Chronos-2 improves peak load forecasting across grid levels
Peak-Aware Short-Term Load Forecasting Across Distribution Grid Aggregation Levels
Machine Learning
Summary
Power companies need to predict electricity demand, especially during times when many people use a lot of power at once, to prevent problems like overloads. The authors tested different forecasting tools on electricity data from the UK and Switzerland to see how well they predict these peak times. They found that a newer model called Chronos-2 is better at forecasting high-demand periods at different parts of the distribution network. This can help operators better manage the grid and avoid issues caused by sudden peak loads. The authors also showed that the method works quickly enough to be used in real-world situations.
What this means in practice
- •For distribution system operators: Improve accuracy of electricity demand forecasts during peak periods to reduce grid congestion and voltage issues across various network aggregation levels.
- •For energy grid software developers: Incorporate Chronos-2 models for faster and more accurate short-term load predictions to enhance grid management tools and decision support systems.
Authors
Souhardya Chattopadhyay, Julian Oelhaf, Antonia Schoening, Jessica Deuschel, Bitan Bhattacharyya, Christian Bergler, Andreas Maier, Siming Bayer
Abstract
For distribution system operators, short-term load forecasting (STLF) supports congestion management, voltage control, and asset protection. Most existing approaches focus on overall accuracy across all time steps and neglect performance during high-demand (HD) periods, where larger forecast errors can increase the risk of congestion and voltage violations. In this paper, we study peak-aware STLF across three operator-relevant distribution grid aggregation levels, area codes (AC), secondary substations (SUB), and low-voltage (LV) feeders, using open datasets from the United Kingdom and Switzerland. We compare statistical baselines, machine learning models (LightGBM and XGBoost), and recent time-series foundation models (Chronos Bolt and Chronos-2) under a peak-aware evaluation framework that reports both overall and HD forecasting performance using NMAE and MAPE. The results show that Chronos-2 achieves the best HD performance across all aggregation levels, with HD-NMAE and HD-MAPE of 0.039 and 4.53% at AC, 0.080 and 9.45% at SUB, and 0.138 and 16.14% at LV, while Chronos-Bolt consistently ranks second best. Compared with the gradient boosted ML models, Chronos-2 reduces mean HD-NMAE by about 20-51% across levels while remaining best or near-best on the overall metrics. A quantile analysis of the probabilistic Chronos outputs further identifies aggregation-specific operating points, and runtime measurements indicate that foundation model inference is fast enough for practical deployment. Overall, the findings highlight peak-aware evaluation and aggregation specific quantile selection as a practical pathway toward more operationally relevant STLF in distribution networks.