电力信息与通信技术2026，Vol.24Issue(3)：33-44,中插1-中插2,14.DOI:10.16543/j.2095-641x.electric.power.ict.2026.03.05
考虑负载时空转移特性的数据中心集群算电协同优化调度
Computing-power Synergistic Scheduling of Internet Data Centre Clusters Considering Spatio-temporal Load Transfer Characteristics
摘要
Abstract
With the growing number of high-energy Internet Data Centers(IDCs)being connected in cities,a key challenge is to design a scientifically grounded computing-power synergistic scheduling paradigm that ensures highly reliable electricity supply for IDCs while fully exploiting their flexibility to reduce energy use and support secure,stable grid operation.Yet existing studies remain insufficient both in characterizing the adjustable features of computing networks and in designing synergistic scheduling and market transaction mechanisms.To address this gap,this paper proposes an optimal computing-power synergistic scheduling method for IDC clusters that explicitly considers spatiotemporal load transfer.First,we develop a flexibility-modeling approach for IDC-cluster resources that accounts for equipment-level operational couplings and provides a systematic,quantitative description of the spatiotemporal regulation capabilities of three classes of flexible resources:IT workloads,cooling/air-conditioning systems,and energy storage.Second,we propose a coordination scheme for IDC-cluster participation in grid operations that integrates demand response and frequency-regulation services.On one side,real-time Distribution Locational Marginal Prices(DLMPs)published by the Distribution System Operator(DSO)guide the Data Center Aggregator(DCA)in planning intra-cluster compute-workload migration;on the other,idle reserve storage capacity across the cluster is aggregated to participate in frequency-regulation transactions,leveraging storage's fast,bidirectional ramping capability.We then construct a bi-level DCA-DSO collaborative optimization model.Finally,case studies demonstrate significant gains in frequency-regulation revenues,reductions in electricity procurement costs,and improvements in green-power utilization,thereby validating the effectiveness and engineering applicability of the proposed model and method.关键词
数据中心/算电协同/时空负载均衡/可调能力估计/调频辅助服务市场/投标调度Key words
internet data centers/computing-power synergistic operation/spatiotemporal load balancing/flexibility estimation/frequency regulation ancillary services market/bidding and scheduling分类
信息技术与安全科学引用本文复制引用
向梓旸,黄淳驿,李康平,梁钢,王涛,李雅洁..考虑负载时空转移特性的数据中心集群算电协同优化调度[J].电力信息与通信技术,2026,24(3):33-44,中插1-中插2,14.基金项目
国家自然科学基金项目(52407126) （52407126）
上海市科委科技创新行动计划(25692113000). （25692113000）
