绿电直连视角下电-碳-算协同控制技术探究OA
Exploration of electricity-carbon-computing coordinated control technologies from perspective of green power direct connection
在"双碳"目标与"东数西算"工程推动下,绿电直连逐渐成为算力设施降低用能碳排放、促进新能源就地消纳的重要方式.然而,新能源出力波动、算力任务调节与碳排放约束相互耦合,传统分域管理模式难以兼顾供电安全、算力服务质量与低碳经济性.为厘清绿电直连场景下电-碳-算协同控制的技术体系,围绕体系架构、耦合建模、多时间尺度调度、动态碳核算及价值转化开展综述.研究表明,电-碳-算协同可形成物理运行、数据支撑、协同调控和市场价值4层架构;现有方法主要包括基础机理、多源不确定性和多主体博弈建模,并逐步形成日前—日内—实时相衔接的调度体系;算力任务迁移、储能与制冷协同调节及动态碳强度反馈,是促进绿电消纳和降低算力碳排放的关键手段.当前仍存在统一建模不足、跨时间尺度协调不充分、算力柔性量化不充分及市场机制衔接不完善等问题.未来应加强数据融合、智能调控、动态碳核算与工程标准化研究,为相关系统的规划、调度和低碳运行提供参考.
Driven by China's"dual carbon"goals and the"East Data,West Computing"initiative,green power direct connection has become an important approach for reducing carbon emissions from energy consumption in computing facilities and promoting local utilization of renewable energy.However,renewable generation output fluctuations,computing task regulation,and carbon emission constraints are strongly coupled,which makes the conventional sector-specific management mode difficult to simultaneously ensure power supply security,computing service quality,and low-carbon economic operation.To clarify the technical framework of electricity-carbon-computing coordinated control under green power direct connection,this review focuses on system architecture,coupling modeling,multi-time-scale scheduling,dynamic carbon accounting,and carbon value conversion.The review shows that electricity-carbon-computing coordination can be organized into a four-layer architecture consisting of physical operation,data support,coordinated regulation,and market value.Existing methods mainly include mechanism-based coupling models,multi-source uncertainty models,and multi-stakeholder game models,while a scheduling framework linking day-ahead,intraday,and real-time stages is gradually being developed.Migration of computing tasks,coordinated regulation of energy storage and cooling systems,and dynamic carbon intensity feedback are key measures for improving renewable energy utilization and reducing the carbon emissions of computing services.Nevertheless,current studies still face limitations in unified modeling,cross-time-scale coordination,quantitative characterization of computing flexibility,and market-mechanism linkage.Future efforts should focus on strengthening physics-data integration,intelligent regulation,dynamic carbon accounting,and engineering standardization.These findings provide a reference for the planning,scheduling,and low-carbon operation of related systems.
王慧;葛磊蛟;付天成;汤奇霖;吴宇阳;赵纹鉴
中国能源建设集团天津电力设计院有限公司,天津 300400天津大学 电气自动化与信息工程学院,天津 300072兰州交通大学 自动化与电气工程学院,兰州 730070天津大学 电气自动化与信息工程学院,天津 300072天津大学 电气自动化与信息工程学院,天津 300072兰州交通大学 自动化与电气工程学院,兰州 730070
能源科技
绿电直连算力负荷多时间尺度新型电力系统自适应协同碳价值转化动态碳核算算电协同
green power direct connectioncomputing loadmulti-time scalenew power systemadaptive coordinationcarbon value conversiondynamic carbon accountingcomputing-electricity coordination
《综合智慧能源》 2026 (9)
40-55,16
新能源系统高端装备和智慧运行教育部工程研究中心开放课题(KFKT-26ERC-02)国家自然科学基金项目(52277118,52577130) Open Project of the Engineering Research Center of New Energy System High-end Equipment and Smart Operation,Ministry of Education(KFKT-26ERC-02)Project Supported by National Natural Science Foundation of China(52277118,52577130)
评论