期刊
IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS
卷 56, 期 5, 页码 5880-5889出版社
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/TIA.2020.2986186
关键词
Batteries; State of charge; Economics; Reactive power; Indexes; Artificial intelligence; Frequency control; Battery swapping station (BSS); fast frequency regulation service (FFRS); policy gradient (PG); reinforcement learning
资金
- U.S. Natural Science Foundation [1800716]
Battery swapping stations (BSSs) have great potential in providing fast frequency regulation service (FFRS) owing to their large battery storage capacity. However, compared to a regular BSS, a BSS providing FFRS faces the following financial risks: first, higher infrastructure investment to support the vehicle-to-grid services; second, higher battery aging costs due to FFRS; third, FFRS causes uncertainties to batteries' charging costs. Under such a context, in this article, we propose an economic risk assessment model for the BSS-based FFRS by comparing its economics with a regular BSS. In this model, the value at risk (VaR) of daily revenue and the long-term return on investment (ROI) of a regular BSS and a BSS providing FFRS are compared. The assessment results are obtained in three steps: first, we develop the operation and economic models for the BSS-based FFRS. Next, the mathematical models of the VAR and ROI analyses are formulated. Finally, the VaR of daily revenue is compared through a statistical analysis of a large number of scenarios. The ROI comparison is conducted by a policy gradient based reinforcement learning algorithm, which can handle the nonconvexity and stochastic dynamics brought by the electric vehicle visits. The practicality of the proposed framework is demonstrated by using the real ancillary market data from utilities and the traffic count data from onsite traffic sensors.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据