awesome-ml-power-systems

KO | EN

gitlite — search

#javascript #python #hacktoberfest #react #ai #typescript #llm #go #golang #android #machine-learning #rust #deep-learning #linux

awesome-ml-power-systems

★ 13 Open GitHub ↗

A curated collection of papers, datasets, software, benchmarks, and resources at the intersection of machine learning and electric power systems.

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# Awesome Machine Learning for Power Systems
[![Awesome](https://awesome.re/badge.svg)](https://awesome.re)

> A curated collection of papers, datasets, software, benchmarks, and resources at the intersection of machine learning and electric power systems.

This repository aims to provide researchers, students, and practitioners with a structured overview of the rapidly growing field of machine learning for electric power systems.

Unlike other paper lists, this repository emphasizes:
- Research categorization
- Reproducibility (code availability)
- Datasets and benchmarks
- Open-source tools
- Research gaps and future directions

Only research articles with publicly available source code and datasets are included in this collection. Articles whose reported results cannot be independently reproduced are removed as reproducibility assessments become available. If you find this repository useful, please consider starring it and citing it.

```bibtex
@misc{aryandoust2026awesome,
  author = {Aryandoust, Arsam},
  title = {Awesome Machine Learning for Power Systems},
  year = {2026},
  publisher = {GitHub},
  howpublished = {\url{https://github.com/ArsamAryandoust/awesome-ml-power-systems}}
}
```

## Introduction

Electric power systems are undergoing a profound transformation driven by:

- Renewable energy integration
- Electrification of transportation and heating
- Distributed energy resources
- Increased uncertainty in system operation
- Growing computational complexity

At the same time, machine learning has experienced rapid advances in:

- Deep learning
- Reinforcement learning
- Graph neural networks
- Generative AI
- Foundation models
- Scientific machine learning

This repository tracks research at the intersection of these two fields. Electrifying energy consumption and meeting that demand with renewable energy sources is a central strategy for mitigating global warming and climate change. This shared objective motivates much of the research conducted at this intersection.

## Getting Started

| Category | Resource |
|---|---|
| Survey | [Tackling Climate Change with Machine Learning](https://dl.acm.org/doi/10.1145/3485128) |
| Survey | [Physics-informed machine learning](https://www.nature.com/articles/s42254-021-00314-5) | 
| Tutorial | [Tutorial on amortized optimization](https://arxiv.org/abs/2202.00665) |
| Tutorial | [Deep Implicit Layers — Neural ODEs, Deep Equilibrium Models, and Beyond](https://implicit-layers-tutorial.org) |
| Textbook | [Probabilistic Machine Learning: An Introduction](https://probml.github.io/pml-book/book1.html) |
| Textbook | [Geometric Deep Learning: Grids, Groups, Graphs, Geodesics, and Gauges](https://geometricdeeplearning.com) |
| Textbook | [Reinforcement Learning and Optimal Control](http://www.athenasc.com/index.html) |
| Textbook | [Reinforcement Learning: An Introduction](http://incompleteideas.net/book/the-book-2nd.html) |
| Lecture | [MIT 6.S191 Introduction to Deep Learning](https://introtodeeplearning.com) |
| Lecture | [Stanford CS224W Machine Learning with Graphs](https://www.youtube.com/watch?v=JAB_plj2rbA&list=PLoROMvodv4rPLKxIpqhjhPgdQy7imNkDn) |
| Tool | [PyTorch](https://pytorch.org) |
| Tool | [JAX](https://docs.jax.dev/en/latest/) |
| Tool | [WandB](https://wandb.ai/site/) |

## Recommended Publication Venues

Unlike many other academic disciplines, computer science, and machine learning in particular, places especially strong emphasis on conference publications. Top-tier conferences often serve as the primary venues for disseminating high-impact research, in part because their shorter publication cycles are well aligned with the rapid pace of research and development in these fields.

For work at the intersection of machine learning, energy systems, and climate, the following venues are particularly relevant:

| Category | Publication Venue |
|---|---|
| Top-tier machine learning conference | [ICLR](https://iclr.cc/) |
| Top-tier machine learning conference | [ICML](https://icml.cc/) |
| Top-tier machine learning conference | [NeurIPS](https://neurips.cc/) |
| Artificial intelligence and machine learning journal | [Nature Machine Intelligence](https://www.nature.com/natmachintell/) |
| Energy journal | [Nature Energy](https://www.nature.com/nenergy/) |
| Energy journal | [Joule](https://www.sciencedirect.com/journal/joule) |
| Climate journal | [Nature Climate Change](https://www.nature.com/nclimate/) |
| Power systems journal | [IEEE Transactions on Smart Grid](https://ieee-pes.org/publications/transactions-on-smart-grid/) |
| Power systems journal | [IEEE Transactions on Power Systems](https://ieee-pes.org/publications/transactions-on-power-systems/) |
| Energy journal | [Applied Energy](https://www.sciencedirect.com/journal/applied-energy) |
| Data-focused journal | [Scientific Data](https://www.nature.com/sdata/) |

## Research Articles

### ML for power systems

| Year | Venue | Title | Application |
|--------|--------|--------|-------- |
| 2024 | NeurIPS (Datasets & Benchmarks) | [PowerGraph: A Power Grid Benchmark Dataset for Graph Neural Networks](https://proceedings.neurips.cc/paper_files/paper/2024/hash/c7caf017cbbca1f4b368ffdc7bb8f319-Abstract-Datasets_and_Benchmarks_Track.html) | Cascading failure forecasts based on grid state |
| 2025 | ICML | [Optimization Proxies using Limited Labeled Data and Training Time: A Semi-Supervised Bayesian Neural Network Approach](https://openreview.net/forum?id=Jnpgx8OzfD) | Learning surrogates/proxies for power-flow and power-system optimization |
| 2025 | NeurIPS (Datasets & Benchmarks) | [PFΔ: A Benchmark Dataset for Power Flow under Load, Generation, and Topology Variations](https://openreview.net/forum?id=Gi1HtsTAkv) | Benchmarking machine-learning methods for power-flow prediction and analysis |
| 2024 | ICML | [Compact Optimality Verification for Optimization Proxies](https://dl.acm.org/doi/10.5555/3692070.3692378) | Structuring a gradient-based primal heuristic formulation to verify and bound worst-case optimality gaps for DC-OPF neural network proxies. |
| 2024 | Google DeepMind | [OPFData: Large-scale datasets for AC optimal power flow with topological perturbations](https://arxiv.org/abs/2406.07234) | Large collection of solved instances for ACOPF with ten grid topologies |
| 2023 | ICML | [Low Complexity Homeomorphic Projection to Ensure Neural-Network Solution Feasibility for Optimization over (Non-)Convex Set](https://proceedings.mlr.press/v202/liang23a.html) | Feasibility projection for neural-network solutions to AC optimal power flow (AC-OPF) |
| 2021 | NeurIPS | [Adversarially Robust Learning for Security-Constrained Optimal Power Flow](https://proceedings.neurips.cc/paper/2021/hash/f0f07e680de407b0f12abf15bd520097-Abstract.html) | Learning-based optimization for security-constrained OPF under adversarial uncertainty |
| 2024 | IEEE Transaction on Smart Grid | [Space-Vector Neural Networks: Efficient Learning From Three-Phase Electrical Waveforms](https://ieeexplore.ieee.org/document/10510414) | Reconstruction of unbalanced three-phase voltage signals and the detection of cyber-attacks in digital substations |
| 2024 | Energy \& Buildings | [Benchmarking reservoir computing for residential energy demand forecasting](https://www.sciencedirect.com/science/article/pii/S0378778824003529) | Autoregressive electric load forecasting |
| 2022 | Nature Machine Intelligence | [Enhanced spatio-temporal electric load forecasts using less data with active deep learning](https://www.nature.com/articles/s42256-022-00552-x) | Spatial electric load forecasting |
| 2026 | Hugging Face | [AI.grids v1](https://huggingface.co/AI-grids) | A growing collection of standardized machine-learning datasets for power-system applications, including electricity consumption, renewable energy, and grid-related benchmarks designed to facilitate reproducible ML research.|
| 2024 | NeurIPS | [PowerPM: Foundation Model for Power Systems](https://proceedings.neurips.cc/paper_files/paper/2024/hash/d0a2279c9f7ded859bcbf878c3c3d1ed-Abstract-Conference.html) | Foundation model / representation learning for power-system tasks |
| 2024 | DeepMind | [Mixture of A Million Experts](https://arxiv.org/abs/2407.04153) | Pioneering mixture of experts model |
| 2022 | NeurIPS | [LIPS - Learning Industrial Physical Simulation benchmark suite](https://proceedings.neurips.cc/paper_files/paper/2022/hash/b3ac9866f6333beaa7d38926101b7e1c-Abstract-Datasets_and_Benchmarks.html) | A modular benchmark suite for evaluating data-driven augmented physical simulators on industrial use cases, such as power grids and pneumatics, across criteria including accuracy, efficiency, robustness, and real-world applicability.|

### ML with hard constraints

| Year | Venue | Title | Application |
|--------|--------|--------|-------- |
| 2026 | Arxiv | [ECO: Energy-Constrained Operator Learning for Chaotic Dynamics with Boundedness Guarantees](https://arxiv.org/pdf/2512.01984) | Kuramoto–Sivashinsky and Navier–Stokes equations |
| 2026 | Arxiv | [ENFORCE: Nonlinear Constrained Learning with Adaptive-depth Neural Projection](https://arxiv.org/abs/2502.06774) | Non-linear equality and ineqaulity constraints |
| 2026 | Arxiv | [RAYEN: Imposition of Hard Convex Constraints on Neural Networks](https://arxiv.org/abs/2307.08336) | Convex constraints |
| 2026 | Arxiv | [HardNet++: Nonlinear Constraint Enforcement in Neural Networks](https://ieeexplore.ieee.org/abstract/document/11579320) | Nonlinear constraints |
| 2025 | Arxiv | [HardNet: Hard-Constrained Neural Networks with Universal Approximation Guarantees](https://arxiv.org/abs/2410.10807) | piecewise constraints |
| 2025 | NeurIPS | [FSNet: Feasibility-Seeking Neural Network for Constrained Optimization with Guarantees](https://proceedings.neurips.cc/paper_files/paper/2025/hash/3874e2be479a9d4e94d4514046c1f934-Abstract-Conference.html) | Convex and nonconvex constraints |
| 2025 | Arxiv | [Enforcing Hard Linear Constraints in Deep Learning Models with Decision Rules](https://arxiv.org/abs/2505.13858) | Linear constraints |
| 2024 | Computers and Chemical Engineering | [Physics-informed neural networks with hard linear equality constraints](https://www.sciencedirect.com/science/article/abs/pii/S0098135424001820) | Linear equality constraints |
| 2023 | Arxiv | [POLICE: Provably Optimal Linear Constraint Enforcement for Deep Neural Networks](https://arxiv.org/abs/2211.01340) | Linear constraints |
| 2023 | Arxiv | [Data-Driven Control with Inherent Lyapunov Stability](https://arxiv.org/abs/2303.03157) | Lyapunov stability |

## Projects and Initiatives

| Name | Description |
|---|---|
| [AI.Grids](https://cresym.eu/ai-grids/) | A European initiative developing open, sovereign AI foundation models for electricity-grid operations, forecasting, planning, and management. |
| [GridFM](https://gridfm.org/) | An open collaboration between power-systems and AI researchers developing foundation models trained on diverse electric-grid data. |
| [PowerAgent](https://poweragent.seas.harvard.edu/) | An open-source community developing LLM-powered tools and agentic AI workflows for power-system analysis, planning, and operation. |

<!--
## Open-Source Software

| Tool | Description |
|--------|-------------|
| pandapower | Open-source power-system analysis framework |
| PyPSA | Python for Power System Analysis |
| MATPOWER | Power flow and optimization toolkit |
| Grid2Op | Benchmark platform for power-grid operation |
| PowerGridworld | RL environments for power systems |

## Open Research Challenges

The following research challenges remain largely open:

| Research Challenge | Description |
|---|---|
| Reliable Deployment | Many methods perform well in simulation but remain difficult to deploy in real-world power systems. |
| Benchmarking | The community lacks widely accepted benchmarks and evaluation protocols. |
| Generalization | Models often struggle under changing operating conditions. |
| Safety and Trustworthiness | Power-system operators require interpretable, robust, and reliable models. |
| Foundation Models for Power Systems | The role of large-scale pretrained models remains largely unexplored. |
| Offline Reinforcement Learning | Safe learning from historical operational data remains a major challenge. |

## Future Trends

Potential high-impact research directions include:

- Foundation models for grid operation
- Multi-agent energy systems
- AI agents for system operation
- Physics-informed generative models
- Foundation datasets for power systems
- Reliable uncertainty quantification
- Human-AI collaboration in control rooms
- Explainable and interpretable AI for power systems
- Neurosymbolic models for power systems
-->

awesome-ml-power-systems

Explore Similar Repositories

Was this content helpful?

Recent Feedback

Download README