A PV–Storage Rolling Optimization Model Based on the Temporal Fusion Transformer

作者：Sheng Zhuo, Jia Mei, Jingfang Xie, Tianlin Jiang, Jianrong Zhu · 年份：2025 · DOI：10.1109/icepg67373.2025.11466949 · 研究领域：Power Systems and Renewable Energy、Microgrid Control and Optimization、Optimal Power Flow Distribution

With the growing penetration of renewable energy resources (RESs) in the power system, addressing the technical and economic challenges of RESs are enhanced. This paper investigates the rolling bidding and scheduling for photovoltaic (PV)-storage power plants based on day-ahead (DA) and real-time (RT) markets. A rolling optimization model is developed under the Nordic sequential trading framework. On the basis of determining the day-ahead market clearing, the Temporal Fusion Transformer (TFT) is employed to forecast the real-time output of PV and quantify the prediction error, to dynamically adjust the compensation capacity and real-time output of storage, and to compensate for the fluctuation of PV generation. Further, the real-time bidding and scheduling strategy of the PV storage plant is recursively solved by rolling the calculation of the prediction time domain based on the latest predicted error. The effectiveness of the proposed method is verified by the case studies.