基于混合监督式学习的输电断面功率极限快速评估方法A fast evaluation method of power limit of transmission section based on hybrid supervised learning
张静,叶琳,刁瑞盛,徐建平,吕勤
ZHANG Jing,YE Lin,DIAO Ruisheng,XU Jianping,LYU Qin
摘要(Abstract)：
快速、准确地实时评估电网输电断面功率极限，对于电网安全、稳定、经济运行至关重要。为此，提出一种基于混合监督式学习算法的自适应方法，通过对历史运行方式的大量仿真分析和特征提取，训练并定期更新人工智能模型，形成电网运行状态与功率极限的准确匹配关系，从而快速、准确地实时评估输电断面功率极限。所提方法采用多种监督式学习预测算法，包括深度神经网络、支持向量回归、梯度提升决策树、随机森林，使用混合模型进行功率极限预测可充分发挥不同算法优势。该方法的有效性在具有真实运行特性的500节点输电网模型中得到验证。
The quick,accurate and instantaneous evaluation of power limit of transmission section is of great significance to safe,stable and economic operation of power grid. The paper proposes an adaptive method based on a hybrid supervised learning algorithm. Through numerous simulations and analysis of historic operation modes,characteristic extraction,training,and regular updating of AI model,the correct matching relationship between grid operation status and power limit is established for fast,accurate and instantaneous power limit evaluation of the transmission section. The proposed method employs multiple supervised learning forecast algorithms,including deep neural networks,support vector regression,gradient boosting decision tree,and random forest. The hybrid models for power limit forecasting can make best of the advantages of the algorithms. The effectiveness of the method is verified on a 500-bus transmission model with real operation characteristics.
关键词(KeyWords)：
人工智能;输电断面;机器学习;功率极限预测;电网运行方式
AI;transmission section;machine learning;power limit forecast;grid operation condition
基金项目(Foundation): 国网浙江省电力有限公司科技项目（5211JH1900M4）
作者(Author):
张静,叶琳,刁瑞盛,徐建平,吕勤
ZHANG Jing,YE Lin,DIAO Ruisheng,XU Jianping,LYU Qin
DOI: 10.19585/j.zjdl.202209009
参考文献(References)：
- [1] MAMUN A,SOHEL M,MOHAMMAD N,et al.A comprehensive review of the load forecasting techniques using single and hybrid predictive models[J]. IEEE Access,2020,8:134911-134939.
- [2] NATARAJAN V,KARATAMPATI P. Survey on renewable energy forecasting using different techniques[C]//Proceedings of the 2nd International Conference on Power and Embedded Drive Control(ICPEDC).Chennai,India:IEEE,2019.
- [3] DIAO R,VITTAL V,LOGIC N. Design of a real-time security assessment tool for situational awareness enhancement in modern power systems[J].IEEE Transactions on Power Systems,2010,25(2):957-965.
- [4] DIAO R,SUN K,VITTAL V,et al.Decision tree based online voltage security assessment using PMU measurements[J].IEEE Transactions on Power Systems,2009,24(2):832-839.
- [5] DIAO R,WANG Z,SHI D,et al. Autonomous voltage control for grid operation using deep reinforcement learning[C]//Proceedings of IEEE PES General Meeting. Atlanta,GA,USA:IEEE,2019.
- [6] ZHANG B,LU X,DIAO R,et al. Real-time autonomous line flow control using proximal policy optimization[C]//Proceedings of the IEEE PES General Meeting.Montreal,Canada:IEEE,2020.
- [7] YAN Z,XU Y.Data-driven load frequency control for stochastic power systems:A deep reinforcement learning method with continuous action search[J].IEEE Transactions on Power Systems,2019,34(2):1653-1656.
- [8] HUANG Q,HUANG R,HAO W,et al.Adaptive power system emergency control using deep reinforcement learning[J].IEEE Transactions on Smart Grid,2019,11(2):1171-1182.
- [9] HAARNOJA T,ZHOU A,ABBEE P,et al.Soft actorcritic:off policy maximum entropy deep reinforcement learning with a stochastic actor[C]//Proceedings of ICML.Stockholm,Sweden:IMLS,2018.
- [10] KUNDUR P.Power system stability and control[M].New York:McGraw-Hill,1994.
- [11] DIAO R,SHI D,ZHANG B,et al.On training effective reinforcement learning agents for real-time power grid operation and control[C]//Proceedings of NeurIPS.Online:Connectionist,2020.
- [12] WANG S,DIAO R,LAN T,et al.A DRL-aided multilayer stability model calibration platform considering multiple events[C]//Proceedings of the 2020 IEEE PES General Meeting. Online:IEEE,2020.
- [13] LAN T,DUAN J,ZHANG B,et al. AI-based autonomous topology control for maximizing time-series available transfer capabilities considering uncertainties[C]//Proceedings of the 2020 IEEE PES General Meeting. Online:IEEE,2020.
- [14] WANG S,DIAO R,XU C.On multi-event co-calibration of dynamic model parameters using soft actor-critic[J].IEEE Transactions on Power Systems,2021,36(1):521-524.