Power System Intelligent Emergency Frequency Control Based on Predictive Boosting Learning
IEEE Transactions on Automation Science and Engineering · 2026 · 23 · 11316-11327
DOI: 10.1109/tase.2026.3697776 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Transactions on Automation Science and Engineering, 2026. Verfasst von Tao Wang und 5 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref). Gefördert von: National Natural Science Foundation of China, Science and Technology Innovation Program of Hunan Province, Hunan Provincial Innovation Foundation for Postgraduate.
Abstract (Originalsprache)
The high penetration of renewables and the low-inertia characteristics of modern power systems pose a major threat to system frequency stability. While a multitude of emergency frequency control schemes have been developed, how to efficiently achieve reliable system frequency control against diverse events in practical large-scale systems remains a significant challenge. Faced with this issue, this study develops an intelligent emergency frequency control (IEFC) scheme via predictive boosting learning. Distinct from empirical task-specific solutions, this scheme establishes a generic “prediction-action search” paradigm that can be flexibly extended to various safety-critical control tasks. Inspired by the predictive control idea in modern control theory, a deep regression network (DRN) is first designed to accurately infer the system frequency stability level after the intervention of combinatorial control actions built upon multiple active power control resources. With the aid of the DRN’s prediction capability, the proposed IEFC scheme can efficiently evaluate the effects of various candidate stability control actions. Owing to this forward-looking capability, the IEFC scheme is able to rapidly screen all possible control actions and select the optimal one for execution before frequency instability occurs. Furthermore, to improve the prediction accuracy of the DRN, a boosting learning algorithm that systematically assembles a group of relatively weak regression models into a much stronger regression model is carefully devised. With the incorporation of this algorithm, the IEFC scheme achieves strong reliability and adaptability for system frequency control against diverse complicated scenarios. In addition, a specialized exit mechanism is integrated to guarantee the scheme’s robustness even under extreme uncertainties. Comprehensive numerical tests are carried out on a realistic regional AC/DC hybrid power system. The results showcase the superiority of the proposed IEFC scheme over competitive baselines in terms of prediction accuracy, control reliability, and computational efficiency, verifying its great potential as a transferable automation paradigm. Note to Practitioners—This paper is motivated by the challenging problem of implementing emergency frequency stability control with both high reliability and efficiency for practical power systems, but it also applies to emergency control tasks involving other stability issues. Following a large disturbance with a significant power imbalance, a practical power system faces a high risk of frequency instability. Due to the complex nonlinear characteristics of the system, it is difficult for practical frequency control schemes to achieve timely and reliable frequency control under various emergency conditions. In this paper, candidate frequency control actions are proactively formulated by intelligently coordinating heterogeneous emergency frequency control resources. When a fault occurs, a deep learning model efficiently predicts the effects of different candidate control actions on system frequency stability, thereby enabling the selection of the optimal one to achieve rapid and precise frequency control. The efficacy of the proposed approach is validated through comprehensive simulations on a realistic regional power grid. However, some implementation details of the formulated control scheme, such as the optimization of control parameters for individual resources, are not considered in the current framework. Future work will focus on refining the proposed approach by further taking such implementation details into account.
Förderung
- National Natural Science Foundation of China
- Science and Technology Innovation Program of Hunan Province
- Hunan Provincial Innovation Foundation for Postgraduate
Zitieren
APA
Wang, T., Zhu, L., Hou, J., Deng, C., Li, J., & Zhang, C. (2026). Power System Intelligent Emergency Frequency Control Based on Predictive Boosting Learning. IEEE Transactions on Automation Science and Engineering, 23, 11316-11327. https://doi.org/10.1109/tase.2026.3697776
Vancouver
Wang T, Zhu L, Hou J, Deng C, Li J, Zhang C. Power System Intelligent Emergency Frequency Control Based on Predictive Boosting Learning. IEEE Transactions on Automation Science and Engineering. 2026;23:11316-11327. doi:10.1109/tase.2026.3697776
BibTeX
@article{Wang2026,
author = {Wang, Tao and Zhu, Lipeng and Hou, Jie and Deng, Chao and Li, Jiayong and Zhang, Cong},
title = {Power System Intelligent Emergency Frequency Control Based on Predictive Boosting Learning},
journal = {IEEE Transactions on Automation Science and Engineering},
year = {2026},
volume = {23},
pages = {11316--11327},
doi = {10.1109/tase.2026.3697776}
}
EUVIMED