面向算力-电力协同的数据中心与储能融合机制综述OA
Review of Integration Mechanism of Data Centers and Energy Storage for Computing Power-Electric Power Coordination
人工智能的迅猛发展导致算力需求和数据中心负荷快速增长,使算力与新能源的高效协同成为必然发展趋势,但这一协同面临算力需求用电特性与新能源出力特性时段错配、数据中心供电稳定性风险加剧,以及匹配算力负荷特性的电力市场机制缺失等核心困难.储能技术可以改善算力与新能源协同时空特性,提升数据中心的供电可靠性和负荷灵活性,并为数据中心参与电力市场提供新机遇.科学高效的数据中心与储能融合机制是算力-电力协同的重要研究课题.文中立足算力-电力协同视角,全面总结数据中心与储能融合机制的研究现状.首先,聚焦数据中心内部配建储能(内部配储)场景,从规划和运行等多方面归纳数据中心合理配建储能、实施有效调控的思路与方案;其次,探讨数据中心共用外部公用储能(外部共储)场景,讨论共享储能和租赁储能等典型模式的交易机制及数学方法;最后,从发展形式、模型构建和交易架构3个方面探讨数据中心分享私有储能(分享私储)的可行方案与研究路径.通过研究现状分析,探索数据中心与储能融合发展新业态,可以为破解数据中心绿色用能难题、推动算力-电力高效协同提供参考.
The rapid development of artificial intelligence has led to a surge in computing power demand and data center loads,making the efficient coordination between computing power and renewable energy an inevitable trend.However,this faces core difficulties such as temporal mismatch between the electricity consumption characteristics of computing power demand and the output characteristics of renewable energy,increased risks in data center power supply stability,and the lack of electricity market mechanisms tailored to load characteristics of computing power.Energy storage technology can improve the spatiotemporal characteristics in the coordination of computing power and renewable energy,enhance the power supply reliability and load flexibility of data centers,and provide new opportunities for data centers to participate in electricity markets.Scientific and efficient integration mechanisms between data centers and energy storage represent a critical research topic for the computing power-electric power coordination.From the perspective of computing power-electric power coordination,this paper comprehensively summarizes the state of the art in integration mechanisms between data centers and energy storage.Firstly,focusing on the scenario of configuring in-site energy storage for data centers(in-site energy storage configuration),it outlines ideas and schemes for rationally configuring and effectively regulating energy storage in data centers from planning and operating perspectives.Secondly,the scenario of sharing off-site public energy storage for data centers(off-site shared energy storage)is explored,and transaction mechanisms and mathematical methods for typical modes such as shared energy storage and leased energy storage are discussed.Finally,feasible solutions and research pathways to share private energy storage for data centers(shared private energy storage)are discussed from three aspects of development pattern,model construction and transaction architecture.By analyzing the current research situation,a new business model of integrating data centers and energy storage is explored,which can provide reference for solving the problem of green energy consumption in data centers and promoting efficient computing power-electric power coordination.
唐文虎;梁秀壮;郭采珊;张泽宇;陈伟威
华南理工大学电力学院,广东省 广州市 510641华南理工大学电力学院,广东省 广州市 510641南方电网能源发展研究院配电网规划研究部,广东省 广州市 510663华南理工大学电力学院,广东省 广州市 510641华南理工大学电力学院,广东省 广州市 510641
算力-电力协同数据中心新能源电力市场储能用能
computing power-electric power coordinationdata centerrenewable energyelectricity marketenergy storageenergy consumption
《电力系统自动化》 2026 (17)
37-51,15
国家自然科学基金资助项目(52577109)广东省重点领域研发计划资助项目(2023B0909040003).This work is supported by National Natural Science Foundation of China(No.52577109)and Key Area Research and Development Program of Guangdong Province(No.2023B0909040003).
评论