点击次数：

论文类型：期刊论文

第一作者：Houbo Xiong

通讯作者：Chuangxin Guo

合写作者：Mingyu Yan,Yi Ding,Yue Zhou

发表刊物：Applied Energy

收录刊物：SCI

刊物所在地：英国

学科门类：工学

一级学科：电气工程

文献类型：J

卷号：333

页面范围：120578

ISSN号：0306-2619

关键字：Concentrated solar power plants；Wind energ；yPower scheduling；Dynamic programming；Multi-stage robust optimization；Thermal energy storage；Robust dual dynamic programming

DOI码：10.1016/j.apenergy.2022.120578

发表时间：2023-03-01

影响因子：11.446

摘要：The concentrating solar power plants (CSP) have well potential in coordinating with the ever-increasing wind energy during power scheduling. However, the existing studies individually design the day-ahead or intra-day optimization of coordinated scheduling between CSP and wind power, which makes the scheduling decisions not optimal in terms of economic and environmental benefits. Additionally, the non-anticipativity of scheduling decisions are not considered in most of them. This paper proposes a novel dynamic programming (DP) formulated multi-stage robust reserve scheduling (DPMRS) model, which is the first attempt to realize the day-ahead and intra-day joint optimization for coordinated scheduling of CSP and wind power. Under the framework of multi-stage adaptive robust optimization (ARO), DPMRS model enforces the non-anticipativity of scheduling. Besides, a convex modelling technique for thermal energy storage (TES) is presented to ensure the tractability of DPMRS model, whose effectiveness is proved mathematically. Moreover, to efficient solve the DPMRS model, a robust dual dynamic programming with accelerated upper approximation (RDDP-AU) solution methodology is developed, and the mathematical proof for its convergence is provided. Numerical studies on the modified IEEE RTS-79 system and a real-world system in Northwest China validate the effectiveness of the proposed scheduling model and solution methodology. The simulation results demonstrate the DPMRS model brings a 17.22% reduction in scheduling cost, and reduces 57.39% curtailment of renewable energy. Compared with the conventional algorithm, the RDDP-AU significantly reduces the computational consumption by 87.56%, and with the error less than 0.074%.

发布期刊链接：https://www.sciencedirect.com/science/article/pii/S0306261922018359