology-perturbed companion is dataset_release_1_nminusone. Licensed CC-BY-4.0 by DeepMind (LICENSE file in the bucket); cite Lovett et al. 2024, OPFData (arXiv:2406.07234), and the PGLib-OPF sources.

How to access this dataset in Python

import gcsfs
fs = gcsfs.GCSFileSystem(token='anon')
fs.ls('gridopt-dataset/dataset_release_1/', detail=False)[:10]

Lists the files with anonymous access. Runs on any machine, with no cloud account.

Data structure

Files

Example file: dataset_release_1/pglib_opf_case10000_goc_0.tar.gz

Analyze this dataset without downloading it

On Scigantic, OPFData (GridOpt) Release 1: 300k Solved AC Optimal Power Flow Problems per PGLib Grid with Load Perturbations opens in a hosted JupyterLab notebook with the data already mounted, so you can read files where they are stored instead of copying them first. Each notebook includes an AI assistant that can see the mounted files and write and run analysis code. Open the dataset page to launch a notebook, or browse the full catalog.

This page is also available as Markdown and as JSON.

Related Energy datasets