Can Time Series Foundation Models Serve as Surrogates for Power Systems Dynamic Simulation?
Organizations: Siemens Data & AI Research
Abstract
High-fidelity dynamic simulation is essential for scientific and engineering applications, but its computational cost limits large-scale scenario screening. We investigate the use of pretrained time-series foundation models (TSFMs) as computationally efficient surrogates to address this limitation. In-context learning and post-training strategies, including retrieval-augmented reference sample selection and physics-informed regularization, are considered in the investigation. Results are benchmarked against a traditional AI model representative of the state-of-the-art in the domain. Experiments using an open-source TSFM and data from a simulated 9-bus power system indicate promising results. Retrieval augmentation improves in-context learning performance, and fine-tuning leads to superior results compared to the traditional AI model, especially when including physics-informed regularization. The fine-tuned TSFMs achieve lower errors and improved computational costs compared to the traditional AI model, and a 563 speedup compared to the original physics-based simulation.
Figures & tables
| Method | (pu) | (deg) | (Hz) | ROCOF (Hz/s) | ||||
|---|---|---|---|---|---|---|---|---|
| MAE | CRPS | MAE | CRPS | MAE | CRPS | MAE | CRPS | |
| Chronos-2 Random | 0.008863 | 0.00488 | 1.1237 | 0.638 | 0.2163 | 0.140 | 0.09723 | 0.0724 |
| Chronos-2 RAG | 0.003780 | 0.00251 | 0.4200 | 0.243 | 0.1012 | 0.0647 | 0.09923 | 0.0553 |
| STGNN Traditional AI | 0.001094 | – | 0.2223 | – | 0.06900 | – | 0.01633 | – |
| Chronos-2 Standard FT | 0.000437 | 0.00140 | 0.06183 | 0.235 | 0.01091 | 0.039 | 0.007901 | 0.030 |
| Chronos-2 Physics FT | 0.000402 | 0.00120 | 0.06283 | 0.198 | 0.01013 | 0.037 | 0.007865 | 0.028 |
Neural Surrogate Modeling
Neural surrogate modeling uses learned models to approximate costly simulations, enabling faster prediction, design, control, and inference.
See all 466 papers