On Probabilistic Flexible Regulation Capability Forecasting for Electric Vehicles

• Published in: 2024 IEEE/IAS 60th Industrial and Commercial Power Systems Technical Conference (I&CPS) pp. 1 - 6
• Main Authors: Ding, Zhaohao, Lu, Qiyuan, Wei, Xuan, Li, Xinran, Guo, Shan, Gong, Chengming, Hu, Yingjian, Han, Yaning
• Format: Conference Proceeding
• Language: English
• Published: IEEE 19.05.2024
• Subjects: Bayesian neural network; Elevtric vehicles; flexible regulation capability; gate recurrent unit; Predictive models; probabilistic forecasting; Probabilistic logic; Probability distribution; Regulation; Renewable energy sources; Uncertainty; Weather forecasting
• ISSN: 2158-4907
• DOI: 10.1109/ICPS60943.2024.10563654

The uncertainties arising from the high penetration of renewable energy generation in the power system may result in a supply-demand gap, especially during peak load periods, wherein electric vehicles (EVs) can play a significant role with noteworthy regulation capability. This paper proposes a Bayesian gate recurrent unit (GRU) approach to predict the probability distribution of the flexible regulation capability (FRC) from EVs. Firstly, the FRC of an EV cluster is depicted based on the charging characteristics of individual EV. Secondly, the Spearman correlation coefficient is employed in the feature extraction step to construct the inputs of the prediction model. Furthermore, the GRU is embedded in the Bayesian neural network to fully take account of the EV charging behavior, ultimately yielding the probability distribution of FRC from EVs. Numerical experiments with real-world data from the Inner Mongolia region are performed to verify the effectiveness of the proposed method.