Power-grid stability predictions using transferable machine learning
Problem
Assessing the dynamic stability (frequency synchronization) of power grids via numerical simulation is computationally too expensive. The study asks whether machine learning can estimate node synchronization stability more cheaply and transferably.
Why it matters
Unpredicted grid instability can trigger blackouts affecting millions of people.
What the AI did
Trained random forest, support vector machine, and artificial neural network models on synthetic power grids, then tested transfer to real-world national grids.
Results
Three ML models predicted node synchronization stability better when trained on heterogeneous rather than homogeneous input-power distribution; real-grid tests (Great Britain, Spain, France, Germany) truncated in abstract.
Impact
Abstract reports better prediction from heterogeneous training and application to four real national grids; no adoption evidence.
Record
| Venue | Chaos An Interdisciplinary Journal of Nonlinear Science |
|---|---|
| Published | 2021 |
| Work type | article |
| Authors | Seong-Gyu Yang, Beom Jun Kim, Seung‐Woo Son, Heetae Kim |
| Citations | 22 (OpenAlex, at import time) |
| References | 70 |
| Model or method | random forest, support vector machine, artificial neural network |
| Organization | not stated in source |
| Primary source | https://doi.org/10.1063/5.0058001 |
| Open access | green |
| Keywords | power grid, synchronization stability, machine learning, transferability, random forest, neural networks |
Importance rating
significant — thousands to millions affected
Abstract addresses synchronization stability of real-world national power grids (Great Britain, Spain, France, Germany); grid stability is critical infrastructure, though blackout impact is not quantified in the abstract.
More in Engineering & Materials
| # | Paper | Year | Citations | Importance |
|---|---|---|---|---|
| 1 | Scaling deep learning for materials discovery | 2023 | 1,471 |
8
|
| 2 | Forecasting Renewable Energy Generation with Machine Learning and Deep Learning: Current Advances and Future Prospects | 2023 | 378 |
8
|
| 3 | Application of Machine Learning in Material Synthesis and Property Prediction | 2023 | 165 |
7
|
| 4 | Machine learning-guided discovery of ionic polymer electrolytes for lithium metal batteries | 2023 | 115 |
7
|
| 5 | Machine Learning for Structural Health Monitoring of Aerospace Structures: A Review | 2025 | 44 |
7
|