高比例新能源电网-分布式数据中心集群协同优化运行研究  

Coordinated operations of highly renewable power systems and distributed data centers

作　　者：李远征[1] 龙信鑫 周纯杰[1] 杨凯[2] 赵勇[1] 曾志刚[1] LI YuanZheng;LONG XinXin;ZHOU ChunJie;YANG Kai;ZHAO Yong;ZENG ZhiGang(School of Artificial Intelligence and Automation,Huazhong University of Science and Technology,Wuhan 430074,China;School of Electrical and Electronic Engineering,Huazhong University of Science and Technology,Wuhan 430074,China)

机构地区：[1]华中科技大学人工智能与自动化学院,武汉430074 [2]华中科技大学电气与电子工程学院,武汉430074

出　　处：《中国科学：技术科学》2024年第1期119-135,共17页Scientia Sinica(Technologica)

基　　金：国家自然科学基金(批准号:62073148,U21B2002);腾讯-犀牛鸟基金(编号:RAGR 20210102)资助项目。

摘　　要：随着新能源并网比例的逐步提升,电网对不确定性新能源的消纳态势日趋严峻,导致弃风弃光频现.数据中心作为新兴的重要电力负荷,其运行能耗高,规模增长迅速,算力及数据资源可灵活调配,具备可挖掘的柔性调节潜力.据此,本文利用分布式数据中心集群增强电力系统柔性调节能力,助力新能源消纳.首先,根据数据中心服务器运行工作状态和制冷设备温控要求,建立单个数据中心的精细化电力负荷模型.然后,考虑数据任务处理的时空可调节性,建立分布式数据中心互联互通的协调模型.在此基础上,提出考虑不确定性的电网-分布式数据中心集群的协同运行框架,研究集群调节能力的统一表征,建立两者顺序决策的双层优化模型,提出双层非线性区间优化方法进行求解.最后,通过算例分析验证了协同运行框架的有效性,得到了不同偏好条件下帕累托解集和分布式数据中心集群响应的变化规律.With the increase in the demand for renewable energy in power grids, the uncertainty of renewable energy poses a great challenge;moreover, it also leads to more frequent wind and solar power curtailments. As emerging and important loads of the power system,data centers have high energy consumption, rapid scale growth, and flexible allocation of computing power and data resources, which supports flexible adjustments. In this study, distributed data center clusters were employed to enhance the flexible adjustment capability of a power system and improve renewable energy absorption. First, a refined power load model for an isolated data center was developed according to the server workload and temperature requirements of the refrigeration equipment. Subsequently, a collaborative operation framework of distributed data center interconnection was established, considering the spatiotemporal adjustability of data processing. A collaborative operation framework considering the uncertainty of the power system and distributed data center cluster was proposed based on this. To study the unified regulation abilities of the cluster, a bi-level sequential optimization model was established. This model was solved by the proposed bi-level nonlinear interval optimization method. Finally,case studies were performed to verify the effectiveness of the collaborative operation framework. Thus, the Pareto solutions of the optimization model and responses of the distributed data center cluster under different preferences were obtained.

关 键 词：数据中心 柔性调节 新能源不确定性 协同运行 区间优化 

分 类 号：TP308[自动化与计算机技术—计算机系统结构] TM73[自动化与计算机技术—计算机科学与技术]